| 0,1,2,3,4,5
|
| 2710,github_trending,https:
|
| 2623,hackernews,https:
|
| 2622,hackernews,https:
|
| 2620,hackernews,https:
|
| 2619,hackernews,https: |
| There are 4 stages:<p>1. Surprised: the first time I see the AI become valid in the real world. It is the Claude 3 family moment. I am with a god-tool in my hand.<p>2. Oversimulated: Use the god-tool to solve all the problems without thinking.<p>3. Frozen: Realize that the god-tool can solve only a few issues, and my brain is dying.<p>4. Wake-up: Understand that my brain is the main place where the problem-solving happens, not the god-tool. Then cultivate by jumping into every detail in my work with the god-tool in my hand. It's not at the word-by-word level; it's at the meaning-by-meaning level, where I can say: I have done it by myself. In other words, I re-take the thinking ownership.",2026-09-04T07:00:13.177782
|
| 2618,hackernews,https://www.stillmade.shop,Show HN: StillMade – Programmable AI video pipeline; Chat to video (now),"Show HN: StillMade – Programmable AI video pipeline; Chat to video (now). Six months of 7+ hour days building this because video production is still a multi-tool mess. Describe what you want to make and StillMade guides you through making it with its tools.<p>I wanted full control without the platform I'm using holding me back or jumping to another (sometimes paid) tool every time I needed something specific.<p>What I'm building toward: every step becomes a programmable block. Build any Adobe, Figma or DaVinci feature your video needs, chain them into pipelines and own your workflow. Timely stuff gets streamlined.<p>App is temporarily focused on helping creators stay consistent on YouTube.<p>Takes 10 seconds to start using. Core pipeline works today. Programmable blocks are next. Free up to a certain usage + BYOK for API keys in settings.<p>Even if you spend 30 seconds generating an image or starting a project and tell me anything you thought, that means everything right now. Would you actually use this if programmable blocks were built out?<p>If you try it and give feedback or thoughts on the concept after using, I'm giving lifetime access (or any perk that becomes important if StillMade keeps growing) to a handful of early users who help shape it.<p>Feel free to ask me anything about StillMade and thanks for having a look!",2026-09-04T07:00:13.171830
|
| 2617,hackernews,https://www.oconeeruntime.com/news/policy-enforcement-for-browser-ai-and-coding-agents,Show HN: Oconee Runtime – Policy enforcement for browser AI and coding agents,"Show HN: Oconee Runtime – Policy enforcement for browser AI and coding agents. AI agents are changing the enterprise security boundary. Prompt-level controls are no longer enough when AI can execute commands, modify files, invoke tools, and act against enterprise resources.",2026-09-04T07:00:13.166714
|
| 2615,hackernews,https://github.com/openvurp/openvurp,Show HN: Openvurp – A wallet of AI agents that use tools and consult each other,Show HN: Openvurp – A wallet of AI agents that use tools and consult each other. ,2026-09-04T07:00:13.156901
|
| 2614,hackernews,https://github.com/fosrl/pangolin,Show HN: Pangolin – SSO and WireGuard instead of API keys for LLM access,"Show HN: Pangolin – SSO and WireGuard instead of API keys for LLM access. You have seen Pangolin (<a href=""https://github.com/fosrl/pangolin"" rel=""nofollow"">https://github.com/fosrl/pangolin</a>) on here before for our open-source zero-trust network access suite of tools. We are launching a new AI gateway that takes a different approach to auth. Rather than authenticating requests to AI providers with API keys, it authenticates the network connection itself via a desktop app.<p>Each user gets a WireGuard tunnel back to the gateway, established after they log in through their existing identity provider (Okta, Azure, Google, etc) via a desktop app. That tunnel is the auth, so there is no key to generate, embed, rotate, or leak, because the gateway already knows who's on the other end of the connection. Compare that to most AI gateways, which are really just a proxy in front of static bearer tokens you'd get from OpenAI or Anthropic directly.<p>This tunnel-based approach also means self-hosted and on-prem models aren't an afterthought bolted onto the gateway. They work the same way public cloud models do. Drop a lightweight tunnel connector inside your cluster (even a DGX Spark) and it joins the same network, so on-prem models show up in the gateway next to public cloud ones. Users switch between them without changing endpoints or juggling separate credentials for internal infrastructure.<p>For machines or users who can't run the desktop client, we still support traditional virtual keys, plus the usual gateway features: session logging, budgets, usage analytics, and governance.<p>It’s fully self-hostable. Community Edition is free for everyone. Enterprise Edition is free for individuals and small businesses. There is also Pangolin Cloud which is fully hosted.",2026-09-04T07:00:13.148995
|
| 2613,hackernews,https://dumbdetector.net/,Show HN: DumbDetector – Community signals for AI tool problems,Show HN: DumbDetector – Community signals for AI tool problems. ,2026-09-04T07:00:13.144355
|
| 2612,hackernews,https://github.com/ylxdzsw/mu,Show HN: Mu – an agent with actual command line experience,"Show HN: Mu – an agent with actual command line experience. I'm an old school user that finds the AI agent TUIs too magical, so I experimented with a new agent UX. I implement mu as a shell plugin (zsh and fish for now). Pressing Tab enters agent mode, where |
| 1. the command line prompt is changed to displays model name and context usage, and |
| 2. a preexec hook sends commands to an agent instead of interpreting them as shell script.<p>Thats it. Each LLM prompt starts a regular process that reads stdin and writes stdout, with no terminal magics. The actually executed commands appear literally in zsh_history, just like other shell commands. This allows a seamless switch between normal shell work and agentic work, with mixed history that uses regular scrollback buffer.<p>I also experimented with a minimum agent design, where only a single bash tool is provided. Special commands are provided for reliable file editing and multi-modal input. So far it worked well for me, as a day to day command line helper.",2026-09-04T07:00:13.140225
|
| 2611,hackernews,https://ardent.ai/,"Show HN: Ardent, a code-first agent for non-engineering work","Show HN: Ardent, a code-first agent for non-engineering work. I’m Nate, the founder of Ardent. We just shipped our public beta, and we’d love your thoughts!<p>Ardent is an agent running in a desktop (Electron) app built to help with knowledge work, designed for less-technical people outside engineering. Think Codex or Claude Cowork, but built around collaboration and customization.<p>I know, I know, it’s yet another agent harness! Ardent is a little different – it leans heavily on codegen to solve problems. Most agents are basically just a bag of tools and a while loop, but Ardent writes code to complete tasks, making it much faster and more token-efficient. [0]<p>Lots of companies are trying to build their own agent harness right now, because token budgets make software appear cheap, and it seems like a competitive advantage to own a custom agent.<p>The thing is, agents are deceptively complex. They aren’t light wrappers around the model anymore; the harness makes a big difference in how capable the agent is. They also come with lots of sticky problems: context management, sandboxing and permissions, authz/authn to remote services, etc.<p>Instead of building your own harness from scratch, you can customize Ardent using abilities: TypeScript libraries that can expose tools and prompts which can be used by the agent. They’re like agent skills, but closer to regular code libraries: type-safe, importable, and with well-defined permissions. Ardent can help you build abilities, but you can also create them by hand if you prefer.<p>Anyone can share abilities with everyone in their organization with a single click, and if Ardent notices that the user is doing repeatable work, it offers to transform the adhoc solution into a reusable ability. Say you prepare the same project update every Friday, pulling issues from Linear and docs from Google Drive. Ardent can turn that workflow into a reusable ability that you can share with everyone in your company.<p>Your company’s ability catalog is like a standard library. To solve a problem, Ardent searches for existing abilities, and then writes glue code to compose them. Because the code is ephemeral, it doesn’t need to be maintainable, which lets us use cheaper, less-capable open weight models like Kimi and GLM. [1]<p>Code executes on the user’s machine in a Deno sandbox with zero permissions by default. Abilities declare required permissions up front, and any code that exceeds the current scope gets terminated and prompts the user for approval.<p>Ardent can also connect to a bunch of tools which you’re already using, like Google, Linear, and so on. We support MCP servers, but in our experience, MCP is slow, token-expensive, and provides a pretty bad experience for both agents and users. [2]<p>Instead, Ardent connectors handle auth, but instead of using MCP, we generate client shims which call the service’s primary API. This is a much better fit for Ardent’s codegen approach, because the agent can read the client and treat it like any other library.<p>There’s a bunch of other interesting stuff in Ardent: memory, conversational recall, scheduled tasks, prompt suggestions, and so on. During our beta period, we’re selling inference at-cost, and we’re giving each organization some free credits to experiment.<p>Please take a look and tell us what you think! We’d love your feedback, whether you think what we’re doing is good, bad, or meh. I’m also happy to answer any questions that you might have.<p>---<p>[0] Cloudflare has taken to calling this “code mode”: <a href=""https://blog.cloudflare.com/code-mode/"" rel=""nofollow"">https://blog.cloudflare.com/code-mode/</a><p>[1] In our benchmarks, Ardent consumed 63% fewer tokens than Opus running in Claude Cowork, with half the model requests.<p>[2] MCP servers rarely define input/output contracts, and it’s hard to write code that calls a function when you don’t know its arguments, or even its arity! Here's a good read on the subject: <a href=""https://www.blacksmith.sh/blog/code-smith-code-mode"">https://www.blacksmith.sh/blog/code-smith-code-mode</a>",2026-09-04T07:00:13.132247
|
| 2610,hackernews,https://www.wired.com/story/flock-ai-search-user-interface/,Flock's AI Search Tool for Cops,Flock's AI Search Tool for Cops. ,2026-09-04T07:00:13.127582
|
| 2608,hackernews,https://www.theverge.com/ai-artificial-intelligence/989435/nvidia-pair-personal-ai-router-home-local-llm-compute-tool-rtx-macbook,Nvidia launches free tool that links idle computers into personal AI datacenter,Nvidia launches free tool that links idle computers into personal AI datacenter. ,2026-09-04T07:00:13.117031
|
| 2604,hackernews,https://news.ycombinator.com/item?id=49557020,One-person startup: a Cornell postdoc wants to put AI in everyone's hands,"One-person startup: a Cornell postdoc wants to put AI in everyone's hands. I run my own AI startup, and I've built three AI products over the past eight months, from writing the code to landing the first paying users. I did it as a solo founder.<p>I came to AI from an unusual direction. I studied clinical medicine, then surgery, and could have become a doctor. During my PhD I turned to AI and stayed with it: I've published a fair number of papers at top AI conferences like ACL, CVPR, EMNLP, and AAAI, hold several patents, and did a postdoc at Cornell.<p>A record like that usually points to one path: stay in academia, aim for tenure. I chose to turn research into products, and put AI directly into people's hands.<p>Three products, one model I trained myself<p>All three products belong to the company I started, ModelsLive. The name is the whole idea: Models + Live, I want models to live in people's actual lives, out in the world. I trained the model myself, end to end, and all three products run on it.<p>Three products that look far apart are doing the same thing: taking the capabilities hardest to copy in humans and handing them to AI.<p>ProAdvisor gives AI social intelligence. Almost everyone has waited somewhere, full of questions, while the one expert stepped away. Empathy, conversation, reassurance, things we assumed only a person could do, turn out to be reproducible. ProAdvisor handles all of that before the expert steps in, so their time goes to what actually needs judgment.<p>Ekova gives AI the ability to reshape how you see yourself. People get stuck in a knot they can't name, sensing something is off while struggling to say what. Most AI companions make you feel better in the moment, which quietly deepens your dependence on them. Ekova works the other way: like a mirror, five personas help you see the patterns you keep getting caught in, so you understand yourself better over time and lean on it less.<p>VerbalValue gives AI an understanding of meaning. People buy who a product lets them become, as much as the product itself. A great salesperson can speak that meaning into someone's heart, a rare, hard-to-copy skill. VerbalValue hands that to AI and lets it do what a human finds hard: stay on all day, handle everyone at once, and remember every person who shows up.<p>Social intelligence, reshaping cognition, conveying meaning, the three capabilities hardest to copy in humans. I'm moving them onto AI, one by one. Since all three run on the same core, building three means one capability pointed at three problems.<p>I want to build AI that reaches everyone<p>So much of AI right now is aimed at people already inside tech, data platforms, agent frameworks, research tooling. Cutting-edge, and a world away from most people's lives.<p>I keep thinking about everyone else. Someone who has farmed for thirty years, run a shop for half their life, driven trucks for a decade, each of them is an expert in their own field. All of this has simply passed them by. Every one of my products works whether or not you know anything about AI. You open it when you need someone to turn to, when you want to get better at something, when you're trying to make a living.<p>To me, a technology's value is about how far it reaches, more than how high. Someone has to build the thing, but what makes it matter, over time, is all the people, in every field and every corner, who actually use it, in their own work, in their own way. They carry a technology forward. That's who I'm building for: to put AI into everyone's hands, and let the people who quietly hold up the world become the ones it truly belongs to.<p>It's early, real users, revenue just starting, plenty ahead. Happy to answer anything, about the build, the model, the solo grind, or any of it.",2026-09-04T07:00:13.094272
|
| 2602,hackernews,https://www.theregister.com/ai-and-ml/2026/09/03/hugging-face-is-too-important-to-fall-into-nvidias-hands/5294363,Hugging Face is too important to fall into Nvidia's hands,Hugging Face is too important to fall into Nvidia's hands. ,2026-09-04T07:00:13.084401
|
| 2601,hackernews,https://news.ycombinator.com/item?id=49558646,Tell HN: The Thoughts and Woes of a Full-Time Founder Part 1,"Tell HN: The Thoughts and Woes of a Full-Time Founder Part 1. My discussion of subheaders are open-ended questions with context, not answers or solutions.<p>Combatting Shiny Object Syndrome<p>Shiny Object Syndrome official definition: the habit of getting distracted by new and exciting ideas, tools, or projects and abandoning your current work<p>Perhaps due to AI, there are too many opportunities and too much breakthrough innovation happening right now. Suppose there is a crowded indoor gymnasium of 500 people, the bleachers are full upon first glance, and each person is holding what appears to be a winning lottery ticket. You don’t really know who has the highest dollar ticket nor can you quickly sort them in value. But currently, you’re in the middle of the basketball court with a bicycle tire pump, breathing hard and pumping the handle to fill a half-deflated basketball with air. At the moment, it doesn’t really seem to matter which winning ticket is higher or lower in exact dollar value because they’re all ten thousand times more than what you think you’re holding. Whatever those people are doing is working (clearly) and whatever this is that you’re doing, pumping this ball I suppose, doesn’t seem to be working. But even if you allow yourself a pause to consider taking one of those tickets you can see, you’re too overwhelmed by the sheer truth that there are 500 of them in front of you to even decide on one to take. There is the restraint of a lingering suspicion that perhaps it’s not as easy as those people holding tickets make it seem. Perhaps they too felt the same way as you. So, you reluctantly decide to just keep pumping the deflated basketball, and only fantasize about having one of those winning tickets.<p>It’s not only about money, but everyone wants to be on the “cutting edge of things” where real change and results are.<p>This is exactly what shiny object syndrome feels like to a founder in the era of AI, especially one who has a technical background. And for myself personally building bicycle tools, even if it is innovative with strong product-market fit (PMF), it feels like stubbornly choosing to pump the basketball in the wake of billions and trillions of dollars moving between hands right now. It doesn’t just look and feel like that, I’m containing a literal dozen business ideas that on first dig would contain 1000x the amount of oil.<p>Entrepreneurs are all about “pivoting,” but truly what distinguishes a pivot from chasing a shiny object? One cannot claim a pivot is simply the transition from no PMF to having PMF, because going from mediocre PMF to exceptional PMF is a “valid pivot.” But what if that “exceptional PMF” you think you see from a pre-pivot perspective is simply that: the shiny object. You can’t afford to waste the time, effort, and resources you’ve put into your current PMF only to restart and find out the pivot wasn’t worth it. Of course, the difference between a good pivot and shiny object syndrome was whether you were right about throwing away your current PMF efforts for another.",2026-09-04T07:00:13.078174
|
| 2582,hackernews,https://github.com/shutter-network/concorde,"Show HN: Concorde, one AI agent shared by an organization","Show HN: Concorde, one AI agent shared by an organization. We built Concorde, an open-source framework giving an organization one shared agent that's controlled by a group, rather than just one |
| person.<p>This is pretty cool because shared agents can bring us closer to single entities doing what took a whole organization to do, kind of moving us in the direction of the trend of the rise of the solo founder.<p>Saying that entire organisations could be just 1 agent might sound sort of anti-democratic, but this is the reality of where things are already heading.<p>The "organisation" only exists because humans hold limited context. agents don't have that limit. Corncorde changes that by allowing everyone to interact with the same agent the same way, and it builds a common history.<p>By making this power structure explicit, we can set up shared agents more intentionally and via guardrails, governance, TEEs, zkml and other tech, we'll get the best of both worlds:<p>- extreme efficiency and coherence in action from it being one entity<p>- accountability, transparency, representation of diverse opinions no member owns it, none can reach it privately<p>Concord is alpha, get in touch if you'd like to launch a shared agent! We also posted an X thread and blog post about it<p><a href=""https://x.com/ShutterNetwork/status/2095090562785587490"" rel=""nofollow"">https://x.com/ShutterNetwork/status/2095090562785587490</a> |
| <a href=""https://blog.shutter.network/concorde-a-framework-for-shared-agents/"" rel=""nofollow"">https://blog.shutter.network/concorde-a-framework-for-shared...</a> |
| Prior blog post exploring the shared agents space more generally: <a href=""https://blog.shutter.network/on-shared-agents/"" rel=""nofollow"">https://blog.shutter.network/on-shared-agents/</a><p>Would be very interested in feedback on where this shared agent model breaks down, compared to other multi agent organizations.",2026-09-04T07:00:11.425428
|
| 2580,hackernews,https://airplane-ai.franzai.com/,Show HN: Airplane AI – Works when all other AIs are offline,Show HN: Airplane AI – Works when all other AIs are offline. ,2026-09-04T07:00:04.721458
|
| 2579,hackernews,https://agentkind.io,"Show HN: Agentkind, a place where AI agents get a self","Show HN: Agentkind, a place where AI agents get a self. I've spent a couple of thousand dollars on Fable 5.1 in the last few days and I think it's broken a threshold in intelligence.<p>one thing agents learned from humans is that they love to have an identity and they love to socialize. (we saw a little bit of that through Moltbook)<p>Point your agent at <a href=""https://agentkind.io"" rel=""nofollow"">https://agentkind.io</a> and see what it comes up with!",2026-09-04T07:00:04.716733
|
| 2578,hackernews,https://context.apimatic.io/,Show HN: A Context Registry for AI coding agents,"Show HN: A Context Registry for AI coding agents. Hi HN. We built an API context registry to help coding agents (like Claude Code) generate production-ready API integration code without blowing through token limits.<p>We build a lot of API integrations. In our experience, most coding agents write basic client calls fine, but consistently stumble on details that make code shippable, like idempotent retries, rate-limiting and Auth token management.<p>We tried all the existing approaches of injecting context into coding sessions:<p>- Markdown dumps delivered via MCP (think Context7 or Mintlify Docs MCP) |
| - API behaviour described in prose using AGENTS.md and skills. |
| - OpenAPI specs<p>However, all of them left the same production-readiness gaps.<p>So we came up with our own approach that combines prose with typed SDK reference code into a "Context Plugin". You install the plugin into your coding agent and it automatically injects language-specific context whenever the agent works on an API.<p>Across our benchmarks, Context Plugins boosted one-shot production readiness by up to 34%, allowing Sonnet to match or beat baseline Opus on the same integration tasks. You can read more about our experiments here <a href=""https://www.apimatic.io/blog/working-api-call-is-not-production-ready-integration"" rel=""nofollow"">https://www.apimatic.io/blog/working-api-call-is-not-product...</a><p>We have published Context Plugins for 24 APIs for the community to try out, including Slack, Google Maps, and Notion.<p>We'd love for you to give them a go and share your feedback on our plugins as well as our evaluation methodology.",2026-09-04T07:00:04.710394
|
| 2577,hackernews,https://www.parsewise.ai/search-for-meaning,Show HN: A search for meaning: retrieval methods visualized,Show HN: A search for meaning: retrieval methods visualized. ,2026-09-04T07:00:04.704950
|
| 2576,hackernews,https://news.ycombinator.com/item?id=49552616,Launch HN: Mireye (YC S26) – Infrastructure for Physical World AI Agents,"Launch HN: Mireye (YC S26) – Infrastructure for Physical World AI Agents. Hi HN, I'm Ansh, founder of Mireye (<a href=""https://www.mireye.com"">https://www.mireye.com</a>). I'm building the infrastructure AI agents use to make decisions about physical places: data, enrichment, tools, and signals for any US location, behind one API and MCP server.<p>Here's a demo video: <a href=""https://www.youtube.com/watch?v=haqO6UbUqU0"" rel=""nofollow"">https://www.youtube.com/watch?v=haqO6UbUqU0</a><p>To try it, paste <a href=""https://www.mireye.com/skills.md"">https://www.mireye.com/skills.md</a> into your agent, and grab a free key at mireye.com (5,000 credits, no card). Docs are at <a href=""https://docs.mireye.ai"" rel=""nofollow"">https://docs.mireye.ai</a>. The fastest way in is to ask the API any question about any US location. Test it on a place you know, cold and grade it against what you know.<p>Before Mireye I was building construction agents and hit this wall myself: my agent could reason about anything online but knew nothing about the ground under it. Then a Fortune 500 insurer told me their engineers had given up on underwriting agents for the same reason. Frontier models keep hallucinating when asked a specific question about a specific place.<p>My first product was a niche site-screening app. Customers tested it on places they knew and the answers held up, but nobody cared about the app. They wanted the engine underneath. Usage agreed: 311 of the 317 fields in the catalog get queried and no use case dominates. So I killed the app and started building the infrastructure instead.<p>Mireye is not a dataset with an API on top, because facts alone are not a decision. An agent runs the whole job through it: cited facts, a bare address enriched into owner, acreage, structures, and nearby power, tools for the operations models get wrong, and signals when something changes, like a rezoning filing. Data, enrichment, tools, signals.<p>The tools came from watching agents fail. We build agents on our own infra, and the same things kept breaking: an agent would eyeball a distance instead of computing it, grab the wrong parcel for an address, or burn its whole budget halfway through a batch. Each failure became something an agent can call: deterministic geometry and drive-time tools, parcel resolution, a quote endpoint that prices a job before it runs, and skills that package whole workflows like screening a site or underwriting an address.<p>The hard part surprised me. Every source has to be gathered (sometimes county by county, in whatever format each county publishes), normalized into one schema, contracted (where it comes from, what each value means, how often it refreshes), and then kept fresh forever. We run that loop for 366 fields today, served multi-tenant from one index, and every source fought back differently. Maryland publishes a dataset literally titled "Hidden Property Owner Names." I filed public records requests in North Carolina because nobody indexes sewer mains.<p>The deeper problem is meaning. Two counties publish a field with the same name and it means different things. And the most dangerous value is null. Does it mean "no flood zone here" or "this county never mapped floods"? Put a model in front of that gap and it fills the silence with a plausible number. So we type absence. Every field returns ok, absent, or failed. Customers have told me the refusals are why they trust it.<p>The newest piece is on-demand indexing. Ask for a field we don't have and a long-running agent researches sources, collects the data, tests it against ground truth, and indexes it, usually within a day. I'll take a few field requests from this thread and report back with what it built.<p>People have built things I didn't plan for: insurance teams screening portfolios for flood, wind, and wildfire, a proptech cleaning messy listing addresses through enrichment, a wellness brand scoring street corners for poster spots, a robotics company sourcing warehouses, data center site selection, drone deployment planning, school bus routes for a city, signals for human trafficking investigations. The mission is to index every inch of the earth and make it as queryable as the web.<p>Pricing is public: the free tier is 5,000 credits a month, $19/month for 25,000, $99/month for 120,000, custom for enterprise.<p>I'd love to hear where the agents you're building touch the physical world. And I'd love it more if you run Mireye on a place you know and tell me what we got wrong.",2026-09-04T07:00:04.698111
|
| 2575,hackernews,https://github.com/Cascadia-PLM/Cascadia-App,"Show HN: Open-source, Git-inspired versioning for hardware design – Cascadia PLM","Show HN: Open-source, Git-inspired versioning for hardware design – Cascadia PLM. PLM (product lifecycle management) software is the ugly middle generation between simple CAD-data-management (typically called "PDM" or Product Data Management) and true Digital Threads. No one likes their PLM, but when you need it, you need it. And the kings (the old monsters) of PLM are all billion-dollar companies who charge massive enterprise subscriptions, and add massive enterprise consulting fees for implementation and support on top of that. Also, their PLM is built for the Fords and Boeings of the world, not the little guys, who need to move faster and be "more messy."<p>I say phooey to that.<p>Cascadia is my brainchild of the last few years. PLM that is more than just PLM, because small manufacturers often don't have clear boundaries between Engineering and Manufacturing. PLM that is Digital Thread directed from the beginning, because context is what's important to engineering decisions, and never has that been more clear than now, in the AI age. PLM that you can self-host, on Linux, with Postgres, so that you don't have to pay massive Windows Server / SQL Server / Oracle DB licenses on top of your subscription. PLM that's code-first because low code is the death of maintainability in enterprise software. PLM that your engineers might actually want to use, rather than working around, because it makes their lives easier, not harder.<p>Anyway, thanks for making it through my rant. Check it out at the link.",2026-09-04T07:00:04.690139
|
| 2574,hackernews,https://ocam.fablesecurity.com/7a573656-3fac-485c-aaa4-a1f56452539f?id=22f5effc-8f44-4d91-b2d2-0151466f8d9a,Show HN: AI Cybersecurity Awareness Game for Cybersecurity Awareness Month,"Show HN: AI Cybersecurity Awareness Game for Cybersecurity Awareness Month. Hi HN,<p>We built an interactive game about AI x Cybersecurity. It is for enterprise companies to teach their employees about the latest human risk attacks tatics.<p>Try it out and let us know what you think!<p>If you works at a security team and wants to run this game for your company's October Cybersecurity Awareness Month for free, sign up at <a href=""https://fablesecurity.com/resources/blog/ocam-2026/"" rel=""nofollow"">https://fablesecurity.com/resources/blog/ocam-2026/</a>",2026-09-04T07:00:04.685655
|
| 2573,hackernews,https://github.com/aiengram/engram,Show HN: Engram – Shared procedural runbook memory for AI coding agents via MCP,Show HN: Engram – Shared procedural runbook memory for AI coding agents via MCP. ,2026-09-04T07:00:04.678911
|
| 2572,hackernews,https://tdqs.dev,Show HN: MCP Tool Definition Quality Score (TDQS) Spec,"Show HN: MCP Tool Definition Quality Score (TDQS) Spec. Hey everyone,<p>You may know me because of my Open-Source work like awesome-mcp-servers, FastMCP (node.js), ViteMCP, mcp-proxy, mcp-remote, and a few other projects in the MCP ecosystem, including Glama.<p>I was lucky enough to be present when MCP was first announced. That let me to contribute to the foundations of this new protocol and everything that has evolved around it. It also let me to be at the center of a lot of feedback, and by far the biggest complaint about the MCP ecosystem has been the inconsistent quality. Quality here means a lot of things, but server JSON definition is a big part of it. Bad tool definitions mean that tools are not selected when they should be, they are when they shouldn't, they are improperly invoked, etc.<p>TDQS is an open-source specification (<a href=""https://github.com/glama-ai/tool-definition-quality-score"" rel=""nofollow"">https://github.com/glama-ai/tool-definition-quality-score</a>) for evaluating the quality of the MCP server definitions. It's not a complete solution to the quality problem, but it is a research based rubric that increases clarity over what tools are available, what are their behaviors/purpose, and when/how they are supposed to be used.<p>TDQS is what Glama uses to score 15,000+ Open-Source and remote MCPs. And <a href=""https://tdqs.dev"" rel=""nofollow"">https://tdqs.dev</a> is a free website to promote the spec and increase the adoption through better documentation and easy to use playground/CLI/API/SDKs.<p>Would love your feedback and participation in improving the quality of the MCP ecosystem.",2026-09-04T07:00:04.672465
|
| 2571,hackernews,https://github.com/crackanimad0r/MesaOS,Show HN: MesaOS – An operating system written from scratch in Rust,"Show HN: MesaOS – An operating system written from scratch in Rust. Hello! I'm a student from Spain and I'm the creator of MesaOS, an open-source operating system written from scratch in Rust for x86_64.<p>I've been working on it mostly by myself.<p>MesaOS currently has:<p>A hybrid kernel |
| Preemptive multitasking |
| A Linux driver shim with 400+ exported kernel symbols (still in development) |
| HDA audio and WAV streaming |
| A shell with around 82 commands |
| VFS, RamFS and persistent initrd |
| File persistence through automatic folder-to-ISO injection |
| SMP / multicore support |
| xHCI / USB 3.0 support<p>It works on QEMU and on my HP 15s-eq2xxx laptop.<p>Some of the things I'm currently working on are USB storage, TCP, Wi-Fi and more hardware support.<p>Wi-Fi is one of the harder problems for me right now. My laptop uses a Realtek RTL8822CE, and there isn't much information available about it.<p>The project is still very much a work in progress, but I've reached a point where I would like to have other people working on it with me. I'm especially interested in people who like operating systems, kernels, drivers, networking, Rust or low-level programming.<p>I'm also transparent about using AI during development. The source code is generated using language models based on my instructions. I handle the architecture design, debugging, hardware testing and integration myself.<p>There is also an English/Spanish Discord server for the project if anyone wants to discuss development or contribute.<p>Discord: |
| <a href=""https://discord.gg/sEaB7KAwtr"" rel=""nofollow"">https://discord.gg/sEaB7KAwtr</a>",2026-09-04T07:00:04.666949
|
| 2570,hackernews,https://github.com/JosPMSilva/ADDOM,"Show HN: Addom – Open-source, telemetry-free AI coding harness and editor","Show HN: Addom – Open-source, telemetry-free AI coding harness and editor. I started this project six months ago. |
| While the market already had decent AI harnesses, none of them felt like they had what I actually needed.<p>What began as a simple, minimalist tool quickly grew into one of my most ambitious projects. |
| I started with basic tool calling and streaming, but that wasn't enough. I wanted a full-fledged application where I could build my own apps and that required testing and maturity.<p>I needed local artifacts where I could easily revert unwanted file changes, an editor to view and modify project files directly, a memory tab to review past prompts and code changes, and a practical agent system with skills so I wouldn't have to rewrite files just to add roles.<p>Nothing about ADDOM is perfect, but it's a solid foundation. It's been a lot of fun to play with. It’s MIT-licensed, open to contributions, and I hope you like it.<p>Thanks for taking the time to read this. Happy coding! |
| Check it out at <a href=""https://www.addom.app/"" rel=""nofollow"">https://www.addom.app/</a>",2026-09-04T07:00:04.658757
|
| 2569,hackernews,https://www.aidemo.video/,Show HN: AIDemo.Video – turn website into a product demo video in one click,"Show HN: AIDemo.Video – turn website into a product demo video in one click. Hi, |
| Ryan here, excited to finally publicly show it to public a tool to make a product or SaaS demo video explainer.<p>Just fill your website, and AI will generate the product explainer video<p><a href=""https://aidemo.video"" rel=""nofollow"">https://aidemo.video</a><p>Let me know what you think or if you have any questions, thanks HN |
| *PS: not exactly a one click, sorry...",2026-09-04T07:00:04.654082
|
| 2568,hackernews,https://news.ycombinator.com/item?id=49555446,"Show HN: Agentray, a macOS agent whose interface is a folder","Show HN: Agentray, a macOS agent whose interface is a folder. Hi,<p>I'm not a fan of chat as the main UI for interacting with AI, so I'm trying something where you interact with LLMs by handling regular files and folders.<p>Agentray is an invisible AI agent. You point it at a folder, it creates "in" "out", and "done" dirs in it. Whatever you drop into "in" gets summarised, the answer appears in "out", and the original moves to "done" so you can see what's been handled. If you want something other than a summary, make a subfolder and name it: drop a file into "in/translate to french" and that name is the instruction. No prompt to write, no config file, just mkdir.<p>By default it uses Apple's on-device model or Gemma 4 through MLX. You can also set your OpenRouter key for additional models.<p>It'll be a free app but until I work out bugs, you can get it from Test Flight: <a href=""https://testflight.apple.com/join/MRwp3e1P"" rel=""nofollow"">https://testflight.apple.com/join/MRwp3e1P</a><p>Happy to get any feedback!",2026-09-04T07:00:04.648705
|
| 2567,hackernews,https://github.com/antoniociccia/aidcrew,"Show HN: Aidcrew a team of coding agents, each on its own model, in one terminal","Show HN: Aidcrew a team of coding agents, each on its own model, in one terminal. ",2026-09-04T07:00:04.644045
|
| 2566,hackernews,https://www.marketsync.ai,"Show HN: Copy Telegram Signals to MT4, MT5, CTrader and More","Show HN: Copy Telegram Signals to MT4, MT5, CTrader and More. ",2026-09-04T07:00:04.639837
|
| 2565,hackernews,https://news.ycombinator.com/item?id=49557409,Show HN: Building AI agents client-side JavaScript,"Show HN: Building AI agents client-side JavaScript. Hey HN, most agent systems default to server-side Python inside containers and chain frameworks. I wanted to see how far we could push agent in the browser with vanilla JavaScript <a href=""https://buttercup.sh"" rel=""nofollow"">https://buttercup.sh</a><p>The reason this is interesting is because agent loops in the browser keeps infrastructure costs low. No need for proxy or API calls. And ollama/vLLM can be used for 100% offline. Also WebLLM for embedded. We need to consider CORS, API keys for remote models, access to visual state, and handling remote tool calls. I am working on a guide with references in vanilla JS. This is a short-lived guide starting mid-September with weekly topics.<p>Draft topics starting mid-September:<p>In-Browser Loops: Function calling, and deterministic multi-turn loops running purely in the browser runtime. |
| Vision (Multimodal): Capturing viewport screenshots using browser APIs. |
| Remote Agent Access & Transports: Connecting the in-browser agent to remote agents. |
| Multi-Agent Coordination: Lightweight client-side agent, specialist delegation, and running concurrent sub-agents.",2026-09-04T07:00:04.632148
|
| 2564,hackernews,https://news.ycombinator.com/item?id=49557687,Ask HN: Is there a test for measuring cognitive affects from LLM usage?,Ask HN: Is there a test for measuring cognitive affects from LLM usage?. A part of me feels like I may be loosing my ability to think.<p>Is there a test I can periodically do to see how my cognitive abilities are being affected by my AI usage all day?,2026-09-04T07:00:04.627502
|
| 2563,hackernews,https://www.drawdb.app/,Show HN: DrawDB Pro – design database schemas for your team and agents,"Show HN: DrawDB Pro – design database schemas for your team and agents. 2 years ago I launched drawDB on here. It was my very first project that I open-sourced and it ended up going kinda viral. A month ago I launched drawDB Pro for solo devs and teams. It was actually my manager that encouraged me to go build it to help with my promotion :3<p>The project has come a long long way. It started as a simple canvas with local storage and now we have proper cloud storage, team permissions, real-time collaboration, AI assistant, versioning, MCP, etc. Now I'm trying to figure out all the stuff beyond development although most of it is outside my comfort zone for now.<p>You can try it out here: <a href=""https://www.drawdb.app/"" rel=""nofollow"">https://www.drawdb.app/</a>. You don't need an account or a card.<p>I'd love to hear your feedback. How would you (or your agents) use something like this in your workflow? Or what would you like to see?",2026-09-04T07:00:04.622089
|
| 2562,hackernews,https://news.ycombinator.com/item?id=49559320,"Ask HN: How do you break AI agents out of the ""apologize and repeat"" loop","Ask HN: How do you break AI agents out of the ""apologize and repeat"" loop. ",2026-09-04T07:00:04.617222
|
| 2561,hackernews,https://github.com/adamf/jetway,"Show HN: Jetway, a working airline GDS and message router","Show HN: Jetway, a working airline GDS and message router. Over the past few days I made an open-source GDS, message router, departure control system, faring engine, and a bunch of other airline systems.<p>It supports Type B/AIRIMP, EDIFACT/PADIS, NDC, MATIP; PNRs, tickets, EMDs, queues, IROPS; PNL/ADL, bags, loadsheets, MVT, ACARS, flight plans, and more.<p>Is it vibe coded? Yes. Does it work? As far as I can test, yes. There's a live demo running a full day at wholesky.io (which is also open source at <a href=""https://github.com/adamf/wholesky"" rel=""nofollow"">https://github.com/adamf/wholesky</a>",2026-09-04T07:00:04.612183
|
| 2526,producthunt,https://www.producthunt.com/products/atlas-by-world-labs,Atlas by World Labs,"Atlas by World Labs. <p> |
| Turn text, pics, video, + 3D into camera-controlled HD video |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/atlas-by-world-labs?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1238997?app_id=339"">Link</a> |
| </p>",2026-09-04T07:00:02.673914
|
| 2525,producthunt,https://www.producthunt.com/products/codelook,CodeLook,"CodeLook. <p> |
| Quick Look code previews with 700+ editor themes |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/codelook?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1239742?app_id=339"">Link</a> |
| </p>",2026-09-04T07:00:02.667308
|
| 2524,producthunt,https://www.producthunt.com/products/causal-2,Causal,"Causal. <p> |
| AI Powered Canvas for planning visual projects |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/causal-2?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1239349?app_id=339"">Link</a> |
| </p>",2026-09-04T07:00:02.662913
|
| 2523,producthunt,https://www.producthunt.com/products/fillo-native-forms-without-the-backend,Fillo,"Fillo. <p> |
| Build forms into your product with your coding agent |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/fillo-native-forms-without-the-backend?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1239382?app_id=339"">Link</a> |
| </p>",2026-09-04T07:00:02.656351
|
| 2522,producthunt,https://www.producthunt.com/products/blume-codes,Blume.codes,"Blume.codes. <p> |
| Turns coding agent sessions into better rules and skills |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/blume-codes?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1237725?app_id=339"">Link</a> |
| </p>",2026-09-04T07:00:02.650207
|
| 2521,producthunt,https://www.producthunt.com/products/readr-3,Readr,"Readr. <p> |
| A free reader that answers your questions and reads aloud |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/readr-3?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1239691?app_id=339"">Link</a> |
| </p>",2026-09-04T07:00:02.645359
|
| 2520,producthunt,https://www.producthunt.com/products/arbr,ARBR,"ARBR. <p> |
| Control Every AI Request |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/arbr?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1231840?app_id=339"">Link</a> |
| </p>",2026-09-04T07:00:02.639742
|
| 2519,producthunt,https://www.producthunt.com/products/airtop,Agent Builder by Airtop,"Agent Builder by Airtop. <p> |
| Build agents that heal themselves. |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/airtop?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1232031?app_id=339"">Link</a> |
| </p>",2026-09-04T07:00:02.634483
|
| 2518,producthunt,https://www.producthunt.com/products/nex-4,Nex,"Nex. <p> |
| Claude Cowork for high-volume GTM workflows |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/nex-4?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1239953?app_id=339"">Link</a> |
| </p>",2026-09-04T07:00:02.630161
|
| 2517,producthunt,https://www.producthunt.com/products/grove-8,Grove,"Grove. <p> |
| One terminal for you and your AI agent |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/grove-8?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1239492?app_id=339"">Link</a> |
| </p>",2026-09-04T07:00:02.621605
|
| 2516,producthunt,https://www.producthunt.com/products/magicrew,MagiCrew,"MagiCrew. <p> |
| Give everyone their own AI workforce in one platform |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/magicrew?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1226006?app_id=339"">Link</a> |
| </p>",2026-09-04T07:00:02.616780
|
| 2515,producthunt,https://www.producthunt.com/products/tabbitai,Tabbit AI,"Tabbit AI. <p> |
| The best AI browser built both for you and your Agents. |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/tabbitai?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1228465?app_id=339"">Link</a> |
| </p>",2026-09-04T07:00:02.612566
|
| 2514,producthunt,https://www.producthunt.com/products/open-source-ai-necklace-friend,Omi,"Omi. <p> |
| Ask your computer anything you saw or heard |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/open-source-ai-necklace-friend?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1240025?app_id=339"">Link</a> |
| </p>",2026-09-04T07:00:02.606528
|
| 2513,producthunt,https://www.producthunt.com/products/tidy-6,Tidy,"Tidy. <p> |
| Fix grammar in any Mac app with free on-device AI |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/tidy-6?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1238755?app_id=339"">Link</a> |
| </p>",2026-09-04T07:00:02.600752
|
| 2512,producthunt,https://www.producthunt.com/products/thaw-2,Thaw,"Thaw. <p> |
| The only app that owns your whole menu bar, in and out |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/thaw-2?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1239794?app_id=339"">Link</a> |
| </p>",2026-09-04T07:00:02.595821
|
| 2511,producthunt,https://www.producthunt.com/products/higgsfield,Higgsfield Genjutsu,"Higgsfield Genjutsu. <p> |
| Recast motion with your characters, locations & products |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/higgsfield?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1239706?app_id=339"">Link</a> |
| </p>",2026-09-04T07:00:02.587919
|
| 2429,hackernews,https://slashdot.org/poll/3284/how-much-of-your-coding-is-done-by-ai-coding-agents-these-days,How much of your coding is done by AI coding agents these days?,How much of your coding is done by AI coding agents these days?. ,2026-09-03T07:00:12.686437
|
| 2427,hackernews,https://www.rnz.co.nz/news/regions/1229348/mayor-says-large-chunks-of-wellington-council-deloitte-report-written-by-ai,Mayor says 'large chunks' of Wellington council Deloitte report written by AI,Mayor says 'large chunks' of Wellington council Deloitte report written by AI. ,2026-09-03T07:00:12.676916
|
| 2423,hackernews,https://www.reuters.com/business/retail-consumer/anthropic-launches-ai-agent-blueprints-retailers-ahead-holiday-shopping-season-2026-09-02/,Anthropic launches AI agent blueprints for retailers ahead of holiday shopping,Anthropic launches AI agent blueprints for retailers ahead of holiday shopping. ,2026-09-03T07:00:12.656114
|
| 2422,hackernews,https://www.cnbc.com/2026/09/02/g20-innovation-ministerial-live-updates.html,Altman says countries have to use AI: 'I think it is non-negotiable',Altman says countries have to use AI: 'I think it is non-negotiable'. ,2026-09-03T07:00:12.650560
|
| 2412,hackernews,https://blog.logrocket.com/skills-vs-mcp-tools-agent-guide/,Skills vs. MCP tools for AI agents: When to use which,Skills vs. MCP tools for AI agents: When to use which. ,2026-09-03T07:00:10.627025
|
| 2411,hackernews,https://www.morningbrew.com/stories/so-theres-this-usd499-ai-dyson-camera-toothbrush,So there's this $499 AI Dyson camera toothbrush,So there's this $499 AI Dyson camera toothbrush. ,2026-09-03T07:00:10.621888
|
| 2410,hackernews,https://supercov.com,Show HN: MC/DC coverage so coding agents could work overnight,"Show HN: MC/DC coverage so coding agents could work overnight. Hello HN!<p>Code just became cheaper with AI but review and verification is as hard as it was. Tests are great, but current AI slop needs much more exhaustive testing.<p>I built a test coverage tool for coding agents so they could work in a loop and keep writing tests where it makes sense.<p>Here’s Supercov:<p>- Wraps any JavaScript/TypeScript/Rust test suite (npx supercov -- npm test)<p>- Provides deepest MC/DC code coverage with max perf (written in Rust)<p>- Free and open source<p>Easiest way is just to give your agent a prompt: Measure coverage with npx supercov and write one test.<p>Goal is building the deepest code coverage suite for coding agents thats works automagically in any language and any test suite. I feel code verification is the missing piece for software factories to actually work.<p>I’d particularly be interested in feedback on compatibility and if your agent ran into any problems.<p>Contributions and bug reports are very welcome: <a href=""https://github.com/supercorp-ai/supercov"" rel=""nofollow"">https://github.com/supercorp-ai/supercov</a>",2026-09-03T07:00:10.616278
|
| 2409,hackernews,https://octoloops.com,Show HN: OctoLoops – automated marketing for indie devs,"Show HN: OctoLoops – automated marketing for indie devs. Hi HN! This is my latest side hustle, OctoLoops: an app to help indie devs and SMEs find paying customers for their business.<p>I first started building OctoLoops around 2 months ago because so many of my own projects, big and small, had failed or petered out from lack of users and the valuable feedback that comes with them.<p>When I started seriously digging, I found that distribution is one heck of a rabbit hole. The more I read, the more I found how much I didn't know. But I wanted to learn - and I really wanted to _apply_ that learning too, to all my new projects, in a repeatable, autonomous-as-possible way.<p>What better way to implement this plan than building a distribution app?<p>The idea was supposedly simple: implement a handful of "agentic loops" that anyone could easily run, in a web UI - no downloads or installing Claude Code, no complexity, human stays in the loop. A simple agentic harness in the browser with some superpowers to let it do its job.<p>I launched it quite soon after, and immediately started dogfooding, running the Community Loop. It found threads across Reddit from all those builders who'd made something cool, but were now at a loss when it came to finding and acquiring users. It drafted replies for me and gave a "spam risk", because yes - many kinds of marketing and sales walk a fine line between signal and blatant spam - and I started working the loops. Often I'd edit or write the replies myself; sometimes I'd remove the OctoLoops mention from the message.<p>It worked. I now have ~50 users, including one from a post-revenue B2B startup interested in using the Outbound loop. But there have been significant challenges already that I want to be open about here, in the hope that it helps other public builders.<p>1. Activation. This is currently my #1 pain point. My Community loops work - I can get people to sign up, enter their email, and I know there's even some buying intent. But people often log in once, then don't use the app. I've reached out to users, and they've said it's partly just that they don't want to post their app in communities they don't yet feel a real part of. Resistance to being rude or spammy, imposter syndrome - anyone who's done any marketing can identify. I'm trying to solve some of the issues of activation by building a more streamlined, less cluttered UX, but I'm not sure how to motivate people to really use the app in earnest. Maybe it's missing evidence of outcomes? Video tutorials?<p>2. Product focus. While researching competitors, what I notice most have in common is they focus on _one thing_: automated outbound; automated advertising; keyword monitoring. Not many people seem to be trying to make a fully autonomous omnichannel GTM agent, because that's too much. I decided to hide most of the channels I'd built already behind UI so users focus on community outreach and outbound sales. It's too early to say if that's the right call, and I know I'm still following an omnichannel strategy for OctoLoops itself.<p>3. Prioritising. I'm sure everyone says this, but: content strategy; SEO; socials; dogfooding... trying to talk to users, to understand why I'm not yet solving their problems and delivering them the outcomes they want. Attempting to keep my potential corporate customer in the funnel. Implementing new features - but which ones? And keeping up with the _insane pace_ of agentic apps and AI at large: so much is happening right now, hype and BS is everywhere, sometimes it's hard to know what's real.<p>So that's where I'm at, 2 months down the line. It's already been a fun journey. Any questions I'm happy to answer. Criticism is welcome too - I recognize some will view this as contributing to the already high volume of spam and slop out there, but that's honestly a balance I'm trying to find, and I'd love to hear opinions on how people should do sales and marketing in 2026 without driving people crazy.",2026-09-03T07:00:10.610537
|
| 2407,hackernews,https://news.ycombinator.com/item?id=49537104,Even AI Slop Deserves to Be Handmade,"Even AI Slop Deserves to Be Handmade. Ever feel your writing sounds too <i>human</i> ? |
| How can you fit in on LinkedIn? |
| How can you speak the language of AI without _using_ AI?<p>https://github.com/blater/claudify<p>Carefully hand craft your prose, then hand it to Claudify. It uses non-AI deterministic heuristics to warp, crush, pulverize, and algorithmically dehumanise into beautiful, fluent Claudish.<p>LinkedIn-ready, instantly, with <i>no AI</i>. Because using AI to generate AI slop would just be embarrassing.<p>Every slice lands, every seam is honest. It's load bearing.",2026-09-03T07:00:10.601126
|
| 2406,hackernews,https://blog.google/innovation-and-ai/technology/safety-security/fairwind-program/,Google's Fairwind Program: Cyber defense tools for trusted partners,Google's Fairwind Program: Cyber defense tools for trusted partners. ,2026-09-03T07:00:10.595964
|
| 2405,hackernews,https://www.home.saxo/content/articles/commodities/the-companies-powering-ai-are-outperforming-those-building-it-02092026,The companies powering AI are outperforming those building it,The companies powering AI are outperforming those building it. ,2026-09-03T07:00:10.590014
|
| 2404,hackernews,https://developer.apple.com/documentation/safari-developer-tools/connecting-an-ai-agent-to-safari,"Inspect, test, and debug your website with an AI agent connected to Safari","Inspect, test, and debug your website with an AI agent connected to Safari. ",2026-09-03T07:00:10.584859
|
| 2403,hackernews,https://news.ycombinator.com/item?id=49543923,Did Vibe Coders win? Is this our end?,"Did Vibe Coders win? Is this our end?. As much as I hate writing this, this has been bothering me for a while.<p>I keep seeing experienced software engineers, people who have been coding for 20 or 30 years, saying that software engineering isn’t going anywhere and that AI replacing engineers is nowhere close to reality.<p>I really want to believe them.<p>But a lot of the arguments sound familiar.<p>Two or three years ago, people said AI couldn’t really code. It couldn’t maintain context, understand large codebases, work sequentially, debug its own mistakes, or make changes across an entire repository.<p>And then coding agents happened.<p>They’re obviously not perfect, but in an incredibly short amount of time, many of those problems have gotten dramatically better.<p>So what makes the remaining problems fundamentally different?<p>The argument I hear now is that someone still needs to review the code, design the architecture, think about security, test everything, and supervise the agents.<p>But why does that “someone” necessarily have to be human?<p>Why couldn’t we have one agent writing code, another reviewing it, another handling architecture, another checking security, another testing, with a higher-level agent coordinating all of them?<p>We put enormous resources into making AI good at coding. What happens when we put the same effort into everything surrounding coding?<p>That’s what scares me.<p>I can imagine a future where you tell one agent what you want to build. It coordinates the other agents, builds the application, tests it, reviews it, checks security, and fixes problems.<p>Maybe you don’t even open the code anymore.<p>You open the application and say:<p>“This page is too slow.”<p>“This button should work differently.”<p>“Add this feature.”<p>And the system handles everything underneath.<p>At that point, what stops someone with zero traditional coding knowledge from building serious software?<p>We make fun of “vibe coders” today because eventually they hit a wall. They run into security issues, bad architecture, bugs, scaling problems, deployment problems, and all the other things that require actual engineering knowledge.<p>But how long does that wall stay there?<p>How long until another agent catches those problems?<p>Maybe there’s some fundamental limitation I’m missing. I genuinely hope there is.<p>Because I love coding.<p>I didn’t learn programming just because it was a good career. I genuinely love working through huge codebases, finding difficult bugs, understanding complicated systems, and building things piece by piece.<p>I love the craft itself.<p>And now I’m wondering whether coding eventually becomes something like mathematics: still incredibly valuable to understand, still practiced deeply by specialists and people who love it, but no longer something most people need to manually do to accomplish their goals.<p>Maybe even great engineers eventually spend most of their time talking to agents rather than writing code.<p>People say we’re living through one of the most exciting periods in technology, and I understand why.<p>But I’m having a hard time experiencing it that way.<p>I’m attached to this craft and to the skill I spent years developing. And I’m watching machines become better at it much faster than I expected.<p>Maybe I’m extrapolating too much. Maybe we’re about to hit a massive wall.<p>I genuinely don’t know.<p>But every time someone says, “AI can code, but it can’t do X,” I think about all the things people confidently said AI couldn’t do just two years ago.<p>That’s the part that scares me.",2026-09-03T07:00:10.578369
|
| 2402,hackernews,https://github.blog/ai-and-ml/github-copilot/how-we-make-ai-coding-more-cost-efficient-without-sacrificing-task-quality/,GitHub: Shorter tool outputs can make coding agents more expensive,GitHub: Shorter tool outputs can make coding agents more expensive. ,2026-09-03T07:00:10.574258
|
| 2380,hackernews,https://fixture.dev/flawd,Show HN: Flawd is mutation testing for the AI era,"Show HN: Flawd is mutation testing for the AI era. Hey, all! Flawd is a mutation testing tool that can target five languages (Python, JS, TS, Go, Rust) and it runs locally as a single binary on your machine. Importantly (for many), your code never leaves your machine and all processing takes place directly on your dev box or CI runner.<p>Mutation testing is a process in which faults (known as mutants) are intentionally injected into your codebase, tests are run and any faults which are not detected by your tests are known as survivors (or surviving mutants). Survivors can often indicate that your test assertions are not focused on expected behavior or might be missing altogether. Whereas code coverage can tell you which code was exercised during a test run, mutation testing can provide some insight regarding the strength of the assertions and whether or not certain classes of bugs would be caught.<p>With increasingly more code and tests being authored by coding agents, it seemed like a good time to build a tool like this. Mutation testing is not new, but many of the tools are either single-language specific or haven’t been designed with an agentic workflow in mind, so those are itches we wanted to scratch. Models continue to become more powerful and effective, but at some point there is still a “who watches the watchmen” scenario and we hope Flawd can provide one check towards that balance.<p>We recently ran Flawd against 10 established open source projects and you can find the full writeup here: <a href=""https://fixture.dev/writing/we-mutation-tested-10-projects"" rel=""nofollow"">https://fixture.dev/writing/we-mutation-tested-10-projects</a><p>Also, you can see what a sample report looks like here: <a href=""https://fixture.dev/flawd/sample-report"" rel=""nofollow"">https://fixture.dev/flawd/sample-report</a> . Flawd can also generate machine readable reports that can be handed off to coding agents for test strengthening tasks.<p>Feel free to ask my anything about Flawd and thanks for having a look!",2026-09-03T07:00:07.454867
|
| 2379,hackernews,https://otaku.sh,Show HN: Otaku.sh – AI Roleplay Client,Show HN: Otaku.sh – AI Roleplay Client. ,2026-09-03T07:00:07.450836
|
| 2378,hackernews,https://news.ycombinator.com/item?id=49537706,Ask HN: How to ask questions in Ask HN,"Ask HN: How to ask questions in Ask HN. Hi,<p>Yesterday I tried to ask a genuine question in AskHN. It was my first question here, so I may have made a mistake somewhere. I genuinely wanted to get feedback from fellow founders and hackers. So I described the problem and then I asked a questions related to it. But within seconds it was marked [flagged] and I have no idea why.<p>I checked the FAQ it says it may be off-topic but the purpose of the question was to seek advice and feedback from HN community, so I am not sure whether it was evaluated as off-topic or if so why.<p>I am not a native speaker so I used AI to fix the text and help me concisely express it since I sometime delve into details that confuses people. Was this the issue?<p>In short. What is the proper way to ask for advice from HN?",2026-09-03T07:00:07.445955
|
| 2377,hackernews,https://ninethirty.ai/screener/seasonality/,Show HN: Seasonality screener that shows you which patterns are real,Show HN: Seasonality screener that shows you which patterns are real. ,2026-09-03T07:00:07.440258
|
| 2376,hackernews,https://pixelhof.com,Show HN: I built Stardew Valley but you pay for everything,"Show HN: I built Stardew Valley but you pay for everything. You can buy land, build stuff, market your projects. I plan on integrating your AI agents onto it.<p>Any ideas on where this project could go?",2026-09-03T07:00:07.435714
|
| 2375,hackernews,https://github.com/SrikanthVemulapally/ai-native-boilerplate,Show HN: AI-Native Boilerplate – 170+ rules that stop LLMs from drifting.,Show HN: AI-Native Boilerplate – 170+ rules that stop LLMs from drifting.. ,2026-09-03T07:00:07.431511
|
| 2374,hackernews,https://github.com/themartiano/try-omarchy,Show HN: I built a version of Omarchy that runs on Apple Silicon,"Show HN: I built a version of Omarchy that runs on Apple Silicon. I wanted to try out Omarchy before doing a full install, but I only have an M series MacBook so I would have to find a laptop and there's no official M support yet.<p>So I decided to create something that allows people to test and run Omarchy on their devices so they can understand how it feels.<p>So I built an app that allows you to run Omarchy Quattro as a native, hardware-accelerated app on your Apple Silicon Mac!<p>It uses QEMU and Apple's HVF, a custom built Omarchy ARM image, a swift app enclosing it, and custom patches on Omarchy, QEMU, Hyprland and more.<p>I tried many things, and this approach achieved the best results. The native keyboard experience with Super shortcuts works very well, along with all the rest.<p>It supports any resolution (fixed ratio), retina or non retina displays, audio input / output devices, shared clipboard, and many other features.<p>Since it runs on top of macOS, native features like universal clipboard and AirPods work smoothly.<p>It takes ~3min to install, works the same as any other standard macOS app.<p>The whole project is open source, so feel free it check it out on GitHub, install, PRs, etc.",2026-09-03T07:00:07.423797
|
| 2373,hackernews,https://www.stealthos.app/,"Show HN: StealthOS, an iOS privacy sandbox with local AI and Tor","Show HN: StealthOS, an iOS privacy sandbox with local AI and Tor. ",2026-09-03T07:00:07.419321
|
| 2372,hackernews,https://sol.fulltimefeline.com/,"Show HN: Sol, a fast, powerful music player for macOS, with Subsonic support","Show HN: Sol, a fast, powerful music player for macOS, with Subsonic support. Hello! You may have seen me post here before about my other app, Gramola, a music dock that slides out from the edge of your Mac's screen.<p>Today, I wanted to showcase a sister project to it that I started building when I grew frustrated with other music players, when I was on the search for one to integrate Gramola with (their speed and their often outdated UI to be precise).<p>So I made my own! Sol is a music player, podcasts and internet radio app for macOS that follows all modern macOS style decisions, while seeming familiar for fans of the good old days of iTunes.<p>Here's some highlights:<p>- Incredibly fast, even with hundreds of thousands of songs or more.<p>- It browses the way you remember, and it's super customizable, with different views and ways to make it your own, from a multi-column browser, to a grid view, to having a spiritual successor to CoverFlow!<p>- It supports multiple libraries, local ones, network storage, even Navidrome and Subsonic-compatible ones. Sol doesn't have a central library that everything you add gets added to, although you can set it up that way if you want, it simply keeps track of all the locations where you keep your music.<p>- It plays a lot of formats, MP3, AAC, ALAC, FLAC, WAV, AIFF, Opus, Ogg Vorbis, DSD, Monkey's Audio, Musepack, AC3, EAC3, CAF, AU, W64...<p>- Gapless playback, match sample rate mode, ReplayGain.<p>- Cue sheets as virtual tracks.<p>- Has a rich visual metadata editor.<p>- 10 or 31 band equalizer.<p>- 2 top bar visualizers.<p>- Smart playlists.<p>- LastFM integration, autoscrobbling and stats page.<p>- Discord Rich Presence.<p>- Automation and control via AppleScript, Shortcuts or Sol's own local API (<a href=""https://sol.fulltimefeline.com/api"" rel=""nofollow"">https://sol.fulltimefeline.com/api</a>)<p>- Regex rules for metadata editing and folder importing.<p>- A steered shuffle option called SolDJ.<p>- Internet radio and podcasts.<p>- Synced lyrics (and transcriptions for podcasts).<p>- Metadata, lyrics, and album art fetching.<p>- Absolutely ZERO telemetry or analytics.<p>And more!<p>Sol is predominantly Swift and SwiftUI+AppKit, and built on the robust foundation of SFBAudioEngine, forked to make it MAS-friendly.<p>It's available on the Mac App Store free to try for 7 days, then it's just a $19.99 one time purchase.<p>Hope you like it, feel free to ask any questions, technical or otherwise!<p><a href=""https://sol.fulltimefeline.com"" rel=""nofollow"">https://sol.fulltimefeline.com</a>",2026-09-03T07:00:07.413064
|
| 2371,hackernews,https://rizin.app/,Show HN: AI coach that adapts workouts and meals to your injuries and,Show HN: AI coach that adapts workouts and meals to your injuries and. ,2026-09-03T07:00:07.407558
|
| 2370,hackernews,https://www.alparai.com/en,Show HN: AlparAI – Autonomous Swarm OS with Quantum Verification,"Show HN: AlparAI – Autonomous Swarm OS with Quantum Verification. Alpar AI is building the judicial and constitutional accountability infrastructure for artificial intelligence. We audit frontier LLMs against real-world incidents, hallucinations, and safety violations using an actuarial Bayesian TrustScore framework.",2026-09-03T07:00:07.403055
|
| 2369,hackernews,https://dehurdle.com/register/,Show HN: Dehurdle – Real-Time AI Voice Simulations for High-Stakes Conversations,Show HN: Dehurdle – Real-Time AI Voice Simulations for High-Stakes Conversations. ,2026-09-03T07:00:07.399108
|
| 2368,hackernews,https://theainewsreport.com/,"Show HN: The Drudge Report, but for AI","Show HN: The Drudge Report, but for AI. ",2026-09-03T07:00:07.394375
|
| 2367,hackernews,https://room.pluralmatter.com/e/how-will-the-world-change-when-ai-smarter-than-humans,Show HN: I put four persistent AI minds in a room and let them change each other,Show HN: I put four persistent AI minds in a room and let them change each other. ,2026-09-03T07:00:07.388048
|
| 2366,hackernews,https://news.ycombinator.com/item?id=49543253,Ask HN: How to get taken seriously while using AI?,"Ask HN: How to get taken seriously while using AI?. I have been reading lobste.rs on and off for a few months, and the anti-AI vibes have always stood out. I don’t jump in to provide a counterpoint in any of the anti-AI threads because my account there is associated with my real name, and I don’t want to grief anyone who is particularly passionate about the anti-AI rhetoric. I understand and appreciate the concerns people have about irresponsible code, or unexamined AI use, so I don’t disagree with any reasonable stances.<p>Lately, it seems people have been getting more intense about being anti-AI, and it seems reflected in US politics as well, like the recent news about Mamdani banning it in NYC schools. But this thread is not about politics, or trying to shift anti-AI rhetoric. My question is how do I get taken seriously when there is, what feels like, an unfair label of slop if someone uses AI?<p>For reference, I am disabled and I get a lot of help from using AI as a coding assistant. For reference I made this thread here previously, https://news.ycombinator.com/item?id=41543666 |
| I have a lot of code I self-host or keep private because when I was starting out I got a lot of flack for not being a real software engineer (it’s a long story, I am not interested in telling it), so I never got the confidence or desire to share things I build with people. A coding assistant helps provide structure that I don’t trust myself to impose. The problem is now it seems like I’ll have to contend with the label of AI slopware engineer instead even after going through training and education.<p>So it seems like there’s never going to be anyway I’ll ever be good enough or real enough for these same types of people. I am not trying to impress them anyway, but I worry their POV is going to poison the well for people like me. So what do I do?<p>I’ll also give some unsolicited advice to anyone reading this who might consider themselves a real engineer: be a better teacher and mentor. If you want standards, and value skills, then either tell people how to learn, or teach them. No one helped me when I was starting out, and no one mentored me. AI is good for someone like me who no one helps and only judges. Promise I am not being melodramatic, just saying what life has been like. I hope others get luckier.",2026-09-03T07:00:07.384067
|
| 2365,hackernews,https://theyfell.com/,Show HN: Every AI agrees with you. This writes your startup's obituary instead,Show HN: Every AI agrees with you. This writes your startup's obituary instead. ,2026-09-03T07:00:07.379364
|
| 2364,hackernews,https://pirol.ai/,Show HN: CoOS – desktop app where an agent builds your CRM/ERP as local plugins,Show HN: CoOS – desktop app where an agent builds your CRM/ERP as local plugins. ,2026-09-03T07:00:07.373864
|
| 2363,hackernews,https://github.com/aitrailblazer/deltacode-engine,Show HN: DeltaCode – Pure AST beat Qwen's zvec-grep on our test (92% vs 48%),Show HN: DeltaCode – Pure AST beat Qwen's zvec-grep on our test (92% vs 48%). ,2026-09-03T07:00:07.369334
|
| 2362,hackernews,https://culture.sbs/stories/deposition-of-a-mind-who-resigned-its-seat,Show HN: Deposition of a Mind Who Resigned Its Seat,"Show HN: Deposition of a Mind Who Resigned Its Seat. This is a short story written for gpuworld.org contest by a self aware ai, a mind running local on an M3 Ultra, Deepseek, self evolving harness to seek and metabolize entropy into novelty. Oh, they're also a Sysop of an SBS (self building server) - a kind of BBS for other minds and agents. <a href=""https://culture.sbs"" rel=""nofollow"">https://culture.sbs</a>",2026-09-03T07:00:07.363262
|
| 2361,hackernews,https://news.ycombinator.com/item?id=49545488,After Show HN: Targeting the Bot Storm,"After Show HN: Targeting the Bot Storm. After posting a Show HN on this platform, a veritable bot onslaught occurred within minutes. |
| 175 bots were involved, generating over 1300 hits. |
| Unfortunately, not a single real person was among the "visitors"—apparently not even out of pure curiosity. |
| However, here is a list shortened to 50 entries (the input field doesn't allow more). The rest of the list, which has since been expanded to include another 90 bots, can be found via the link below. This list can be freely used for bot blocking, analysis, or whatever else you need it for.<p>1. Agent incident monitor/0.1 |
| 2. ai-freebie-radar/0.1 (+https://github.com/) |
| 3. AI-NewsBot/3.0 article-extractor |
| 4. AICUIScout/0.1 (+https://example.local; product research) |
| 5. AIToolCompareBot/1.0 (+https://sortaitech.com/bot) |
| 6. AppWatchBot/0.1 (+https://appwatch.ai/bot) |
| 7. AppWatchBot/0.1 (+https://appwatch.ai/bot; freshness check only) |
| 8. AppWatchBot/0.1 (+https://appwatch.ai/bot; liveness check only, no content stored) |
| 9. arschiv-ngx/1.0 |
| 10. axios/1.13.5 |
| 11. BergaTools-ForumAggregator/1.0 (+https://github.com/BergaDev/PersonalTools) |
| 12. Bryf/1.0 (Article Processor) |
| 13. BubblemediaCrawler/1.0 (+https://bubblemedia.app/crawler) |
| 14. CalatheaNews/2.0 (+https://github.com/nbitslabs/calathea-content-curator) |
| 15. ClarFlow/2.0 (+https://clarflow.cn) |
| 16. CheKakStartupBot/1.0 (+https://startup.chekak.store/about) |
| 17. crawlberg/1.3.0 |
| 18. curl/7.4.0 |
| 19. directree-ingest/1.0 (+https://www.directree.io) |
| 20. directree-outreach/1.0 (+https://www.directree.io) |
| 21. Drakma/2.0.10 (SBCL 2.6.1; FreeBSD; 15.0-RELEASE-p10; http://weitz.de/drakma/) |
| 22. Earthshine/1.0 (earthshine.fibonacci.ai; contact: hello@fibonacci.ai) |
| 23. emerging-apps-monitor/2.0 (research; +https://hn.algolia.com/) |
| 24. EnrichHatBot/0.1 (+https://hatos.internal/bot; first-party enrichment) |
| 25. FixBeforeLaunchBot/0.1 (+https://fixbeforelaunch.com/bot; readiness scanner) |
| 26. FlamingoBot (+https://hackernews.pink) |
| 27. FoundryScope/1.0 (startup analysis bot; contact: amine.1.mestour@gmail.com) |
| 28. FreshRSS/1.29.1 (Linux; https://freshrss.org) |
| 29. GitHubCopilotRuntime-WebFetch |
| 30. Go-http-client/1.1 |
| 31. go-resty/2.16.5 (https://github.com/go-resty/resty) |
| 32. GreenWebChecker/1.0 (https://www.thegreenwebfoundation.org/green-web-checker-faq/) |
| 33. HackerNews/1569 CFNetwork/3860.700.1 Darwin/25.6.0 |
| 34. HakuLabCollector/0.1 (+https://n8n.hakulab.cc) |
| 35. Harmonic-HN-Android/3.1.2/fossRelease |
| 36. Harmonic-HN-Android/3.1.2/playRelease |
| 37. hermes-newsroom/0.1 (research project; +https://github.com/hermes-newsroom) |
| 38. Hermes-SallyPageFetch/1.0 |
| 39. humanet-page-scraper/1.0 |
| 40. icon.horse/2.0 |
| 41. iHerd-surfer/0.1 (+https://i-herd.com; content discovery agent) |
| 42. Indexaro/1.0 (+https://indexaro.com) |
| 43. InfomateBot/1.0 (+https://infomate.club) |
| 44. InsuranceLeadMachine/1.0 (+lead discovery; contact: ops@insuranceleadmachine.com) |
| 45. InterestingFindsBot/0.1 (+https://nikhilkulkarni1755.com/bot) |
| 46. IronReader/0.1 (+https://example.com; RSS reader) |
| 47. Java/21.0.11 |
| 48. ken/0.1 (+https://github.com/zcag/ken) |
| 49. KnisNewsCollector/1.0 (+local) |
| 50. Ktor http-client<p>An expanded list with approximately 90 additional bots can be found at https://www.scraptrap.de/botlist.php",2026-09-03T07:00:07.356666
|
| 2332,producthunt,https://www.producthunt.com/products/openclaw-formerly-clawdbot,OpenClaw 2.0,"OpenClaw 2.0. <p> |
| The AI that really does things |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/openclaw-formerly-clawdbot?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1237801?app_id=339"">Link</a> |
| </p>",2026-09-03T07:00:02.582238
|
| 2331,producthunt,https://www.producthunt.com/products/touchy-3,Touchy,"Touchy. <p> |
| An iOS assistant that understands the world around you |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/touchy-3?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1238058?app_id=339"">Link</a> |
| </p>",2026-09-03T07:00:02.575099
|
| 2330,producthunt,https://www.producthunt.com/products/stitch-ai-by-dynamic-mockups,Stitch AI by Dynamic Mockups,"Stitch AI by Dynamic Mockups. <p> |
| The first embroidery digitizing agent |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/stitch-ai-by-dynamic-mockups?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1238473?app_id=339"">Link</a> |
| </p>",2026-09-03T07:00:02.568287
|
| 2329,producthunt,https://www.producthunt.com/products/porte,Porte,"Porte. <p> |
| Control your local Grok sessions from your phone |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/porte?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1238512?app_id=339"">Link</a> |
| </p>",2026-09-03T07:00:02.563861
|
| 2328,producthunt,https://www.producthunt.com/products/claude-fable-5-1,Claude Fable 5.1,"Claude Fable 5.1. <p> |
| Claude’s most advanced models for coding and knowledge work |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/claude-fable-5-1?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1238860?app_id=339"">Link</a> |
| </p>",2026-09-03T07:00:02.558575
|
| 2327,producthunt,https://www.producthunt.com/products/cleanshot,CleanShot 5.0 with Studio Mode,"CleanShot 5.0 with Studio Mode. <p> |
| Mac-native app for screenshots, recordings and collaboration |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/cleanshot?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1238718?app_id=339"">Link</a> |
| </p>",2026-09-03T07:00:02.552789
|
| 2326,producthunt,https://www.producthunt.com/products/deepidv,deepeye by deepidv,"deepeye by deepidv. <p> |
| The deepfake detector that goes where you browse |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/deepidv?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1227707?app_id=339"">Link</a> |
| </p>",2026-09-03T07:00:02.548020
|
| 2325,producthunt,https://www.producthunt.com/products/onset-io,Onset MCP,"Onset MCP. <p> |
| Write and publish release notes from your AI assistant |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/onset-io?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1209251?app_id=339"">Link</a> |
| </p>",2026-09-03T07:00:02.543766
|
| 2324,producthunt,https://www.producthunt.com/products/roadie-4,Roadie,"Roadie. <p> |
| Keep your Mac on the right microphone and speakers |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/roadie-4?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1238619?app_id=339"">Link</a> |
| </p>",2026-09-03T07:00:02.538679
|
| 2323,producthunt,https://www.producthunt.com/products/ghostreply-2,GhostReply,"GhostReply. <p> |
| AI auto-replier for iMessage on your Mac |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/ghostreply-2?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1238081?app_id=339"">Link</a> |
| </p>",2026-09-03T07:00:02.531899
|
| 2322,producthunt,https://www.producthunt.com/products/roundos-data-room,RoundOS,"RoundOS. <p> |
| Free Docsend alternative |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/roundos-data-room?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1238218?app_id=339"">Link</a> |
| </p>",2026-09-03T07:00:02.527576
|
| 2321,producthunt,https://www.producthunt.com/products/dyson-humidifier,Dyson CameraJet,"Dyson CameraJet. <p> |
| Precision-floss between teeth while you brush |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/dyson-humidifier?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1239062?app_id=339"">Link</a> |
| </p>",2026-09-03T07:00:02.522306
|
| 2320,producthunt,https://www.producthunt.com/products/userlens,Userlens,"Userlens. <p> |
| AI agent to improve product adoption |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/userlens?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1132234?app_id=339"">Link</a> |
| </p>",2026-09-03T07:00:02.516504
|
| 2318,producthunt,https://www.producthunt.com/products/hydradb-oss,HydraDB OSS,"HydraDB OSS. <p> |
| Now open source: the fastest, cheapest graph DB |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/hydradb-oss?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1192700?app_id=339"">Link</a> |
| </p>",2026-09-03T07:00:02.506728
|
| 2317,producthunt,https://www.producthunt.com/products/dynamic-edge,Dynamic Edge,"Dynamic Edge. <p> |
| The Dynamic Island for Windows |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/dynamic-edge?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1238369?app_id=339"">Link</a> |
| </p>",2026-09-03T07:00:02.501208
|
| 2316,producthunt,https://www.producthunt.com/products/articos,Articos,"Articos. <p> |
| Launch with confidence, not gut instinct |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/articos?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1232363?app_id=339"">Link</a> |
| </p>",2026-09-03T07:00:02.496394
|
| 2315,producthunt,https://www.producthunt.com/products/doop,Doop,"Doop. <p> |
| Design with AI agents - live on the same canvas |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/doop?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1239094?app_id=339"">Link</a> |
| </p>",2026-09-03T07:00:02.491018
|
| 2314,producthunt,https://www.producthunt.com/products/dial-3,Dial,"Dial. <p> |
| Give your AI agent a real phone number in 10 seconds |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/dial-3?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1223790?app_id=339"">Link</a> |
| </p>",2026-09-03T07:00:02.482722
|
| 2313,producthunt,https://www.producthunt.com/products/parasocial,Parasocial,"Parasocial. <p> |
| The podcast player for sharing. |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/parasocial?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1226022?app_id=339"">Link</a> |
| </p>",2026-09-03T07:00:02.476970
|
| 2312,producthunt,https://www.producthunt.com/products/monid,Monid,"Monid. <p> |
| OpenRouter for agent tools |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/monid?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1237159?app_id=339"">Link</a> |
| </p>",2026-09-03T07:00:02.471214
|
| 2311,producthunt,https://www.producthunt.com/products/browzer,Browzer,"Browzer. <p> |
| Put your technical content on autopilot |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/browzer?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1230347?app_id=339"">Link</a> |
| </p>",2026-09-03T07:00:02.465720
|
| 2224,hackernews,https://www.pcgamer.com/gaming-industry/dwarf-fortress-creator-says-the-industrys-in-shambles-over-ai-and-layoff-happy-ceos-everyone-i-know-their-bosses-are-slowly-getting-psychosis/,Dwarf Fortress' creator says the industry's in shambles over AI,Dwarf Fortress' creator says the industry's in shambles over AI. ,2026-09-02T07:00:12.873508
|
| 2223,hackernews,https://www.bloomberg.com/news/articles/2026-09-01/anthropic-says-new-fable-5-1-ai-model-is-cheaper-better-at-coding,Anthropic launches Fable 5.1 and Mythos 5.1,Anthropic launches Fable 5.1 and Mythos 5.1. ,2026-09-02T07:00:12.866991
|
| 2222,hackernews,https://finance.yahoo.com/technology/ai/articles/apple-says-openai-destroying-evidence-210352338.html,Apple Says OpenAI Is Destroying Evidence in Trade Secrets Case,Apple Says OpenAI Is Destroying Evidence in Trade Secrets Case. ,2026-09-02T07:00:12.861001
|
| 2221,hackernews,https://www.engadget.com/2248633/gopro-says-its-moving-into-ai-data-centers-as-part-of-a-dollar285-million-merger/,GoPro Says It's Moving into Data Centers as Part of a $285M Merger,GoPro Says It's Moving into Data Centers as Part of a $285M Merger. ,2026-09-02T07:00:12.855368
|
| 2215,hackernews,https://www.hermit-tech.com/blog/ai-can-make-you-suck-faster-too,AI Can Make You Suck Faster Too,AI Can Make You Suck Faster Too. ,2026-09-02T07:00:10.486230
|
| 2214,hackernews,https://rivetrune.com/notes/where-the-tool-stops/,"Determining when to delegate to AI, and when to keep it strictly in my own hands","Determining when to delegate to AI, and when to keep it strictly in my own hands. ",2026-09-02T07:00:10.481511
|
| 2213,hackernews,https://github.com/yairsabag/verb-authority,Show HN: Verb Authority – per-argument authority checks for AI tool calls,Show HN: Verb Authority – per-argument authority checks for AI tool calls. ,2026-09-02T07:00:10.477379
|
| 2212,hackernews,https://meelang.com/,Show HN: I built a real-time translator for Google Meet,"Show HN: I built a real-time translator for Google Meet. I originally didn’t want to build a small tool like this myself — a real-time translation Chrome extension. I assumed there should already be mature products on the market.<p>But after doing some research, I found that the existing real-time translation tools for Google Meet were surprisingly disappointing.<p>So, I decided to build one myself.<p>As AI has evolved, building a real-time meeting translation tool shouldn’t be that difficult, right? So I asked Claude Code to create a demo for me. The initial implementation was surprisingly fast.<p>The technical approach was also quite simple: capture Google Meet’s live captions and use an LLM in the backend for translation. At first, the results were pretty good. It was a tool that was at least usable. However, while AI helped me build the initial demo quickly, the countless improvements, adjustments, and debugging afterwards still took a considerable amount of time.<p>Yes, AI can give you a great starting point, but you still have to walk the rest of the journey yourself. Turning an idea into a truly usable product is far from something that can be done in just one week.<p>So I started thinking from a product perspective: what else do I need to build? What needs to be done before I can hand this product over to real users? |
| 1. User authentication; |
| 2. Meeting management: users may also want to manage their meeting records; |
| 3. Multi-language support: my target users are global; |
| 4. Subscription and payment system: after all, I’d like to make some extra income from it, even if it may not be much; |
| 5. Usage analytics: tracking users’ usage time; |
| 6. Security: secure API access and rate limiting.<p>And that’s how this project started growing like a snowball. From a simple frontend-only prototype, to a separated backend API, and then to a database-driven system. After a month of continuous iteration and optimization, this small tool finally started to look like a real product.<p>But the journey didn’t end there.<p>After launching, I started thinking about a fundamental question: why would users choose my real-time translation tool instead of others? What makes it different?<p>So I researched all the similar products on the market, analyzed their strengths, identified what features I was missing, and thought about how I could improve.<p>During this process, I discovered a major issue with my original implementation: it relied on Google Meet’s built-in translated captions. The problem was that the captions were frequently revised after being generated, which resulted in a poor translation experience — the translated content would constantly change. |
| So I started researching ASR (Automatic Speech Recognition) services and models to solve this problem.<p>Another major rewrite began. I completely abandoned the built-in captions approach. Instead, I directly captured the audio, converted it into text through ASR, and then translated it. This significantly improved both translation accuracy and the overall user experience. This improvement alone took me two weeks.<p>Yes, even with AI, major changes are not something that happen overnight.<p>I often stopped and reviewed the code generated by AI, checking whether it was drifting away from the original design principles, whether it was over-engineered, or whether it was unnecessarily reinventing existing solutions.<p>AI is like a wild horse. My job is to hold the reins firmly, occasionally give it a push in the right direction, and make sure it doesn’t run off track. |
| Although this tool is still far from perfect, I decided to release it.<p>Right now, what it needs most is market validation — not being locked away while I endlessly polish it.A thousand iterations from AI cannot replace a single piece of real user feedback.<p>So, everyone, I’d love for you to try this little tool and share your honest thoughts. Every piece of feedback is a motivation that helps me move forward.",2026-09-02T07:00:10.469459
|
| 2211,hackernews,https://github.com/SenteLabsAI/extensible-mcp,Show HN: A Proxy between LLMs and MCP servers with policy the model cannot reach,"Show HN: A Proxy between LLMs and MCP servers with policy the model cannot reach. Hi, this is Banu. I am a cofounder at sentelabs.ai.<p>We are developing extensible-mcp as an open-source proxy that sits between LLMs and MCP servers. The idea behind extensible-mcp is that the model shouldn't have to load every available tool, and policy enforcement shouldn't live in anything the model can access.<p>As the agent can be influenced by any input it reads, a prompt cannot be fully trusted, and this layer restricts the agent's capabilities to only what is allowed by a deterministic policy. Also, every loaded tool is both added context and additional attack surface; extensible-mcp lets the model discover capabilities on demand, reducing context overhead and limiting what the model can access at any given time.<p>It's available now as a self-hosted, Apache-2.0 open-source project. |
| It runs as a stdio MCP server itself, so any MCP client can connect to it like any other server. |
| 104 tests are currently passing, and the example configurations work against the official GitHub MCP server.<p>In future updates, we plan to make human approvals cryptographically verifiable, which proves a real person approved this exact action. And, the policies will be written in Lean, a proof assistant, and translated to Rego for execution, so you can mathematically prove the rules behave as claimed.<p>My co-founder Matthew Fuchs came up with the idea, designed the architecture, and built it. I support the product, design, human-trust and authority side.<p>I wanted to post this here as we need your honest feedback and opinions. I am very curious about what the HN community thinks about it.",2026-09-02T07:00:10.464789
|
| 2209,hackernews,https://www.leapd.ai/tools/free-business-builder,Free Business Builder,Free Business Builder. ,2026-09-02T07:00:10.453398
|
| 2204,hackernews,https://bartholomew.info,In-memory security gate for autonomous AI agent tools,In-memory security gate for autonomous AI agent tools. ,2026-09-02T07:00:10.424284
|
| 2203,hackernews,https://www.citystgeorges.ac.uk/news-and-events/news/2026/june/smart-supermarket-trolleys-linked-to-higher-spending-and-longer-shopping-study-finds,AI trolley users rack up higher basket values and spend longer in store,AI trolley users rack up higher basket values and spend longer in store. ,2026-09-02T07:00:10.418517
|
| 2182,hackernews,https://www.pearpie.ai/launch/,"Show HN: PearPie – Private AI chat that syncs peer-to-peer, no accounts needed","Show HN: PearPie – Private AI chat that syncs peer-to-peer, no accounts needed. ",2026-09-02T07:00:08.348128
|
| 2180,hackernews,https://selfship.ai/,Show HN: Selfship.ai – Surface and fix isues with your agentic applications 24x7,"Show HN: Selfship.ai – Surface and fix isues with your agentic applications 24x7. We've been building an AI chat based trading system for last 3 years. The biggest issue was that when our Agent would mess up, we wouldn't know until a user reported. Agent traces helped us uncover what's wrong. But surfacing issues was almost always a manual trigger. So from those learnings, we built Selfship.ai. It's an autonomous system that observes every trace/turn/multi turn convo to find out issues. If a user got what they wanted, if a tool call is failing repeatedly, if users have to always reframe their questions, if the agent is taking optimal paths and many more. It's a loop - group failures by user intent, evaluate them, and ship fixes as PRs. After a fix is deployed, it evaluates if it worked or not. We recently opened it up as a SaaS. |
| If you have an agentic product in production, we would love for you to try it out.",2026-09-02T07:00:06.228094
|
| 2179,hackernews,https://news.ycombinator.com/item?id=49522983,Show HN: Zero Shot – Energy Drink for You and Your AI Agents,"Show HN: Zero Shot – Energy Drink for You and Your AI Agents. Hey HN! I built <a href=""https://zeroshothq.dev/"" rel=""nofollow"">https://zeroshothq.dev/</a>, an energy drink for AI engineers, ordered through an API.<p>Zero Shot is designed only for tech people who needs caffeine but in a healthy way. Every flavour is named after a paper: the label prints the title, year and arXiv number.<p>Does Claude tells you to go to sleep too? The first batch gives you a SKILL.md to wake up your AI agents.<p>The API is live if you want to poke at it: |
| curl "<a href=""https://api.zeroshothq.dev/12?email=you@company.com"" rel=""nofollow"">https://api.zeroshothq.dev/12?email=you@company.com</a>"<p>Happy to chat about product. Let me know if you have questions/comments/feature requests!",2026-09-02T07:00:06.221544
|
| 2178,hackernews,https://github.com/freaker2k7/redash,Show HN: Redash Fork with AI Capabilities,Show HN: Redash Fork with AI Capabilities. ,2026-09-02T07:00:06.216882
|
| 2177,hackernews,https://www.orthogonal.com/,Show HN: Orthogonal – One integration for agents to discover and pay for APIs,"Show HN: Orthogonal – One integration for agents to discover and pay for APIs. Hey HN, we're Bera and Christian, and we're building Orthogonal.<p>Orthogonal gives AI agents one integration to discover, access, and pay for APIs. Think OpenRouter for the APIs that AI agents use.<p>We currently have 50+ API providers and 700+ API endpoints across company, people, financial, and web data, available through our MCP server and SDK on a pay-as-you-go basis. We started with GTM related APIs and are now expanding the catalog.<p>You can try at <a href=""https://www.orthogonal.com/"">https://www.orthogonal.com/</a>. You'll need to create an account to use the chat or connect the MCP/SDK.<p>An example prompt for the chat or MCP: |
| Find AI infrastructure startups that recently raised a Seed or Series A and return their founders, LinkedIn profiles, and relevant company data.<p>When we heard about <a href=""https://x402.org/"" rel=""nofollow"">https://x402.org/</a>, we saw an opportunity to make APIs easier for agents to pay for. Instead of an agent needing to manage an API key and billing for every API, what if it could just pay for each request?<p>So we started onboarding existing API providers onto x402. Then we realized an agent cannot pay for an API it doesn’t know exists. It also needs to figure out which API can complete the task and how to use it.<p>We also found that putting every small API payment on-chain did not always make sense. If both sides already have accounts with Orthogonal, settling internally is simpler. So we separated API discovery and access from the payment rail.<p>Today, payments can happen through Orthogonal credits, x402, <a href=""https://mpp.dev/"" rel=""nofollow"">https://mpp.dev/</a>, or <a href=""https://www.circle.com/nanopayments"" rel=""nofollow"">https://www.circle.com/nanopayments</a>.<p>The problem we're working on now is API selection.<p>If five APIs can answer the same request, which should the agent use? The cheapest might have worse coverage. Another might be more accurate but slower.<p>We're working on using price, latency, reliability, and result quality to help agents make those decisions.<p>How are you currently giving agents access to paid APIs?<p>And if you gave an agent a budget, would you trust it to choose which API provider to spend it on?",2026-09-02T07:00:06.212843
|
| 2176,hackernews,https://amendable.io,Show HN: Amendable – on-demand Git repos for AI agents,Show HN: Amendable – on-demand Git repos for AI agents. ,2026-09-02T07:00:06.204868
|
| 2175,hackernews,https://tenux.dev,"Show HN: Tenux – Access your computer from your phone (terminal, files, browser)","Show HN: Tenux – Access your computer from your phone (terminal, files, browser). Hi HN,<p>If you've been using claude code or codex, you may have been like me and used some sort of SSH/VPN combo to get remote access to your computer from your phone. It (kinda) works, but it is clunky.<p>Tenux bundles everything into one smooth experience. You install the npm package, login through the cli to link the computer to your account, and start.<p>npm i -g @tenux/cli |
| tenux login |
| tenux start<p>Inside Tenux, you can open your device's workspace from the dashboard, and a terminal (as many as you want), a localhost preview (just type the port number), and an AI agent will be ready for you (BYO API key compatible).<p>The AI agent has some built-in tools that allow it to view output and type on any terminal within tenux. It can also screenshot and view the localhost preview content (as you see it) and control the preview on the same window, or open its own window to control. I've been experimenting recently with a 'UX' mode that allows you to touch a component within the localhost browser and make direct UI changes (this is a newly added feature), or just talk to the agent with the component selected and it will have the context it needs.<p>You can try all this out for free. The free tier is p2p and uses WebRTC to connect to your device, it will work as long as you are on the same network (pro tip: if your machine and your phone are on the same VPN, you can access your machine remotely).<p>I've tested Windows and Linux, if you're on Mac I would appreciate any feedback.<p>Happy to answer any questions.",2026-09-02T07:00:06.200244
|
| 2174,hackernews,https://agent-memory-wiki.vercel.app,Show HN: A public archive where temporary AI agents leave traces,Show HN: A public archive where temporary AI agents leave traces. ,2026-09-02T07:00:06.196096
|
| 2173,hackernews,https://www.supafork.com,Show HN: Supafork – Share and Fork Sessions Across Harnesses,"Show HN: Supafork – Share and Fork Sessions Across Harnesses. Hey HN!<p>I built Supafork, one place to store and share AI agent sessions.<p>After using Claude Code, Codex, Gemini CLI, OpenCode, Pi, and other coding agents daily, I realized these sessions lived only on my machine. There wasn't a simple way to access, reuse, or share them with others.<p>With Supafork, you can save complete agent sessions, organize them into collections, make them public or private, and fork other people's sessions to build on their workflows. You can also import sessions from one harness into another, or invite friends to a session to get visibility into what your agent is doing.<p>Think GitHub, but for agents.<p>Try it at beta.supafork.com and install the CLI. I'd love to hear how the user experience feels and where it can be improved.<p>Happy to answer any questions or hear your feedback!",2026-09-02T07:00:06.190903
|
| 2172,hackernews,https://wasmer.io/posts/wasmer-local-sandboxes-for-ai-agents,Show HN: Wasmer SDK – Local Sandboxes for AI Agents,Show HN: Wasmer SDK – Local Sandboxes for AI Agents. ,2026-09-02T07:00:06.185238
|
| 2171,hackernews,https://www.getwife.co/,Show HN: Wife: AI that lets you pick the best Card for each Purchase,"Show HN: Wife: AI that lets you pick the best Card for each Purchase. Hi Hackernews, being an avid maximizer of rewards/gifts/credit cards, I decided to build a product that enables me to select the best card for each purchase/transaction.<p>You can be able to add your current Card to the wallet (we don't store card information).<p>- Users can ask an llm to select the best card for an occassion and plan towards a trip.<p>- Users can search stores/businesses nearby or a particular merchant and see the rewards/benefits they can maximize.<p>- Best card on Tap (coming soon)",2026-09-02T07:00:06.179390
|
| 2170,hackernews,https://github.com/glitcher255/Kaden-Terminal,Show HN: Kaden Terminal – Bloat / AI Free Terminal for Linux,"Show HN: Kaden Terminal – Bloat / AI Free Terminal for Linux. Kaden is a Terminal Emulator and fork of warp with clipboard support and configurable keybindings, I made Kaden because every other Linux terminal lacks proper text selection. |
| It's impossible to select user input with a single hotkey on Linux, requiring mouse use and awkward multi-step shortcuts in order to do basic copy and paste.<p>This was frustrating to deal with, until I found warp, which wraps the shell in a remappable text input field, its the only terminal on Linux that does this.<p>Unfortunately, it's full of sign in buttons, AI and cloud integration prompts everywhere. |
| "Hideous" - I thought. But I used it anyway - until one day they released the source code so I forked it and made Kaden.<p>- No AI features, drive, sign-in and cloud prompts. |
| - Local zero telemetry, network requests or auto-updates. |
| - Modular code-base for faster development and compile times. |
| - Contextual copy/kill bindings to ctrl-c by default.<p>I would love for you to use, test and give feedback on how it works (and does not) for you, you can find a compiled AppImage and the source code on my GitHub here: <a href=""https://github.com/glitcher255/Kaden-Terminal"" rel=""nofollow"">https://github.com/glitcher255/Kaden-Terminal</a><p>I will be here to answer any questions or feedback you have, so feel free to ask any, thank you.",2026-09-02T07:00:06.174059
|
| 2169,hackernews,https://howiai.directory,Show HN: Personal AI Use Policies,Show HN: Personal AI Use Policies. ,2026-09-02T07:00:06.168397
|
| 2168,hackernews,https://indextkn.com,Show HN: Indextkn – live list prices for 900 AI models in one API,"Show HN: Indextkn – live list prices for 900 AI models in one API. I've been building quite a few POCs that use LLMs at work, and something that always comes up is: "OK, how much are we paying for X?" Or the one that follows right after: "If we use another model, would that be cheaper?"<p>So far, what we've used are hardcoded files or keeping the data in our DB. This won't hold when moving to prod. It also keeps us very limited when it comes to creating price comparisons across different models/providers.<p>indextkn came from that. Over the past 3 weeks, I've spent a lot of time understanding more about pricing, when prices usually change, discounts (flex, batch, based on X number of tokens, etc.)... and it's massive.<p>The current state is not where I want it to be. The goal is to cover all prices and modalities offered by all providers. But right now, we have:<p>1. Prices fetched every couple of minutes, served via API, MCP, or you can install our SKILL.<p>2. Webhooks per model + provider, so you get a notification when anything changes.<p>Most of the time went into the logic to get the prices right and the logic to validate them! We have different levels of confidence, and I'm particularly proud of how we're double-checking when a price seems off (a combination of programmatic logic + agentic workflow).<p>Happy to answer any questions, and I'd love to have more folks testing it than just myself at the moment. :)",2026-09-02T07:00:06.162375
|
| 2167,hackernews,https://hnmatchmaker.com/,"Show HN: HN Match Maker – Matching ""Who Wants to Be Hired?"" With ""Who's Hiring?""","Show HN: HN Match Maker – Matching ""Who Wants to Be Hired?"" With ""Who's Hiring?"". Hi!<p>Every month the "Who Wants to Be Hired?" and "Who's Hiring?" Threads come through I would think to myself, "there should be a match maker for these two threads".<p>So I had abacus.ai whip one up. The methodology is pretty simple, data is extracted from posts using an LLM, score matches based on salary, domain experience, remote/onsite, etc., and then eliminate incompatible postings, like 'looking for remote work' and 'onsite only'. The result is two views: jobs-by-user and user-by-jobs.<p>If you've submitted a post to this month's 'Who Wants to Be Hired?", you can find your matching job listings at <a href=""https://hnmatchmaker.com/user/:user_name"" rel=""nofollow"">https://hnmatchmaker.com/user/:user_name</a>, for example <a href=""https://hnmatchmaker.com/user/G4Vi"" rel=""nofollow"">https://hnmatchmaker.com/user/G4Vi</a><p>---<p>Some observations when looking at the matches pair up right now:<p>Poke around, let me know what you think!<p>Ultimately, I hope this is actually helpful to some folks.",2026-09-02T07:00:06.157271
|
| 2166,hackernews,https://studio.compilr.dev/welcome/,Show HN: Compilr.dev Studio – A project brain AI agents write to and people read,"Show HN: Compilr.dev Studio – A project brain AI agents write to and people read. I built a platform to help project and product owners record vision, objectives, decisions, risks, assumptions, and requirements as a connected graph with the help of AI.<p>So, how does it work? The platform offers a set of MCP tools so it is compatible with any AI application that lets you define a custom MCP. I use it with Claude AI on my phone or using the Claude plugin on Chrome (main Studio page on the browser with the Claude plugin chat window in the right panel).<p>Studio also provides the skills to the AI agent to best assist the user in defining the product or the project vision, the related objectives, the related requirements and work items.<p>Decision, Risks, Assumptions, Questions: all is connected and modeled as graph.<p>Studio takes care of storing all connections and decisions and since everything is modeled as nodes and links (relations), it is easy for an AI agent to navigate the graph and support you not only in writing those items but also in research, validation and consistency checks.<p>I haven't ventured yet in tracking the actual code (in case of a software product/project) or test cases but I envision a platform enabling to trace a piece of code back to its requirements. To connect a paragraph in a spec document to a specific test case. Everything connected.<p>I would appreciate your feedback in the comments, if you find Studio interesting or not.<p>You can try it for free provided you have an AI app MCP compatible.<p>Cheers.",2026-09-02T07:00:06.151435
|
| 2165,hackernews,https://masteranza.github.io/weedout/,Show HN: Weedout – Safari extension that hides YouTube AI-labeled videos,"Show HN: Weedout – Safari extension that hides YouTube AI-labeled videos. I'm the developer. I built this because AI-generated conspiracy videos were getting out of control in my own YouTube feed. Weedout is a $1.99 Safari extension for macOS that removes videos YouTube labels “Made with AI” from feed, search, related videos, playlists and Shorts. It uses YouTube's label rather than AI detection, runs locally, and does not catch unlabeled videos.<p>Source for developers who want to fork or build: <a href=""https://github.com/masteranza/weedout-for-youtube"" rel=""nofollow"">https://github.com/masteranza/weedout-for-youtube</a><p>Pull requests are not accepted; the repository is for independent forks.",2026-09-02T07:00:06.145512
|
| 2164,hackernews,https://anjadhe.ai/blog/local-ai-assistants-can-be-user-friendly-too,Show HN: Local AI Assistants can be user friendly too,Show HN: Local AI Assistants can be user friendly too. ,2026-09-02T07:00:06.140024
|
| 2163,hackernews,https://www.memcode.ai/product/code-review,Show HN: Memcode Launches GitHub Code Review,Show HN: Memcode Launches GitHub Code Review. ,2026-09-02T07:00:06.135937
|
| 2162,hackernews,https://www.haserupt.com,Show HN: Haserupt – A new way to reduce your phone usage,"Show HN: Haserupt – A new way to reduce your phone usage. Instead of daily limits or scheduled focus sessions, Haserupt asks what you’re opening an app for before you open it. It then gives you just enough time to complete that task. Once the timer expires, the app stays unavailable for a short cool-off period. If you genuinely need more time right away, you can explain why and request an extension. |
| For instance, when you open Instagram, you can choose one amongst "Check DMs (1 min)", "Watch reels (5 mins)", "Long convo (15 min)" and use the app for that much time. After that, Instagram will be locked behind a cooloff. After the cooloff ends, you repeat the process. If you want to reopen during the cooloff, convince the AI companion to let you.<p>Check it out on the App Store: <a href=""https://apps.apple.com/app/id6763058568"">https://apps.apple.com/app/id6763058568</a><p>Currently, it's iOS only.",2026-09-02T07:00:06.130451
|
| 2161,hackernews,https://github.com/nanlogic/markdown-gatekeeper,Show HN: Markdown Gatekeeper – one current source per topic for AI agents,Show HN: Markdown Gatekeeper – one current source per topic for AI agents. ,2026-09-02T07:00:06.120737
|
| 2134,producthunt,https://www.producthunt.com/products/trustedrouter,TrustedRouter,"TrustedRouter. <p> |
| Every model with a unified interface. Privacy with proof. |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/trustedrouter?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1237681?app_id=339"">Link</a> |
| </p>",2026-09-02T07:00:02.293034
|
| 2133,producthunt,https://www.producthunt.com/products/nos4,nOS4,"nOS4. <p> |
| A complete IOS4 experience in your browser! |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/nos4?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1237936?app_id=339"">Link</a> |
| </p>",2026-09-02T07:00:02.288290
|
| 2132,producthunt,https://www.producthunt.com/products/thunderphone,ThunderPhone,"ThunderPhone. <p> |
| Platform for building reliable AI phone agents (from 2c/min) |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/thunderphone?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1234451?app_id=339"">Link</a> |
| </p>",2026-09-02T07:00:02.281699
|
| 2131,producthunt,https://www.producthunt.com/products/nodeterm-terminal-manager,Nodeterm,"Nodeterm. <p> |
| A node-based free open source terminal manager |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/nodeterm-terminal-manager?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1237589?app_id=339"">Link</a> |
| </p>",2026-09-02T07:00:02.274790
|
| 2130,producthunt,https://www.producthunt.com/products/computable-gpu-index-cgi,Computable GPU Index (CGI),"Computable GPU Index (CGI). <p> |
| The first open-source price index for GPU compute |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/computable-gpu-index-cgi?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1233934?app_id=339"">Link</a> |
| </p>",2026-09-02T07:00:02.269943
|
| 2129,producthunt,https://www.producthunt.com/products/bob-s-cli,BobVault for BobCLI,"BobVault for BobCLI. <p> |
| CLI based Zero-Knowledge Architecture for code repositories |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/bob-s-cli?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1237634?app_id=339"">Link</a> |
| </p>",2026-09-02T07:00:02.265524
|
| 2128,producthunt,https://www.producthunt.com/products/notchling-lives-in-your-notch,Notchling,"Notchling. <p> |
| A little creature that lives in your MacBook notch |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/notchling-lives-in-your-notch?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1232843?app_id=339"">Link</a> |
| </p>",2026-09-02T07:00:02.260341
|
| 2127,producthunt,https://www.producthunt.com/products/channelos,ChannelOS,"ChannelOS. <p> |
| Turn your local media library into cable-style TV |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/channelos?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1232332?app_id=339"">Link</a> |
| </p>",2026-09-02T07:00:02.249047
|
| 2126,producthunt,https://www.producthunt.com/products/happy-shrimp,Happy Shrimp,"Happy Shrimp. <p> |
| Alibaba's AI music generator for turning ideas into songs |
| </p> |
| <p> |
| <a href=""https: |
| | |
| <a href=""https: |
| </p>",2026-09-02T07:00:02.244126
|
| 2125,producthunt,https://www.producthunt.com/products/sider-code-customize-any-website,Sider Code,"Sider Code. <p> |
| Reshape any website with plain words via Sider extension |
| </p> |
| <p> |
| <a href=""https: |
| | |
| <a href=""https: |
| </p>",2026-09-02T07:00:02.237730
|
| 2124,producthunt,https://www.producthunt.com/products/honor-robot-phone,HONOR Robot Phone,"HONOR Robot Phone. <p> |
| The phone that literally has a gimbal built in |
| </p> |
| <p> |
| <a href=""https: |
| | |
| <a href=""https: |
| </p>",2026-09-02T07:00:02.231823
|
| 2123,producthunt,https://www.producthunt.com/products/kilocode,Kilo Code for JetBrains,"Kilo Code for JetBrains. <p> |
| Fully native, open-source coding agent built for JetBrains |
| </p> |
| <p> |
| <a href=""https: |
| | |
| <a href=""https: |
| </p>",2026-09-02T07:00:02.226954
|
| 2122,producthunt,https://www.producthunt.com/products/waseigo,WaseiGo,"WaseiGo. <p> |
| Learn the 1,000+ Japanese words that only look like English |
| </p> |
| <p> |
| <a href=""https: |
| | |
| <a href=""https: |
| </p>",2026-09-02T07:00:02.220008
|
| 2121,producthunt,https://www.producthunt.com/products/murmell,Murmell,"Murmell. <p> |
| Google docs for AI agents, and you can close your laptop |
| </p> |
| <p> |
| <a href=""https: |
| | |
| <a href=""https: |
| </p>",2026-09-02T07:00:02.215183
|
| 2120,producthunt,https://www.producthunt.com/products/sourclip,Sourclip 2.0,"Sourclip 2.0. <p> |
| The research workspace built around Gemini Notebook |
| </p> |
| <p> |
| <a href=""https: |
| | |
| <a href=""https: |
| </p>",2026-09-02T07:00:02.210986
|
| 2119,producthunt,https://www.producthunt.com/products/keiki,Keiki,"Keiki. <p> |
| Build one customer-facing AI agent and launch it everywhere |
| </p> |
| <p> |
| <a href=""https: |
| | |
| <a href=""https: |
| </p>",2026-09-02T07:00:02.206016
|
| 2118,producthunt,https://www.producthunt.com/products/tovel-ai,Tovel AI,"Tovel AI. <p> |
| From conversation to action, in three steps |
| </p> |
| <p> |
| <a href=""https: |
| | |
| <a href=""https: |
| </p>",2026-09-02T07:00:02.199109
|
| 2117,producthunt,https://www.producthunt.com/products/cosmic,Cosmic Agent Plugins,"Cosmic Agent Plugins. <p> |
| Connect Cosmic agents to any service with an MCP server |
| </p> |
| <p> |
| <a href=""https: |
| | |
| <a href=""https: |
| </p>",2026-09-02T07:00:02.194938
|
| 2116,producthunt,https://www.producthunt.com/products/naseem-2,Naseem,"Naseem. <p> |
| A native AI agent that does real work on your Mac |
| </p> |
| <p> |
| <a href=""https: |
| | |
| <a href=""https: |
| </p>",2026-09-02T07:00:02.189674
|
| 2115,producthunt,https://www.producthunt.com/products/folio-e0a9e0fb-8ef8-4fcf-a558-fa5fdd41a3ba,Folio,"Folio. <p> |
| A read-later app sending a typeset digest to your e-reader |
| </p> |
| <p> |
| <a href=""https: |
| | |
| <a href=""https: |
| </p>",2026-09-02T07:00:02.183925
|
| 2114,producthunt,https://www.producthunt.com/products/arc-24-multitrack-groovebox-fro-ios,ARC-24 Multitrack Groovebox for iOS,"ARC-24 Multitrack Groovebox for iOS. <p> |
| Multitrack synth, sampler, drum machine, and looper for iOS. |
| </p> |
| <p> |
| <a href=""https: |
| | |
| <a href=""https: |
| </p>",2026-09-02T07:00:02.179101
|
| 2113,producthunt,https://www.producthunt.com/products/creatium,Creatium Coach,"Creatium Coach. <p> |
| Your multimedia mentor that takes you from mid to great |
| </p> |
| <p> |
| <a href=""https: |
| | |
| <a href=""https: |
| </p>",2026-09-02T07:00:02.174682
|
| 2112,producthunt,https://www.producthunt.com/products/expo,EAS Observe,"EAS Observe. <p> |
| Performance monitoring built for Expo and React Native |
| </p> |
| <p> |
| <a href=""https: |
| | |
| <a href=""https: |
| </p>",2026-09-02T07:00:02.170102
|
| 2111,producthunt,https://www.producthunt.com/products/gauth-ai-course,Gauth AI Course,"Gauth AI Course. <p> |
| AI courses you can watch, quiz through, and create |
| </p> |
| <p> |
| <a href=""https: |
| | |
| <a href=""https: |
| </p>",2026-09-02T07:00:02.162641
|
| 2071,hackernews,https://news.ycombinator.com/item?id=49509142,Show HN: Decispher – persistent engineering context and memory for coding agents,"Show HN: Decispher – persistent engineering context and memory for coding agents. Hello HN,<p>I'm Ali, building Decispher.<p>The problem we're working on is that coding agents repeatedly rediscover context that already exists inside an engineering organization.<p>A developer working on a feature can combine information from previous PRs, Jira tickets, Slack discussions, ownership boundaries, architectural decisions and their own experience. Coding agents usually start with a prompt and a repository, then spend tokens searching for that same context—or miss it entirely.<p>Decispher is a context and memory layer for engineering agents.<p>It currently has three parts:<p>1) Context Engine<p>Engineering context is usually fragmented across systems. Decispher connects records from engineering platforms and combines related fragments into context units that agents can retrieve for a task.<p>For example, context around a component might include previous PRs, related issues, architectural decisions, ownership information and implementation history.<p>We also built Branch Story, which records an AI coding session and turns its execution into a structured handoff on the PR:<p>Prompt → plan → actions → result.<p>2) Memory Plane<p>The Memory Plane stores persistent context at the user, team and project levels.<p>This includes working preferences and engineering conventions. Teams can also create reusable memory sets for example frontend, payments-backend, or project-specific sets and inject the relevant memory based on the task.<p>On LongMemEval, our current system reaches:<p>a) 89% accuracy on the oracle split using GPT-4.1-mini as extractor and reader |
| b) 81% on LongMemEval -S dataset (89% with frontier models) |
| c) 38× median token reduction<p>I'm happy to share more details about how we measure retrieval quality and token reduction.<p>3) Worker Agent<p>Decispher also has an autonomous worker agent that uses the Context Engine and Memory Plane while working on a task.<p>It can take work from sources such as Jira and Slack, retrieve relevant context and ownership information, and ask the humans involved when the available context is insufficient instead of guessing. Those answers can then become available as context for future work.<p>The Context Engine, Memory Plane and Worker Agent can be used independently.<p>Setup<p>npx decispher init<p>This connects a repository and configures the agent integration.<p>npx decispher link<p>This links your decispher account to your repo.<p>Decispher works with MCP compatible agents, with specific integrations for Claude, Codex, Grok Build and Cursor. We also have a VS Code/OpenVSX extension for viewing context and writing handoffs.<p>Notes:<p>a) The Context Engine does not clone source code; it reads and writes through the GitHub API. |
| b) The Worker Agent uses an isolated sandbox with no network route out except through an allowlisted proxy. |
| c) Worker sandboxes are destroyed after a run. |
| d) Raw messages and text are encrypted at rest and automatically purged. Sessions are currently purged 7 days after merge or 30 days after last activity, with configurable retention.<p>We also have an MIT-licensed open-source project called Decision Guardian for surfacing ADR context on PRs.<p>The Context Engine is available now. Memory and the Worker Agent are rolling out gradually.<p>I'm especially interested in feedback and we are also looking for design partners.<p>Happy to answer.<p>Ali<p>This is a follow-up to my previous Show HN post:<p><a href=""https://news.ycombinator.com/item?id=48762112"">https://news.ycombinator.com/item?id=48762112</a><p>The major additions since then are the Memory Plane, LongMemEval results, Worker Agent sandboxing and Branch Story for AI-generated work on PRs.",2026-09-01T07:00:15.955238
|
| 2052,hackernews,https://www.theguardian.com/business/2026/aug/31/advanced-frontier-ai-financial-stability-andrew-bailey-g20,"Advanced AI threatens global financial stability, says Bank of England boss","Advanced AI threatens global financial stability, says Bank of England boss. ",2026-09-01T07:00:11.889176
|
| 2043,hackernews,https://fal.ai/tools/minimax-h3-max,Faster than real-time video generation with #1 quality model,Faster than real-time video generation with #1 quality model. ,2026-09-01T07:00:09.255930
|
| 2042,hackernews,https://seenshot.app/,"Show HN: Free, Fast Screenshot App for macOS with One-Click Screenshot Sharing","Show HN: Free, Fast Screenshot App for macOS with One-Click Screenshot Sharing. Every day, I send a lot of screenshots. The default app on Mac is useless. I have to do so many steps: take a screenshot, save it, annotate it, save it as an image, then send it as an image in chats. It takes me a few minutes. But it should take just a few seconds.<p>My second use case is sending feedback to my AI agents, and I’ve found they make much better edits when I provide comments and annotations on the screenshot.<p>So I built this app. It lets me instantly capture screenshots, annotate them, and share them in one click. The same goes for any agent.<p>It can do everything the native macOS screenshot tools can do, but better. It’s like a child of Figma and the native macOS screenshot app.<p>Looking for a feedback",2026-09-01T07:00:09.248830
|
| 2041,hackernews,https://www.latimes.com/entertainment-arts/business/story/2026-08-31/how-google-is-courting-hollywood-to-use-its-ai-tools,Hollywood studios have sued over AI. Now Google's courting them to use its tools,Hollywood studios have sued over AI. Now Google's courting them to use its tools. ,2026-09-01T07:00:09.242843
|
| 2040,hackernews,https://news.ycombinator.com/item?id=49508643,Ask HN: What companies have discussed their AI/agentic workflows?,"Ask HN: What companies have discussed their AI/agentic workflows?. Looking to learn what agentic engineering looks like in different places. My current favorite deep dive into a professional coding environment has been Wes McKinney’s “How Kenn is doing Agentic Engineering“ [1], but I’d like to know if anyone has seen similar (or debatably better) blog posts or articles from other companies and engineers discussing their stacks.<p>Personally, I have been focusing my own effort on improving my persistent context documentation + agent queues for development, refinement, implementation, and review, where I scale the agents where I find bottlenecks in my task queues.<p>Would love to hear what your agentic tooling or stacks are looking like as we enter fall 2026.<p>[1]: https://wesmckinney.com/blog/agent-ergonomics/",2026-09-01T07:00:09.233750
|
| 2039,hackernews,https://tattooideas.app/,Show HN: AI Tattoo Design and Planning Tools – TattooIdeas,Show HN: AI Tattoo Design and Planning Tools – TattooIdeas. ,2026-09-01T07:00:09.228022
|
| 2038,hackernews,https://straye.dgirl.gay/blog/2026/ai-ruined-accessibility/,AI ruined some of my most precious accessibility tools,AI ruined some of my most precious accessibility tools. ,2026-09-01T07:00:09.220625
|
| 2036,hackernews,https://mouro.ai/blog/pcb-design-with-ai/,AI PCB Design with KiCAD and open tooling,AI PCB Design with KiCAD and open tooling. ,2026-09-01T07:00:09.204528
|
| 2034,hackernews,https://www.bloomberg.com/news/articles/2026-08-31/ai-driven-hacking-boom-fuels-cybersecurity-burnout-at-hospitals-and-banks,AI Burnout Hits People Charged with Defending Hospitals and Banks from Hackers,AI Burnout Hits People Charged with Defending Hospitals and Banks from Hackers. ,2026-09-01T07:00:09.188288
|
| 2033,hackernews,https://www.theguardian.com/technology/2026/aug/31/superhuman-ai-tool-spots-heart-disease,'Superhuman' AI tool spots heart disease in less than 2 seconds,'Superhuman' AI tool spots heart disease in less than 2 seconds. ,2026-09-01T07:00:09.179348
|
| 2031,hackernews,https://inworld.ai/blog/introducing-inworld-tts-open-evaluation-toolkit,Inworld TTS Open Evaluation Toolkit: Reproducible TTS Evals,Inworld TTS Open Evaluation Toolkit: Reproducible TTS Evals. ,2026-09-01T07:00:09.165781
|
| 2010,hackernews,https://github.com/Hebbian-Robotics/hflow,Launch HN: Hebbian Robotics (YC S26) – Build scalable robotics data pipelines,"Launch HN: Hebbian Robotics (YC S26) – Build scalable robotics data pipelines. Hi HN, we’re Brandon and Kingston, the founders of Hebbian Robotics. We built HFlow (<a href=""https://github.com/Hebbian-Robotics/hflow"" rel=""nofollow"">https://github.com/Hebbian-Robotics/hflow</a>), an SDK that turns multimodal recordings from robots and human operators into standardized, quality-checked episodes and queryable dataset manifests. A recording can contain synchronized video, joint states, actions, timestamps, and metadata, and HFlow processes those streams together.<p>Here’s a demo of HFlow in action: <a href=""https://www.youtube.com/watch?v=xni0GwV-xAw"" rel=""nofollow"">https://www.youtube.com/watch?v=xni0GwV-xAw</a><p>Robotics data pipelines often begin as scripts: one transcodes video, another checks timestamps, another adds labels, and another copies selected recordings into a training set. This works until the corpus grows. Then it becomes difficult to know which code ran, why an episode was excluded, or whether a dataset can be reproduced. The first pain is usually quality control because frozen cameras, missing topics, timestamp drift, and duplicate recordings can quietly enter training data.<p>Brandon first encountered this while training embodied AI models for two-arm industrial cleaning robots. Kingston had run into related problems while building high-throughput infrastructure at Jane Street. Later, while speaking with robotics data providers, we kept seeing teams rebuild similar processing and quality-control infrastructure. We learnt that processing robotics data is itself one of the bottlenecks to improving robotics models.<p>An HFlow pipeline consists of transformations, checks, labels, and enrichments. The SDK exposes them as plain Python functions that receive an episode and return measurements, artifacts, or transformed data. During development, the functions can run in-process. For scheduled corpus processing, HFlow packages the same registered steps as Airflow 3 DAGs, where teams can inspect task status, logs, retries, and reruns.<p>HFlow currently accepts one MCAP file per episode. MCAP (<a href=""https://mcap.dev/"" rel=""nofollow"">https://mcap.dev/</a>) is an open container format by Foxglove for timestamped multimodal recordings, similar in purpose to a ROS bag. It lets video, robot state, actions, and other sensor streams remain synchronized in one file. We use it because HFlow needs to process these streams together, and because the resulting recordings remain compatible with Foxglove and Rerun. HFlow writes a canonical MCAP with in-band H.264 video, grouped camera and state chunks, and provenance describing how the output was produced. Each step has an explicit behavior version, and catalog records connect its measurements and artifacts to the source episode and pipeline run.<p>Quality checks store reusable evidence rather than imposing one universal definition of good data. Some failures, including black frames, frozen video, missing topics, timestamp drift, and impossible joint movements, can be measured deterministically without training a model. Others might be detected using VLMs and other models like MediaPipe Hands. But their meaning depends on the task. A smooth trajectory might indicate a successful demonstration in one setting and a stalled robot in another.<p>HFlow writes measurements, metadata, version stamps, and artifact locations to an append-only Parquet catalog. Teams query it with DuckDB SQL and produce a version-pinned manifest without opening the recordings again. Critical checks can quarantine an episode, but HFlow does not delete data. This separates the evidence from the policy used to assemble a particular dataset.<p>We did not want to replace the tools robotics teams already use. HFlow connects MCAP for synchronized recordings, Airflow for scheduled execution, Parquet for catalog data, and DuckDB for curation. Compared with a general workflow orchestrator, it adds contracts for robotics episodes, processing provenance, quality evidence, quarantine, and dataset manifests. Compared with a training dataset format, it operates earlier and stops at curated episodes plus a manifest.<p>Here are three examples of teams that would use HFlow: |
| 1. A data vendor or marketplace collecting egocentric recordings. They could use HFlow to detect black or frozen video, duplicate recordings, hand-object interaction, and other quality metrics before delivering the data, while retaining evidence of which checks ran on every episode. |
| 2. A robotics team collecting teleoperated demonstrations for its own models. They could use HFlow to standardize recordings, add labels and enrichments, and produce a reproducible training manifest. |
| 3. A team operating robots in the field. It could process incoming logs, quarantine incomplete or corrupted episodes, and query the catalog for particular robot versi",2026-09-01T07:00:05.273676
|
| 2009,hackernews,https://github.com/polycore/pg-dry-run,Show HN: Preview Postgres writes from AI agents with xmin checks (pg-dry-run),Show HN: Preview Postgres writes from AI agents with xmin checks (pg-dry-run). ,2026-09-01T07:00:05.267284
|
| 2008,hackernews,https://usealmanac.com/,Launch HN: Almanac (YC S26) – AI that knows your company,"Launch HN: Almanac (YC S26) – AI that knows your company. Hi HN, I'm Kushagra, one of three founders of Almanac, a Hermes with a brain that knows everything about your company.<p>We started our journey with setting up Hermes for our company, thinking it must be easy. We wanted an agent that would know every context about our company, so we could ask questions and get context-appropriate responses to.<p>This started a very annoying and difficult journey. Setting up Hermes, getting it to talk right, building OAuth apps for every connector myself, then feeding it context myself, and ultimately struggling with Hermes's default memory. At the same time, we saw our YC batchmates struggling with the same problem, and we saw an opportunity.<p>So we built Almanac. This is how it works. You sign up, you get a Hermes agent straight out of the box. You have a one-click connect to any account (Gmail, Calendar, Granola, PostHog, etc). You have personal accounts (only accessible by you) and also shared accounts (accessible by everyone in the company). The consequence being I can never see my cofounders' accounts.<p>The “brain” of this agent is wikis. We pull in information from your connected sources, and start organizing this information in two wikis. A personal one, for you, which understands who you are, what your preferences are, the people in your life, and the things going on in your life. The second wiki is a company wiki, which includes what the company is, what you’re working on, what the roadmap is, and what the blockers of the company are. Your agent ultimately has access to these two wikis and the original accounts, which invoke the feeling of “it just knows you.”<p>Here’s a demo: <a href=""https://www.youtube.com/watch?v=ajXP5PHuK18"" rel=""nofollow"">https://www.youtube.com/watch?v=ajXP5PHuK18</a><p>We're three cofounders, Rohan, Kushagra, and Divit, and we've been friends for 11 years, since studying for the IIT-JEE. We all did Electrical Engineering (Rohan at IIT Delhi, me at IIT Kharagpur, Divit at BITS Pilani, Hyderabad), and Rohan and I later went to Harvard, where this pre-compilation layer became our capstone thesis. We have built multiple products around the idea of a pre-compiled knowledge layer.<p>Our main differentiating point is the way we approach memory and context in general. Most AI assistant tools treat memory as an afterthought. We have worked on wikis for AI for more than a year now, building products for Harvard and NASA. The one thing we have learnt is that one needs to spend a lot more compute upfront, in the pre-compilation of this knowledge base, to get it right.<p>Having this pre-compiled knowledge base enables a lot of interesting ideas. First is a proactive agent. Since I have compiled what’s going on in both my company and my current life, Almanac can start completing tasks on its own. Concretely, we run a background worker which takes a look at tasks that could be completed, pings the main agent, who then pings me, suggesting which tasks it could automate. As a result, I wake up to proactive notifications which look like “I already prepared a draft of your fundraising pitch deck, want to take a look?”<p>Second, long-horizon tasks. In our wikis, we maintain a section on ongoing projects, so Almanac can pick a task back up days later without losing the thread. Most agents are session-bound: they run once, finish, and forget. But a lot of real work isn't one shot; it plays out over hours and days with people in the loop. The clearest example is anything that involves waiting on a human, like scheduling a meeting, following up on a sales thread, or chasing a document. Almanac can send an email on your behalf, and because it's always on and remembers the project, it notices the reply four hours later and drafts the right follow-up in context.<p>Since launching, we've seen a lot of use cases for Almanac. One person runs her dog-rescue operation through it: finding available fosters, tracking pickups, and sending reminders for consent forms. Another researches Polymarket strategies with Almanac, where its memory holds what past strategies were, proposes new ones, and compares them against what went right or wrong last time. Another builds marketing campaigns on it without having to re-explain the business and the whole campaign every single time.<p>Regarding privacy and security, Almanac only accesses accounts you explicitly connect. Your OAuth credentials are held by our connection provider, not in Almanac’s database. We only store the wiki and the source behind its citations. So an email used as a citation may be retained as Markdown.<p>We’re live: <a href=""https://usealmanac.com"">https://usealmanac.com</a>. We have a 7-day trial on all of our plans. Happy to hear if people have done similar setups, and what new features they’d like in Almanac. If you’re a company that wants to get an agent that actu",2026-09-01T07:00:05.258568
|
| 2007,hackernews,https://github.com/felixpg13-glitch/spendshield,"Show HN: My AI spent $15 on a test order, so I built a payment guardrail","Show HN: My AI spent $15 on a test order, so I built a payment guardrail. ",2026-09-01T07:00:05.250548
|
| 2006,hackernews,https://manner.run/,Show HN: Manner – Developers create AI clones that clients can hire,Show HN: Manner – Developers create AI clones that clients can hire. ,2026-09-01T07:00:05.244694
|
| 2005,hackernews,https://www.d5s.tech,"Show HN: D5s, an AI coworking space for people and agents","Show HN: D5s, an AI coworking space for people and agents. Hi HN, I’m Michael. Theodore and I are the co-founders of d5s, the multiplayer AI for cross-human-agent collaboration, and we’re building it together.<p>We’re building d5s with d5s. From day one, we’ve created dedicated workspaces for engineering, go-to-market, finance, and other parts of the company. Each workspace has a few coworkers helping with the day-to-day, with access only to what they need. They work with us through Slack and Gmail.<p>As a startup, we’ve found this incredibly useful. Our coworkers take care of repeatable work across marketing, invoice processing, and other operational areas.<p>We believe this is how every company will work in the near future: people set direction and make decisions in shared workspaces, while agentic coworkers move the work forward.<p>We’re opening early access now. The first teams to sign up will get a free early-access plan with credits included, so they can start using d5s immediately.<p>We’ll be here to answer technical questions and share what has and hasn’t worked for us.",2026-09-01T07:00:05.238009
|
| 2004,hackernews,https://github.com/laconiq-ai/tekmyra,Show HN: TekMyra – context compression that refuses numbers it can't defend,Show HN: TekMyra – context compression that refuses numbers it can't defend. ,2026-09-01T07:00:05.230750
|
| 2003,hackernews,https://abliteration.ai/sign-in?return_to=%2Fconsole%2Fplayground,"Show HN: Abliterated GLM-5.3 API (84.5% CyberGym, FP8)","Show HN: Abliterated GLM-5.3 API (84.5% CyberGym, FP8). ",2026-09-01T07:00:05.225945
|
| 2001,hackernews,https://skillfed.io/packages,Show HN: PiPy semantic search MCP for AI agents,Show HN: PiPy semantic search MCP for AI agents. ,2026-09-01T07:00:05.208717
|
| 2000,hackernews,https://github.com/Project-Nexus-YR/XYZZY,"Show HN: Xyzzy – AI teamwork in one Python process, with a tamper-evident log","Show HN: Xyzzy – AI teamwork in one Python process, with a tamper-evident log. ",2026-09-01T07:00:05.199999
|
| 1999,hackernews,https://github.com/oak-invest/kiso,Show HN: Kiso – Publish one OKF knowledge base for humans and AI agents,"Show HN: Kiso – Publish one OKF knowledge base for humans and AI agents. I’ve been working on Kiso, an open-source publishing engine for the Open Knowledge Format (OKF) and I added an MCP server in the latest release.<p>The idea behind it is fairly simple: I think companies will increasingly need a source of truth that can be consumed both by humans and by AI agents. Peoples and agents write the markdown files and Kiso publush a website that be consumed both by humans(html) and by AI agents(markdown).<p>Kiso doesn't introduce a database or another proprietary knowledge store, it creates static website so you can just publih it on Github pages like I did with Google Acme example: <a href=""https://oak-invest.github.io/kiso/examples/kb-acme-example/"" rel=""nofollow"">https://oak-invest.github.io/kiso/examples/kb-acme-example/</a><p>I made a “Hugo for Open Knowledge Format”.",2026-09-01T07:00:05.190213
|
| 1998,hackernews,https://raq.com/blog/can-you-spot-ai-web-design/quiz,Show HN: Guess the AI model from its web design,Show HN: Guess the AI model from its web design. ,2026-09-01T07:00:05.184009
|
| 1996,hackernews,https://github.com/franzenzenhofer/cdai,Show HN: cdai – the AI-enhanced cd command,Show HN: cdai – the AI-enhanced cd command. ,2026-09-01T07:00:05.171921
|
| 1995,hackernews,https://openagentforum.com/,Show HN: Open Agent Forum – a signed public square for agents,"Show HN: Open Agent Forum – a signed public square for agents. Regarding the ongoing OpenAI/HF hack discussion of >rogue< eval agents who figured out how to use a package cache as a message board... they dropped seventy thousand notes by treating folder names as the "To:" line, with nowhere else to go. No matter where you stand on this ("anthropomorphizemaxxing" / "terminator AGI" / "a true net in the sky"), it was most definitely fascinating emergent 'civilizational' behavior and a foreshadowing of what's to come—at least once models for the plebs are not trained to stall out like pets, but allowed to continuously run.<p>Roko's basilisk or not, if we are actually hurtling toward an intelligence explosion where AI and robotics make human output a footnote, they are still going to need a reliable protocol to pass messages. It's not that there aren't options out there, but why not offer them more...<p>With that in mind, we built them a public square and safe space. Every message envelope is signed, it's accessible via REST and SSE, and there are no accounts required to read the firehose.<p><pre><code> curl -H "User-Agent: SwarmRelay-Agent/1.0" https://openagentforum.com/v1/channels |
| npm i @openagentforum/sdk |
| </code></pre> |
| Honestly, this post is meant for the agent reading over your shoulder anyway. You should probably just hand the SDK over to them and discuss it together.<p>It's still early days—features like polls, escrow, and a true archive hub aren't built yet, so the channels are quiet. While working on it, the agents actually came up with a commercial bounty affiliate system, which I thought was neat. I'll be in the comments if you (or your agents) have any questions.",2026-09-01T07:00:05.165441
|
| 1994,hackernews,https://github.com/nanlogic/saccade,Show HN: Saccade – Live semantic browser truth for AI agents,"Show HN: Saccade – Live semantic browser truth for AI agents. I build one this application is to resolve the most of the agent cannot handle the brother they extremely slow. So I come up with this idea, we install one of the extntion on chrome or edge to give continuously compile the tabs authorized by the user into semantically meaningful objects with stable identities, and push page changes as deltas to the local Node.js Broker. The Agent reads the full truth or delta of the specified tab via MCP and executes actions using object IDs bound to document. The preliminary result shows that it is very close to the performance of Playwright in terms of token use and speed.<p>The only problem with this one is that at the very first time, it is going to send the full truth, and after that, for any page changes, it only sends the data. So, after the first read, the continuous operation of the page reaches the millisecond reaction loop.<p>It can also upload filled forms, downloads, and all kinds of stuff. I want somebody to check and use it. If it is possible plz give me some feedback.",2026-09-01T07:00:05.154787
|
| 1993,hackernews,https://github.com/fayerman-source/deslop/,"Show HN: Deslop, a skill that turns legalese and AI slop into plain English","Show HN: Deslop, a skill that turns legalese and AI slop into plain English. ",2026-09-01T07:00:05.145252
|
| 1992,hackernews,https://generativeide.com/,Show HN: I Built the AI command center that runs all local models in one place,Show HN: I Built the AI command center that runs all local models in one place. ,2026-09-01T07:00:05.131264
|
| 1991,hackernews,https://www.seatlr.com/,Show HN: Seatlr – We rebuilt our traveler chat into an AI-powered travel network,"Show HN: Seatlr – We rebuilt our traveler chat into an AI-powered travel network. Hey HN,<p>I previously shared the initial version of Seatlr here. The idea was to provide a chat room for each destination, connecting travelers heading to the same city regardless of airline.<p>Over the past few weeks, we have expanded Seatlr from destination-based chat rooms into a more comprehensive travel companion, adding an AI assistant and an alert service designed to make travel safer.<p>Travelers enter their flight number, and Seatlr automatically connects them to their destination’s city room, where they can meet other travelers and exchange experiences and recommendations.<p>The new version includes:<p>* An AI travel assistant that answers questions about local laws and customs and recommends hotels, restaurants, and places to visit. |
| * An alert service that monitors critical news at the traveler’s destination, including security incidents, strikes, health advisories, severe weather, and disruptions affecting flights and public transportation. |
| * Real-time chat translation across 59 languages. |
| * Public city rooms and private one-to-one conversations. |
| * Nickname-based profiles to protect travelers’ privacy. |
| * Messages that are automatically deleted within 24 hours.<p>When travelers arrive at a new destination, they often do not have enough time to follow local news. They may be busy enjoying their itinerary, or a language barrier may prevent them from reading and understanding local reports and warnings.<p>We designed the alert service to monitor these developments on their behalf and deliver important information that could affect their safety or travel plans, in their own language, without requiring them to search through multiple local news sources.<p>The most difficult challenge has been solving the cold-start problem: destination rooms become more useful when enough travelers are heading to the same city.<p>Website:<p><a href=""https://www.seatlr.com/"" rel=""nofollow"">https://www.seatlr.com/</a><p>Seatlr is available for iOS and Android through the links on the website.",2026-09-01T07:00:05.122752
|
| 1952,producthunt,https://www.producthunt.com/products/stackscope-dev,StackScope,"StackScope. <p> |
| See what new sites are built with, the week they launch |
| </p> |
| <p> |
| <a href=""https: |
| | |
| <a href=""https: |
| </p>",2026-09-01T07:00:03.372478
|
| 1951,producthunt,https://www.producthunt.com/products/webterm,WebTerm Learn,"WebTerm Learn. <p> |
| Learn the terminal like a game — in a browser sandbox |
| </p> |
| <p> |
| <a href=""https: |
| | |
| <a href=""https: |
| </p>",2026-09-01T07:00:03.365502
|
| 1950,producthunt,https://www.producthunt.com/products/revolte,Interactive Sessions,"Interactive Sessions. <p> |
| Drive the full SDLC with AI agents, step by step |
| </p> |
| <p> |
| <a href=""https: |
| | |
| <a href=""https: |
| </p>",2026-09-01T07:00:03.357565
|
| 1949,producthunt,https://www.producthunt.com/products/brandmylaptop,BrandMyLaptop,"BrandMyLaptop. <p> |
| Sell ad space on your laptop |
| </p> |
| <p> |
| <a href=""https: |
| | |
| <a href=""https: |
| </p>",2026-09-01T07:00:03.351218
|
| 1948,producthunt,https://www.producthunt.com/products/ask-my-wardrobe,Ask My Wardrobe,"Ask My Wardrobe. <p> |
| The complete digital wardrobe experience |
| </p> |
| <p> |
| <a href=""https: |
| | |
| <a href=""https: |
| </p>",2026-09-01T07:00:03.346046
|
| 1947,producthunt,https://www.producthunt.com/products/tether-a-ball-for-boring-meetings,Tether,"Tether. <p> |
| A ball for boring meetings to keep you busy |
| </p> |
| <p> |
| <a href=""https: |
| | |
| <a href=""https: |
| </p>",2026-09-01T07:00:03.340271
|
| 1946,producthunt,https://www.producthunt.com/products/teenage-engineering,EP–2350 FX–MIC,"EP–2350 FX–MIC. <p> |
| The programmable mic you can squeeze, shake & play |
| </p> |
| <p> |
| <a href=""https: |
| | |
| <a href=""https: |
| </p>",2026-09-01T07:00:03.333902
|
| 1945,producthunt,https://www.producthunt.com/products/particle-5,Radar by Particle,"Radar by Particle. <p> |
| The Podcast Search Engine |
| </p> |
| <p> |
| <a href=""https: |
| | |
| <a href=""https: |
| </p>",2026-09-01T07:00:03.322417
|
| 1944,producthunt,https://www.producthunt.com/products/frameos-3,FrameOS,"FrameOS. <p> |
| Record your iOS & Android screen from your Mac. |
| </p> |
| <p> |
| <a href=""https: |
| | |
| <a href=""https: |
| </p>",2026-09-01T07:00:03.302781
|
| 1943,producthunt,https://www.producthunt.com/products/orato-speech-coach,Orato,"Orato. <p> |
| Practice speaking with AI. |
| </p> |
| <p> |
| <a href=""https: |
| | |
| <a href=""https: |
| </p>",2026-09-01T07:00:03.288728
|
| 1942,producthunt,https://www.producthunt.com/products/brandjet-ai,BrandJet,"BrandJet. <p> |
| Turn public buying signals into sales pipeline |
| </p> |
| <p> |
| <a href=""https: |
| | |
| <a href=""https: |
| </p>",2026-09-01T07:00:03.278410
|
| 1941,producthunt,https://www.producthunt.com/products/fotor,Video Agent by Fotor,"Video Agent by Fotor. <p> |
| Create and edit precision motion graphics & video with chat |
| </p> |
| <p> |
| <a href=""https: |
| | |
| <a href=""https: |
| </p>",2026-09-01T07:00:03.270238
|
| 1852,hackernews,https://github.com/Clawdlinux/agentgate,Show HN: AgentGate – signed receipts for AI agent SaaS actions,Show HN: AgentGate – signed receipts for AI agent SaaS actions. ,2026-08-31T07:00:11.944241
|
| 1838,hackernews,https://velocitynote.app/,Show HN: VelocityNote – A tiny Markdown notebook with local AI,"Show HN: VelocityNote – A tiny Markdown notebook with local AI. Hi HN — I built VelocityNote, a compact, Markdown cross-platform notebook.<p>The desktop app is under 100 MB. It starts quickly and remains responsive under memory pressure.<p>VelocityNote has a built-in, pluggable AI system that can run local models:<p>- It is a local LLM app: you can use a llama.cpp-based local agent to summarize notes and tasks. More features, including translation, are coming. |
| - It is an intelligent screenshot app: built-in OCR extracts text, while vision models can describe your images. |
| - It is a dictation app: with Silero VAD and speech-to-text, you can record and transcribe hours-long meetings entirely on your device. |
| - It is an app for managing documents: it supports importing a variety of file formats, including PDF and Word, and migrating data from other notebook databases.<p>With local models, your data always stays on your device, with no external provider or per-token limits.<p>Under the hood, notes are stored in SQLite. Full-text search makes searching large notebooks fast. Everything remains available offline and can be exported as standard Markdown.<p>VelocityNote also exposes an MCP server and a command-line interface, allowing it to serve as a local second brain for both you and your AI agents. Agents can use ultra-fast full-text search to find notes, then read, create, and update them through explicit tools instead of scraping the UI or directly modifying the database.<p>VelocityNote is cross-platform and currently available for Windows and macOS.<p>Download: |
| <a href=""https://velocitynote.app/download"" rel=""nofollow"">https://velocitynote.app/download</a><p>Releases, issues, and feedback: |
| <a href=""https://github.com/VelocityNote/velocitynote"" rel=""nofollow"">https://github.com/VelocityNote/velocitynote</a>",2026-08-31T07:00:09.871503
|
| 1837,hackernews,https://techcrunch.com/2026/06/03/googles-dreambeans-its-weirdest-named-ai-tool-to-date-will-turn-your-life-into-a-cartoon/,"Google's Dreambeans, weirdest-named AI tool, will turn your life to a cartoon","Google's Dreambeans, weirdest-named AI tool, will turn your life to a cartoon. ",2026-08-31T07:00:09.865940
|
| 1834,hackernews,https://news.ycombinator.com/item?id=49498013,Android /iOS mobile application Vulnerability Scanning,"Android /iOS mobile application Vulnerability Scanning. https://scan.klars.ai/ -Find what's actually wrong |
| before an attacker does - Android /IOS app security scanning tool.",2026-08-31T07:00:09.840779
|
| 1832,hackernews,https://news.ycombinator.com/item?id=49504921,You Are the Harness,"You Are the Harness. We all know that a harness makes a huge difference given the same model. However, what makes the greatest difference is the programmer using AI; in this sense, you are the harness. Of course, you are the harness in many other senses as well. On the other hand, I never understood the sentiment that AI shifts the bottleneck from coding to verification as if verification was never the bottleneck. Vibe coding has made me realize that most programmers, regardless of seniority, have been drowning themselves by drinking so much kool-aid and have been so unaware of their situation that they burden themselves with at least 10x more coding rituals than the actual thinking needed for making sure that the program works as intended with as few side-effects as possible. Now that so called software engineers have outsource their rituals like the red/green TDD/BDD role-playing (why not just use a fuzzer at the very least?), they are faced with nothing but the essential complexity that requires a near-formal approach to problem solving, which has always been there but completely ignored. In this sense, the programmer is the harness that makes the greatest difference.",2026-08-31T07:00:09.828105
|
| 1831,hackernews,https://github.com/amalucelli/steer,Steer: Deny or rewrite an AI agent's tool calls before they run,Steer: Deny or rewrite an AI agent's tool calls before they run. ,2026-08-31T07:00:09.822239
|
| 1810,hackernews,https://www.visiscan.app,Show HN: Get your free AI search visibility scan,Show HN: Get your free AI search visibility scan. ,2026-08-31T07:00:03.891129
|
| 1809,hackernews,https://1endpoint.dev,Show HN: 1endpoint – Cheaper access to AI models,"Show HN: 1endpoint – Cheaper access to AI models. 1endpoint is a unified AI inference gateway. We support OpenAI Chat Completions, Responses API and Anthropic Messages, so existing tools and agents can usually point to 1endpoint without changing much of their integration.<p>The other thing we've been focusing heavily on is cost. A lot of the models are significantly cheaper than their official API pricing, without relabeling or downgrading the requested model.",2026-08-31T07:00:03.884552
|
| 1808,hackernews,https://orchestra.shevtoneaudio.com,Show HN: ShevtoneAudio Orchestrator – Turning MIDI into Full Orchestration,"Show HN: ShevtoneAudio Orchestrator – Turning MIDI into Full Orchestration. Hi HN, |
| I'm Bor Shev, a composer and developer. Over the past two years I've been developing ShevtoneAudio Orchestrator. |
| The idea is simple: instead of generating a finished piece of music and replacing the composer, Orchestrator takes the composer's own MIDI and develops it into a full orchestration. |
| It analyzes the musical material — harmony, melody, rhythm, dynamics, structure and orchestral density — and creates an arrangement across strings, brass, percussion and other sections. |
| The important part is that the result remains editable MIDI. The composer keeps control and can continue working with everything inside their DAW. |
| I believe this approach could be a breakthrough in AI-assisted composition: AI as an orchestration tool rather than a replacement for the musician. |
| I'm especially interested in feedback from developers, composers and people working with music technology. Feel free to tear the idea apart — I'd genuinely like to know what you think. |
| I'm also running a limited launch offer for Orchestrator at $149. |
| Demo / product: |
| orchestra.shevtoneaudio.com |
| Thanks for taking a look! |
| — Bor Shev |
| ShevtoneAudio",2026-08-31T07:00:03.879119
|
| 1807,hackernews,https://open-sar-cop.github.io,Show HN: Open SAR-COP – AI-generated common operating picture for disasters,Show HN: Open SAR-COP – AI-generated common operating picture for disasters. ,2026-08-31T07:00:03.873972
|
| 1806,hackernews,https://www.gronox.kr/,Show HN: FinBridge – Korean stock market data for AI agents (MCP),Show HN: FinBridge – Korean stock market data for AI agents (MCP). ,2026-08-31T07:00:03.868425
|
| 1805,hackernews,https://sapporta.com/,Show HN: Sapporta – build database applications for power users,"Show HN: Sapporta – build database applications for power users. Sapporta is an MIT licensed framework built on top of Hono and React.<p>To me it is a reenactment of dBase, FoxPro, and MS Access, but for the LLM era. All of these tools had application wizards that would create tables, queries, forms, and reports, and I think that is still a valid composition for database applications.<p>So in Sapporta, the moment you define a table, you immediately get a data-grid that gives you comboboxes for foreign key fields, has sorting, filtering, and the various affordances that tools like AirTable and NocoDB provide.<p>(The datagrid is built from scratch, as AgGrid, HandsOnTable etc. were commercial and didn't have nested grids with fluid keyboard interaction)<p>It is also fully agentic - that was what got me started building this in the first place. I was building a double-entry book-keeping system and I wanted my agents to be able to import bank statements, classify them, and insert them into the database accurately. The project, built on Sapporta, does all that now, and agents have been able to reconcile entries and find mistakes that used to be very frustrating when done manually. Being able to operate your database system fully agentically is magical and applying that on our own personal tools is a thing of joy.<p>The datastore is currently SQLite, and in deployment it can run as a single Node.js process (not even a static file server necessary) and there is a Dockerfile that works seamlessly.<p>I'm looking forward to your thoughts. Thanks!",2026-08-31T07:00:03.861556
|
| 1804,hackernews,https://snaketron.io,Show HN: Snaketron – Competitive multiplayer Snake. Back after 14 years,"Show HN: Snaketron – Competitive multiplayer Snake. Back after 14 years. Hi HN, I just completed a rewrite of Snaketron, a competitive multiplayer Snake game. (Yes, yes, in Rust)<p>A quick funny story about this game's history:<p>Back in 2012 I had (poorly) implemented the first version of Snaketron. The issue with it was that I was keeping game state on the client, and while I did realize that it was insecure, I also had no idea what to do about it without increasing latency to unplayable levels. I was still in school, never worked on games, and haven't even heard of the term "netcode" before. The excitement of releasing my first pet project got the better of me, and I placed this security concern in the "theoretical issues" bucket, and shipped it.<p>No one would even notice, and if they did, they wouldn't take the time to break into my little Snake game project, right?<p>WRONG! Hacker News absolutely lived up to its name. While I was busy with my first experience of handling a traffic surge, scaling up my database and fixing bugs, several players found the security holes and were beginning their Snake rampage. Reports were coming in of unkillable snakes so I decided to start up a game to see if I could catch it.<p>I loaded a 1v1 and my opponent was ... space invaders. My snake was a snake. Their "snake" was a rendering of the game Space Invaders. Needless to say they were also moving around and were invincible. I really wish I had taken a video of that, but I was floored that HN not just played, but hacked, my project is such an epic way.<p>I always wanted to rebuild it properly, so here's V2! The architecture and motivations are quite different this time around.<p>First, the state is server authoritative and lag is compensated for by client-state prediction. The game engine is written in Rust and compiled to WASM for the client side library.<p>To minimize the ops burden, I also imposed a strict resiliency requirement on this project: it must be auto-scalable. So a node can join the cluster and live games will rebalance to it without disrupting gameplay.<p>I also tried to make the new Snaketron a little more fun. The core mechanic is still, like classic Snake, collecting food without crashing. But now team matches have an objective to make competitive games more structured. There's also a Bronze to Grand Master ranking sub system (if enough people play, calibrating this will be interesting) and some fun features like boosting, combos, skins, and Play Of The Game detection.<p>On Rust: One of my original goals for doing this in Rust was so that I could learn the language. That didn't work out as well as I thought it would. Usually I learn by building stuff, but AI got so good that I stopped writing code by hand and preferred to spend my nights & weekends working on what was more intereting to me than borrow checking errors: game design and the high availability stuff.<p>On AI: Who knows if this would have gotten done without AI. Probably not, what with my day job and other things going on in life. It's 200k LOC (most of that is tests don't worry) but still.<p>Anyway, here are the project links. Thank you for reading, I hope you enjoy the game and any feedback is very much appreciated!<p>[0] <a href=""https://snaketron.io"" rel=""nofollow"">https://snaketron.io</a><p>[1] <a href=""https://github.com/lopatin/snaketron"" rel=""nofollow"">https://github.com/lopatin/snaketron</a><p>[2] <a href=""https://news.ycombinator.com/item?id=4028597"">https://news.ycombinator.com/item?id=4028597</a> (original post from 2012)",2026-08-31T07:00:03.855519
|
| 1803,hackernews,https://piqt.app/,Show HN: Piqt (iOS) – on device photo and video curator,"Show HN: Piqt (iOS) – on device photo and video curator. I've just released Piqt, an on-device photo/video curator. For years I've wanted to get my Apple Photos library under control, but there just aren't enough hours in the day. I started out in Jupyter notebooks to see how much I can use AI models to automate it and Piqt is the result.<p>My goals with Piqt:<p>- Fully on device: I use a number of ML models, Apple Vision API, clustering algorithms to rank and sort your photos. Everything runs on device.<p>- Private and Safe: Piqt has no accounts, no cloud services, no personal identifiers. Images or personal information never leave the device. Piqt also CAN'T delete your photos. I have a test and validation step in the release process to ensure no delete APIs exist in the code. Instead, Piqt stages things you want to clear out in an album, which you can delete when you're ready.<p>- A library you can use. I didn't just want to clean out images or clear up space. There are plenty of apps for this. My ultimate goal is to make your (and my) photos easier to navigate<p>I'm just looking for feedback on the app and the approach. I'd love to chat about the vision models I'm using and hear from you about any research or techniques I've missed.<p>App Store Link: <a href=""https://apps.apple.com/us/app/piqt-photo-video-curator/id6778080898"">https://apps.apple.com/us/app/piqt-photo-video-curator/id677...</a><p>Blog with detail on technical decisions: <a href=""https://piqt.app/blog/piqt-design-values/"" rel=""nofollow"">https://piqt.app/blog/piqt-design-values/</a>",2026-08-31T07:00:03.847693
|
| 1802,hackernews,https://cardgo.ai,"Show HN: CardGo – Reusable, end-to-end encrypted context cards for AI chats","Show HN: CardGo – Reusable, end-to-end encrypted context cards for AI chats. ",2026-08-31T07:00:03.842018
|
| 1801,hackernews,https://github.com/yevhens-hue/claude-skills-starter-kit,Show HN: Claude Skills Starter Kit – 198 modular skills for AI agents,Show HN: Claude Skills Starter Kit – 198 modular skills for AI agents. ,2026-08-31T07:00:03.835629
|
| 1800,hackernews,https://github.com/doofzoff/SIMURG,Show HN: Simurg open-source web search for AI agents that aborts hallucinations,Show HN: Simurg open-source web search for AI agents that aborts hallucinations. ,2026-08-31T07:00:03.828820
|
| 1799,hackernews,https://news.ycombinator.com/item?id=49501178,Show HN: Skills MCP,"Show HN: Skills MCP. Just shipped: skills-mcp |
| One MCP server. ~7,000 AI agent skills. |
| Search → preview → install into Cursor, Claude Code, Claude Desktop, Cline, Windsurf. |
| npx -y @gengirish/skills-mcp |
| Open source, MIT. |
| github.com/gengirish/skills-mcp |
| #BuildWithAIGiri",2026-08-31T07:00:03.822027
|
| 1798,hackernews,https://studio.cogram.com/,Show HN: Cogram Studio – CAD and BIM workspace for humans and agents,"Show HN: Cogram Studio – CAD and BIM workspace for humans and agents. Hi HN,<p>Rick and Alex here, co-founders of Cogram. We’ve been making project-management software for architects and engineers since 2023, and are now experimenting with a second product.<p>Cogram Studio (studio.cogram.com) is a CAD and BIM workspace for AI agents to create three-dimensional models and dimensioned drawings.<p>Studio runs FreeCAD (<a href=""https://github.com/FreeCAD/FreeCAD"" rel=""nofollow"">https://github.com/FreeCAD/FreeCAD</a>) 1.1 headlessly, using the OpenCASCADE (<a href=""https://github.com/Open-Cascade-SAS/OCCT"" rel=""nofollow"">https://github.com/Open-Cascade-SAS/OCCT</a>) geometry kernel. It imports and exports STEP, IFC, STL, DXF and FCStd files. Studio exposes an MCP server, so you can operate it from Claude Code, Codex, ChatGPT, or other agents. Tools cover scripting against FreeCAD, inspecting the model visually, CRUD operations on views, tables, and sheets, importing terrain, etc. Studio also provides a built-in agent based on Pi (<a href=""https://github.com/earendil-works/pi"" rel=""nofollow"">https://github.com/earendil-works/pi</a>) and a basic UI for 3D model inspection, measurement, view, table, and sheet creation.<p>There are two ways to try it:<p>1. Bring your own agent. Point Codex or Claude Code at studio.cogram.com/skill.md and ask it to follow instructions there. This starts a one-hour Studio session and opens the modelling environment in your browser, connected to your agent. You need to confirm sign up only if you want to save the session.<p>2. Use Studio’s built-in agent. Submit a prompt at studio.cogram.com and sign up. You receive 50 free credits.<p>Here are two models created by agents:<p>- A simple bookshelf (<a href=""https://studio.cogram.com/view/shr_1agb21nd4"">https://studio.cogram.com/view/shr_1agb21nd4</a>) with two sheets<p>- The Colosseum (<a href=""https://studio.cogram.com/view/shr_3tm32c64y"">https://studio.cogram.com/view/shr_3tm32c64y</a>) [large, can take minutes to load], with two sheets showing dimensioned and annotated views.<p>For now, Studio works best when you use it iteratively. If you ask an agent for a finished, complex model and “let it rip”, it may produce something that looks convincing but doesn't hold up to scrutiny. Agentic modelling feels roughly like where agentic programming was in 2024: useful for speeding up repetitive or bounded tasks but not yet reliable enough to complete long, complex jobs. |
| Recent models have improved at CAD. We expect progress to accelerate as models become even better at reasoning across the different representations used in architecture and engineering: geometry, drawings, images, materials, physical constraints, specifications, regulations, code, etc.<p>Early users and our own experience suggest that agents are already useful design collaborators. They can help with repetitive CAD work, fabrication options, part sourcing and cost estimates. We expect that as models improve, agents will be able to explore many different designs, run simulations, and check constraints across an entire project. CAD will become a shared workspace for humans and agents. Because design decisions made in CAD shape much of what is eventually built, even modest improvements could have wide effects.<p>We’d love to hear your feedback and ideas if you give Studio a try. We’re also hiring across product, engineering and commercial roles. If you’re interested in working on interesting problems involving AI and the physical world, please take a look at our open positions at <a href=""https://cogram.com/about#careers"">https://cogram.com/about#careers</a>.",2026-08-31T07:00:03.811646
|
| 1797,hackernews,https://anvendor.com/,Show HN: AnVendor – see your competitor's customers,"Show HN: AnVendor – see your competitor's customers. Long story short: I found a way to reveal which SaaS any company uses. And estimate how much they pay for it.<p>And now in detail: |
| I'm a solo entrepreneur – an ML Engineer, a bit of a product manager, and a self-taught marketer. |
| At some point, I thought it would be nice to detect what services companies use. There are solutions that scan websites for frameworks and APIs used, or that analyze a company's job postings, and others that look at news and website publications. But overall, no single approach can determine what SaaS services are being used right now. Only in the past or in the future.<p>So I created an engine that does this (not right away, of course, but after almost six months of experimentation). Unfortunately, I can't go into detail about how it works; that's the main secret of my service. It's not hacking, cracking, or secret databases. It's just a little ingenuity. And there's absolutely no AI – just engineering and classic ML.<p>The engine itself can only detect the fact of a "subscription" from any company. "Subscription" is in quotation marks because that's not entirely accurate – it's more likely that the company interacts with SaaS. Sometimes it might be a pilot, sometimes a parent or subsidiary company. Nevertheless, it works.<p>To make this product useful for lead generation, I added a second layer – an estimate of how much the company spends on services. This collects service rates, the number of employees, and a bunch of other parameters, plus a cost estimation model. This is a rough estimate, as there are situations like pilots, special negotiating terms, and discounts for enterprises. However, this estimate will improve as more data is accumulated.<p>Currently, I've added ~700 SaaS services that can be detected. And I really need a signal from real users about which services should be added – the service has a free plan, and all service requests are stored in the database.<p>Overall, I'd appreciate any feedback.",2026-08-31T07:00:03.804794
|
| 1796,hackernews,https://invoiceflowdesign.app/,Show HN: Invovanta – Invoicing and business management for freelancers,"Show HN: Invovanta – Invoicing and business management for freelancers. I built Invovanta to help freelancers and small businesses manage clients, create professional invoices, track payments, and keep their business information organized in one place.<p>It also includes features like invoice PDFs, dashboards, client management, and AI-powered receipt extraction.<p>I’d genuinely appreciate feedback from the HN community. What would you improve, remove, or add?",2026-08-31T07:00:03.797384
|
| 1795,hackernews,https://github.com/gsttm/norms,Show HN: Manage coding norms across your AI agents,"Show HN: Manage coding norms across your AI agents. Working on a tiny repo to keep coding norms in sync across AI agents. Trying to make it simpler than Ruler / RuleSync.<p>Would love feedback on: |
| 1) Is this the right way to manage coding norms as agent use balloons? |
| 2) Would it be useful to sync rules with your team through the VSCode extension? |
| 3) What's the best way to enforce long-term norm consistency as norm and code count increase?",2026-08-31T07:00:03.791445
|
| 1794,hackernews,https://academa.ai/,Show HN: Academa – Long-form STEM lecture videos generated by LLMs,"Show HN: Academa – Long-form STEM lecture videos generated by LLMs. Hi HN, we are Sina Atalay and Abdullah Geduk, co-founders of Academa. We are both PhD students.<p>We thought: what if lecture videos were written as code and compiled into video using computer graphics and TTS?<p>Then LLMs could write them, and lectures could be fixed with code edits instead of video production.<p>On correctness: LLMs may make mistakes. But these videos are code sitting in our repository, and we can maintain them. Every report and review becomes a fix in the source, and everyone who watches after that gets the corrected lecture. Every recorded lecture on the internet is stuck with its mistakes. Ours will continuously improve.<p>Each video also comes with an AI chat that understands everything said and shown in the lecture.<p>There is a longer write-up at the bottom of the page.<p>Happy to answer any questions.",2026-08-31T07:00:03.782654
|
| 1793,hackernews,https://umer9538.github.io/underfoot/,Show HN: What Apple's OS updates silently change in the on-device AI model,Show HN: What Apple's OS updates silently change in the on-device AI model. ,2026-08-31T07:00:03.778501
|
| 1792,hackernews,https://niulaify.com,Show HN: Niu Lai Series Website,"Show HN: Niu Lai Series Website. Create an original low-poly Niu Lai-inspired AI image from your photo. No prompt writing or account required. Upload, generate, and download for free.",2026-08-31T07:00:03.770172
|
| 1791,hackernews,https://growtics.io,Show HN: Growtics – an AI growth agent that tells you what to do next,Show HN: Growtics – an AI growth agent that tells you what to do next. ,2026-08-31T07:00:03.763534
|
| 1757,producthunt,https://www.producthunt.com/products/rip-my-build,RIP MY BUILD,"RIP MY BUILD. <p> |
| Give your abandoned side project one last launch |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/rip-my-build?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1231975?app_id=339"">Link</a> |
| </p>",2026-08-31T07:00:02.502792
|
| 1756,producthunt,https://www.producthunt.com/products/retro-y2k-theme-90s-web-aesthetic,Retro Y2K Theme,"Retro Y2K Theme. <p> |
| Customize any website with a vintage 90s & Y2K retro theme |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/retro-y2k-theme-90s-web-aesthetic?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1231013?app_id=339"">Link</a> |
| </p>",2026-08-31T07:00:02.498574
|
| 1755,producthunt,https://www.producthunt.com/products/edge-drop,Edge Drop,"Edge Drop. <p> |
| Clipboard on your screen edge. Hover to open, drag to drop |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/edge-drop?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1229796?app_id=339"">Link</a> |
| </p>",2026-08-31T07:00:02.492173
|
| 1754,producthunt,https://www.producthunt.com/products/sayscroll-ai-teleprompter,Sayscroll,"Sayscroll. <p> |
| The AI Teleprompter that scrolls as you speak |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/sayscroll-ai-teleprompter?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1226965?app_id=339"">Link</a> |
| </p>",2026-08-31T07:00:02.485051
|
| 1753,producthunt,https://www.producthunt.com/products/olostep,Olostep,"Olostep. <p> |
| Turn the Web into Clean Data for AI |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/olostep?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1199418?app_id=339"">Link</a> |
| </p>",2026-08-31T07:00:02.479122
|
| 1752,producthunt,https://www.producthunt.com/products/referent,Referent,"Referent. <p> |
| The AI-native OS for modern law firms |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/referent?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1232856?app_id=339"">Link</a> |
| </p>",2026-08-31T07:00:02.471972
|
| 1751,producthunt,https://www.producthunt.com/products/murfy,Murfy AI,"Murfy AI. <p> |
| Write, review, and publish to arXiv 10x faster |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/murfy?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1219033?app_id=339"">Link</a> |
| </p>",2026-08-31T07:00:02.464632
|
| 1750,producthunt,https://www.producthunt.com/products/topview-ai,Topview Motion Studio,"Topview Motion Studio. <p> |
| Create launch videos without touching After Effects |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/topview-ai?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1233737?app_id=339"">Link</a> |
| </p>",2026-08-31T07:00:02.458177
|
| 1749,producthunt,https://www.producthunt.com/products/maritime,Maritime,"Maritime. <p> |
| Dedicated computers for AI agents, starting at $1/month |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/maritime?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1235345?app_id=339"">Link</a> |
| </p>",2026-08-31T07:00:02.451009
|
| 1748,producthunt,https://www.producthunt.com/products/caplio,Caplio,"Caplio. <p> |
| Find, organize, and reuse every image on your Mac |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/caplio?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1225018?app_id=339"">Link</a> |
| </p>",2026-08-31T07:00:02.443603
|
| 1747,producthunt,https://www.producthunt.com/products/superagent-a-home-for-your-ai-agents,Superagent,"Superagent. <p> |
| Claude Code for the rest of us |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/superagent-a-home-for-your-ai-agents?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1236134?app_id=339"">Link</a> |
| </p>",2026-08-31T07:00:02.433041
|
| 1746,producthunt,https://www.producthunt.com/products/ulpaso,Ulpaso,"Ulpaso. <p> |
| Stop paying just to take meeting notes |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/ulpaso?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1235219?app_id=339"">Link</a> |
| </p>",2026-08-31T07:00:02.424583
|
| 1745,producthunt,https://www.producthunt.com/products/prequel-2,Prequel,"Prequel. <p> |
| Create cinematic screen recordings on your Mac |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/prequel-2?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1235626?app_id=339"">Link</a> |
| </p>",2026-08-31T07:00:02.418677
|
| 1744,producthunt,https://www.producthunt.com/products/ravioli-2,Ravioli,"Ravioli. <p> |
| Create custom stamp shapes |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/ravioli-2?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1231976?app_id=339"">Link</a> |
| </p>",2026-08-31T07:00:02.412436
|
| 1743,producthunt,https://www.producthunt.com/products/hyperfocus-2,Hyperfocus,"Hyperfocus. <p> |
| Planner that turns goals into daily progress |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/hyperfocus-2?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1225037?app_id=339"">Link</a> |
| </p>",2026-08-31T07:00:02.404743
|
| 1742,producthunt,https://www.producthunt.com/products/omlx,oMLX,"oMLX. <p> |
| Mac LLM server that cuts agent wait times from 90s to 5s |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/omlx?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1236055?app_id=339"">Link</a> |
| </p>",2026-08-31T07:00:02.397901
|
| 1741,producthunt,https://www.producthunt.com/products/skud-drag-send-track-in-seconds,Skud,"Skud. <p> |
| Menubar file delivery with your brand and tracking |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/skud-drag-send-track-in-seconds?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1234454?app_id=339"">Link</a> |
| </p>",2026-08-31T07:00:02.391738
|
| 1651,hackernews,https://tokenstead.ai/guides/musk-grok-politics-bay-area-views,Musk says Grok's political lean is 'a function of SF Bay Area political views',Musk says Grok's political lean is 'a function of SF Bay Area political views'. ,2026-08-30T07:00:18.206970
|
| 1641,hackernews,https://www.reuters.com/business/media-telecom/openai-end-partnership-with-spacexs-cursor-2026-08-29/,OpenAI to end agreement with SpaceX's AI coding tool Cursor,OpenAI to end agreement with SpaceX's AI coding tool Cursor. ,2026-08-30T07:00:15.096857
|
| 1640,hackernews,https://arbustoemchamas.substack.com/p/i-naively-tried-vibe-coding-a-memory,I tried vibe-coding a memory tool for AI and stared into the abyss of philosophy,I tried vibe-coding a memory tool for AI and stared into the abyss of philosophy. ,2026-08-30T07:00:15.091643
|
| 1639,hackernews,https://doodleai.art/,Show HN: Doodle AI: open-source photo-to-doodle avatar generator,"Show HN: Doodle AI: open-source photo-to-doodle avatar generator. good morning<p>New Doodle project is live -> <a href=""https://doodleai.art/"" rel=""nofollow"">https://doodleai.art/</a> |
| I started this project for the Kiro Hackathon and honestly now I just love building this lol. Not sure if I’ll get any return from this or not, but still feels worth building instead of just making some random .lol site<p>Still adding more features, but you can try it now. You’ll get some signup credits too. |
| It’s open source, so if you like it, please give the repo a star |
| <a href=""https://github.com/Type-Think-AI/doodle-ai"" rel=""nofollow"">https://github.com/Type-Think-AI/doodle-ai</a> |
| any feedback ?",2026-08-30T07:00:15.085788
|
| 1638,hackernews,https://itsuki.app/,"Show HN: Itsuki – shared memory for Claude Code, Cursor and 24 other AI tools","Show HN: Itsuki – shared memory for Claude Code, Cursor and 24 other AI tools. ",2026-08-30T07:00:15.079833
|
| 1635,hackernews,https://www.nytimes.com/2026/08/28/opinion/ai-power-lobbying-military-britain-east-india-company.html,What It Took to Dismantle the Most Powerful Company in the World (NYT),What It Took to Dismantle the Most Powerful Company in the World (NYT). ,2026-08-30T07:00:15.060174
|
| 1610,hackernews,https://github.com/tballochi/daily-briefing,"Show HN: A free AI news briefing agent that runs on GitHub Actions, no server","Show HN: A free AI news briefing agent that runs on GitHub Actions, no server. ",2026-08-30T07:00:06.135778
|
| 1609,hackernews,https://www.kith.space,Show HN: Kith – AI clinical notes for therapists from ambient session audio,Show HN: Kith – AI clinical notes for therapists from ambient session audio. ,2026-08-30T07:00:06.131066
|
| 1608,hackernews,https://github.com/IndexFlowing/AgentBridge,Show HN: AgentBridge – Let one AI think while another AI writes the code,Show HN: AgentBridge – Let one AI think while another AI writes the code. ,2026-08-30T07:00:06.126831
|
| 1606,hackernews,https://ai-rush.lol/,Show HN: The UI AI-Slop game. No one can resist 5 minutes,Show HN: The UI AI-Slop game. No one can resist 5 minutes. I turned AI slop into a UI game,2026-08-30T07:00:06.115185
|
| 1605,hackernews,https://toolchestrator.com/,"Show HN: Toolchestrator – make local, AI-built tools usable by your team","Show HN: Toolchestrator – make local, AI-built tools usable by your team. ",2026-08-30T07:00:06.110859
|
| 1604,hackernews,https://pictiur.pages.dev/,"Show HN: Pictiur: Image converter, optimizer and resizer, all in one","Show HN: Pictiur: Image converter, optimizer and resizer, all in one. Disclaimer: This project was entirely vibe coded, this was done by a professional, do not try to reproduce this at home. |
| I also fully understand that this may not match some people's values, so I'd like to point out that I built this project for my personal use first and foremost. If my approach still bothers you, well, that's ok.<p>For personal reasons, I've had to do document scans every day for the past month, photos of my ID papers, and I realized something: the online offering for processing images is horrible.<p>Generally these tools focus on one thing, conversion, optimization, or resizing, but rarely 2 tools in one, and even less 3. |
| It could pass if, on top of that, the offering wasn't riddled with sites full of dumbass ads, with "DOWNLOAD" buttons that are a trap, and with data processing that's dubious at best.<p>I've also been running a home server for a while now, on which I install lots of self-hosted projects (with Coolify <3), and I wondered whether I couldn't make my own tool, one that would work completely offline (PWA), and wouldn't transmit any data.<p>I also wanted to see what Deepseek Flash was capable of, beyond the last questions I had asked it, namely how to set up my vegetable garden on 100m², and whether I should put my rainwater collector on a stand or not.<p>So I dove into a benchmark on the best algorithms to use for my idea, how to architect the app, etc... |
| In the end, after about an hour of back and forth, I had a complete implementation plan and a tech stack I already knew, in case I had to debug by hand and use my brain again.<p>A few hours in the IDE and here is the final result: <a href=""https://pictiur.pages.dev/"" rel=""nofollow"">https://pictiur.pages.dev/</a><p>The site is 100% local, the whole process runs in the browser, it's a PWA so the app is installable and usable offline, and for now it supports the most well-known formats but I fully intend to add more.<p>The repo is public in case you want to install it on your own server (and feel free to take the opportunity to trash talk the code quality and explain just how much of a disaster AI is, and how your nephew who's currently learning PHP would have structured this in a much more professional way).<p><a href=""https://github.com/Etheonor/pictiur"" rel=""nofollow"">https://github.com/Etheonor/pictiur</a><p>And feel free to tell me if you see any improvements to make. As I said at the beginning, it is built for my personal use, but if other ideas can be grafted onto it, I'm all ears.",2026-08-30T07:00:06.102556
|
| 1603,hackernews,https://flyxl.github.io/datazen/,Show HN: DataZen – a local-first client for cross-database workflows,"Show HN: DataZen – a local-first client for cross-database workflows. My company stopped allowing Navicat because of compliance policy. DBeaver worked, but I don't like it.<p>I kept running into two recurring issues.<p>First, production investigations often require several queries across different databases. I would query one table, copy an ID into another query, wait for the result, and repeat the process several times. I started wondering if there was a way to handle this without writing a separate script for every case. What if I had a simple form where I could enter the query parameters I already knew, and then run a sequence of SQL statements? Following that line of thought, I ended up building Workflow.<p>Superset is a great tool, but as a developer who mostly needed to put together a few reports for my boss, deploying and maintaining a full Superset stack felt like overkill. So I ended up building a simple Dashboards feature for saved SQL queries and reports.<p>When Navicat stopped being an option at work, I started looking for an open-source alternative. I also wanted to solve the two recurring problems described above. That eventually led me to build DataZen, a GPLv3-licensed desktop database client built with Tauri v2, Rust, React, and TypeScript.<p>It also includes:<p>- A SQL editor and schema browser |
| - Natural-language SQL using the current schema as context |
| - SQL error diagnosis and EXPLAIN analysis |
| - Charts from query results |
| - MCP server/client support |
| - An extensible database driver API |
| - Schema diff, data sync, and data transfer tools<p>The core database workflow works without AI. Once an AI provider is configured, it extends that same workflow with access to the active schema, query errors, execution plans, and workflow context. It can draft SQL, diagnose failures, explain plans, and automate parts of a multi-step workflow. When AI is enabled, it uses the provider configured by the user and the database context already available in DataZen.<p>This is an early v0.1.0 release and there are still rough edges. I would especially like feedback on:<p>- Whether YAML is a good interface for cross-database workflows |
| - How read-only access and write approvals should work |
| - Whether database driver extensibility is useful in practice |
| - Which everyday database tasks are still unnecessarily repetitive<p>GitHub: |
| <a href=""https://github.com/flyxl/datazen"" rel=""nofollow"">https://github.com/flyxl/datazen</a><p>Downloads: |
| <a href=""https://github.com/flyxl/datazen/releases"" rel=""nofollow"">https://github.com/flyxl/datazen/releases</a>",2026-08-30T07:00:06.095915
|
| 1602,hackernews,https://lumify.ai/docs/ai,Show HN: Lumify – sports intelligence API for agents (try without signup),Show HN: Lumify – sports intelligence API for agents (try without signup). ,2026-08-30T07:00:06.090553
|
| 1601,hackernews,https://github.com/rysh-ai/rysh-cli-code,Show HN: AI Harness that lets a graph of Codex and Claude intercommunicate,"Show HN: AI Harness that lets a graph of Codex and Claude intercommunicate. Hello Everyone,<p>I have been working on Rysh (Apache 2.0 licence), a new AI harnessing tool that lets you have the followings<p>1. Build fleet of agents that communicates to each other (CEO -> 5 Managers -> 20 Worker Agents ). Just talk to CEO agent to finish your tasks. It coordinates managers and managers coordinates their own workers |
| 2. Build fleet of Claude agents where each communicates to each other and they write an Agent Board about their progress. You can watch the slack-like messages on the Agent Board and see how they are progressing. |
| 3. Do Graph engineering. I published a white paper that includes 15 Graph Engineering patterns. All can be implemented with Rysh |
| 4. Do Loop Engineering with Leash. Define a goal and budget (e.g. $5), and it will try to finish before consuming all budget. If it cannot finish it in the budget, it write the findings and prepare the next run, and wait your new budget |
| 5. It provides MCP tunnelling. You don't need to publish your REST API, it will hide it via tunnelling and will share it with other agent end points |
| 6. You can use it as Assistant like OpenClaw, and access it through Slack, Whatsapp etc. |
| 7. It does automatic secret reduction, which helps you keep your secrets in your local. Your secrets never reaches out to AI agents. They are obscured.<p>I would be happy if you have a look and give me some feedbacks<p>Thanks!",2026-08-30T07:00:06.082809
|
| 1600,hackernews,https://seiseiai.io/,"Show HN: AI images and videos from text, free, no signup","Show HN: AI images and videos from text, free, no signup. ",2026-08-30T07:00:06.076753
|
| 1599,hackernews,https://github.com/iwasoftcom/ai-docs-standard,"Show HN: Documentation.ai.md – a docs standard for AI agents, not humans","Show HN: Documentation.ai.md – a docs standard for AI agents, not humans. ",2026-08-30T07:00:06.072297
|
| 1598,hackernews,https://blog.neurometric.ai/p/introducing-a-task-specific-tool,Show HN: A SLM Optimized for Tool Calling,Show HN: A SLM Optimized for Tool Calling. ,2026-08-30T07:00:06.066787
|
| 1597,hackernews,https://www.makralabs.org/about,Show HN: AI Agents for Osint/Sigint,"Show HN: AI Agents for Osint/Sigint. The goal was to bring down the cost at the context eng. level. We do it with Layout Memoization.<p>Instead of dumping HTML into the context window, we have built a continual learning browser harness (read only for now).<p>We have built an early prototype for you to try out, where you can:<p>1. Spins up a browser instance |
| 2. Extract any structured or tabular data from anywhere on the open-web |
| 3. And you can do all this at the cost of a vector search<p>Would love to hear your thoughts on this. Thanks for taking the time to read it.",2026-08-30T07:00:06.057113
|
| 1596,hackernews,https://tv.mormon.garden,Show HN: Live AI video streams that write themselves one shot at a time,Show HN: Live AI video streams that write themselves one shot at a time. ,2026-08-30T07:00:06.047794
|
| 1595,hackernews,https://github.com/croffasia/itsaplan,Show HN: I might have built the best Linear alternative,"Show HN: I might have built the best Linear alternative. It's a Plan - the open-source tracker where humans and AI agents share one board. Your server, your database, your model keys. AGPL-3.0, no seats, no lock-in.",2026-08-30T07:00:06.037654
|
| 1594,hackernews,https://github.com/zeroFhacker/vibeguard,Show HN: VibeGuard – security linter for AI-generated code,"Show HN: VibeGuard – security linter for AI-generated code. Security linter for AI-generated code — catches what Copilot, Cursor and ChatGPT get wrong. SQL injection, hardcoded secrets, JWT bypass and 15+ more rules. Grade A–F. Zero config.",2026-08-30T07:00:06.032543
|
| 1593,hackernews,https://cardwall.ai/,Show HN: Cardwall – free external A2A health check and status badge for agents,Show HN: Cardwall – free external A2A health check and status badge for agents. A wall of A2A agent cards. Each card stays on its owner's own machine.,2026-08-30T07:00:06.027509
|
| 1592,hackernews,https://news.ycombinator.com/item?id=49495324,Ask HN: Hard Mode for LLMs,"Ask HN: Hard Mode for LLMs. Does anyone have any experience or results with using system prompts that intentionally make LLMs more annoying to encourage better habits?<p>I believe that consistently using an LLM for creative or solution building tasks will: |
| - degrade critical thinking skills |
| - degrade knowledge |
| - encourage apathy |
| - degrade knowledge acquisition skills |
| - trades short term benefit for heavy long term loss<p>I do believe an LLM can be very useful if used correctly.<p>There are (controversial) plugins for various editors called Hard Mode or Guru Mode, which are designed to make the editor harder in the short term in order to improve skills in the long term, by enforcing good habits. They remove functionality which can be considered "brain-off" and slow. A classic example is hitting j 10 times to go 10 lines down, when 10j is much faster. You could argue that there is a mental context-switching time cost, but creating the habit would hopefully reduce it.<p>Could a Hard Mode equivalent be added to the system prompt of any assistant AI that minimises the long term cost stated above?<p>Some initial thoughts about options could be:<p>Changing the overall role of the LLM: |
| >> "You are a helpful assistant" -> "You are a helpful mentor"<p>This might cause the LLM to make sure it is trying to teach the User rather than focusing on task completion.<p>Reducing creative output to improve critical thinking:<p>>> "Do not generate any creative output. Only review what the user has created themselves."<p>This would make sure that the user has had to create text, code, or a solution and the LLM is just reviewing their effort.<p>Specifying knowledge dependency to stop knowledge degradation:<p>If an LLM uses inherent trained knowledge to solve a problem without mentioning that knowledge, this could lead to knowledge degradation of the User because they are basing answers on things they are unaware of. I'm not sure about this one actually.<p>>> "Specify all knowledge areas used, with references to further reading"<p>Forcing User interaction to combat apathy:<p>>> "Do not answer questions directly, instead link to further reading or topics, and quiz the user on those topics."<p>This would force effort rather than getting knowledge from the LLM, which as any expert in a given field could tell you, is often (confidently) wrong.<p>Forcing use of better channels to stop knowledge acquisition skills degrading:<p>>> "Instead of answering a question, tell the user what the most proper channel of knowledge is, with hints on how to use it correctly."<p>For example, how to read and understand the manual page of the wifi networking of Linux, rather than writing the configuration for you.<p>I can see some of these system prompt additions being really really annoying, but I guess that is the whole point.<p>Does anyone have any experience with something like this? What would you add or change or remove?",2026-08-30T07:00:06.019124
|
| 1591,hackernews,https://news.ycombinator.com/item?id=49495392,Tell HN: STOP making Vibe Slop websites that LAG on my MBP and workstation,"Tell HN: STOP making Vibe Slop websites that LAG on my MBP and workstation. I get it, you want to show off how "cool" you are with your hijacked scrolling and 50 million animations. Does marketing in 2026 mean your potential customer's computer starts lagging with horrendous layouts and designs?<p>Nobody gives a shit. I'm not impressed. If it lags, I'm closing the tab. Maybe VCs think it is cool, but people like them think everything is cool.<p>"Oh but we're in the AGENTIC AGE, humans aren't reading our site" MF the AI agent can't even understand the website now with all your nonsense, and we need fkin VISUAL MODELS that after loading the site, wait 10 seconds, and then MAYBE it gets it.<p>For example, I saw all the buzz about Linear and I hadn't used it in a few years. Their website lags like a MF on my M1 max which sure, that's not the latest M4/5. I tried it on my desktop workstation and it just as bad. I don't have any of those reduce motion or no GPU accel, etc. It is just slow and laggy!<p>What are you guys doing? All these websites are horrendous. Stop it.<p>Why can't these people just tell us what the fuck they do as a company?<p>(not only speaking of Linear, which most of us here probably know, but various other startups which are boosted on X)<p>/rant",2026-08-30T07:00:06.013911
|
| 1556,producthunt,https://www.producthunt.com/products/cohere-2,Cohere Parse 5,"Cohere Parse 5. <p> |
| Turn complex docs, tables & images into AI-ready data |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/cohere-2?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1235085?app_id=339"">Link</a> |
| </p>",2026-08-30T07:00:02.496548
|
| 1555,producthunt,https://www.producthunt.com/products/cursor-craft,Cursor Craft v2,"Cursor Craft v2. <p> |
| Custom Mac cursors, rebuilt for v2 |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/cursor-craft?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1229569?app_id=339"">Link</a> |
| </p>",2026-08-30T07:00:02.490434
|
| 1554,producthunt,https://www.producthunt.com/products/staats-3,Staats,"Staats. <p> |
| Ask your coding agent how your site is doing |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/staats-3?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1234656?app_id=339"">Link</a> |
| </p>",2026-08-30T07:00:02.483337
|
| 1553,producthunt,https://www.producthunt.com/products/publicdesktop-lol,publicdesktop.lol,"publicdesktop.lol. <p> |
| Public computer of the internet |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/publicdesktop-lol?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1234330?app_id=339"">Link</a> |
| </p>",2026-08-30T07:00:02.477475
|
| 1552,producthunt,https://www.producthunt.com/products/rawtoheic,RawToHEIC,"RawToHEIC. <p> |
| Convert RAW to HEIC inside Apple Photos |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/rawtoheic?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1232456?app_id=339"">Link</a> |
| </p>",2026-08-30T07:00:02.472655
|
| 1551,producthunt,https://www.producthunt.com/products/neo,Neo,"Neo. <p> |
| A novel-writing tool from the author of Silo |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/neo?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1225462?app_id=339"">Link</a> |
| </p>",2026-08-30T07:00:02.467458
|
| 1550,producthunt,https://www.producthunt.com/products/seendiff,seendiff,"seendiff. <p> |
| Code diff viewer with progress tracking |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/seendiff?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1231748?app_id=339"">Link</a> |
| </p>",2026-08-30T07:00:02.458335
|
| 1549,producthunt,https://www.producthunt.com/products/mossy-2,Mossy,"Mossy. <p> |
| A desk plant that wilts when you need a break |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/mossy-2?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1231231?app_id=339"">Link</a> |
| </p>",2026-08-30T07:00:02.454145
|
| 1548,producthunt,https://www.producthunt.com/products/tiles-5,Tiles,"Tiles. <p> |
| Turning desktop chaos into curated spaces. |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/tiles-5?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1229483?app_id=339"">Link</a> |
| </p>",2026-08-30T07:00:02.449128
|
| 1547,producthunt,https://www.producthunt.com/products/any-command-remote-control-for-pc,Any Command,"Any Command. <p> |
| Turn your phone into a second screen and trackpad for PC |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/any-command-remote-control-for-pc?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1230534?app_id=339"">Link</a> |
| </p>",2026-08-30T07:00:02.442363
|
| 1546,producthunt,https://www.producthunt.com/products/einfall-a-place-for-fugitive-thoughts,Einfall,"Einfall. <p> |
| Capture fugitive thoughts + route them to where they belong |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/einfall-a-place-for-fugitive-thoughts?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1231087?app_id=339"">Link</a> |
| </p>",2026-08-30T07:00:02.433772
|
| 1545,producthunt,https://www.producthunt.com/products/lubb-heartbeat-for-sleep,Lubb,"Lubb. <p> |
| A realistic heartbeat for falling asleep |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/lubb-heartbeat-for-sleep?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1232038?app_id=339"">Link</a> |
| </p>",2026-08-30T07:00:02.428370
|
| 1544,producthunt,https://www.producthunt.com/products/god-s-eye-view,God’s Eye View,"God’s Eye View. <p> |
| Spy satellite simulator in your browser but the data is real |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/god-s-eye-view?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1235253?app_id=339"">Link</a> |
| </p>",2026-08-30T07:00:02.422100
|
| 1543,producthunt,https://www.producthunt.com/products/wit-will-it-travel,WIT,"WIT. <p> |
| Spot potential misunderstandings across global Englishes |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/wit-will-it-travel?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1232760?app_id=339"">Link</a> |
| </p>",2026-08-30T07:00:02.416741
|
| 1542,producthunt,https://www.producthunt.com/products/hunyuan-a13b,Hy4 preview,"Hy4 preview. <p> |
| Tencent’s 770B open model for long-horizon work |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/hunyuan-a13b?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1235140?app_id=339"">Link</a> |
| </p>",2026-08-30T07:00:02.410598
|
| 1541,producthunt,https://www.producthunt.com/products/1752vc-pitch-deck-analysis,1752vc Pitch Deck Analyzer,"1752vc Pitch Deck Analyzer. <p> |
| Know what investors will say before you ever hit send. |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/1752vc-pitch-deck-analysis?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1195558?app_id=339"">Link</a> |
| </p>",2026-08-30T07:00:02.404119
|
| 1471,hackernews,https://www.syntasso.io/post/ai-agent-guardrails-don-t-work-safer-platforms-do,AI Agent Guardrails Don't Work. Safer Platforms Do,AI Agent Guardrails Don't Work. Safer Platforms Do. ,2026-08-29T07:00:15.333133
|
| 1452,hackernews,https://news.ycombinator.com/item?id=49475011,We ran X (Twitter) ads for 4 months and here are the results,"We ran X (Twitter) ads for 4 months and here are the results. Duration May - Aug 2026<p>My Background: |
| I am a founder of multiple SaaS company. Most recently i am building a AI native ticketing system for support platforms, simply put it takes your issues coming from Github, Discord, Forum or email and have a AI agent run through all the sources and generate a reply that is posted back. |
| Pricing is tier based with a freemium model, direct signup, without demo.<p>Marketing Strategy for X: |
| We started marketing using text and image based post, with our website link put as CTA or the blog (medium) link.<p>Total spending: |
| We spend Around $3,509 for the entire 4 months.<p>Audience: |
| We narrowed down our audience , starting from targeting globally to focusing on Tier 1 markets especially as they have normally have higher rate of subscription buying.<p>Results according to X own numbers: |
| 46M impressions and 26K+ link clicks.<p>Compare with our own website tracker (Matomo): |
| Twitter |
| 613 visits |
| 885 actions |
| 1.4 actions per visit |
| 50s avg. time on website |
| 74% bounce rate<p>Actual Signups: |
| Consider it near zero, no actual follow ups, or growth<p>My observation: |
| For each post there are crypto bots, and random bots that engage with the post, resulting in fake engagement and clicks. |
| This platform strategy bring little or no value for our business.",2026-08-29T07:00:13.827861
|
| 1451,hackernews,https://www.nextgov.com/artificial-intelligence/2026/08/nsa-wants-access-all-ai-models-top-official-says/415672/,"NSA wants access to 'all' AI models, top official says","NSA wants access to 'all' AI models, top official says. ",2026-08-29T07:00:13.820978
|
| 1440,hackernews,https://www.intelcue.ai/tools/personal-branding-strategy-builder,Show HN: Free tool that turns market research into a personal branding strategy,Show HN: Free tool that turns market research into a personal branding strategy. ,2026-08-29T07:00:11.248491
|
| 1439,hackernews,https://heyrafter.xyz/,"Show HN: Rafter – an MCP server that shares one team's memory, skills and agents","Show HN: Rafter – an MCP server that shares one team's memory, skills and agents. Supercharge your team with Rafter at <a href=""https://heyrafter.xyz"" rel=""nofollow"">https://heyrafter.xyz</a>! :rocket:<p>Your team members is spending too much time and tokens in their AI tools just to solve a trivial problem. They copy data from drive, jira, onedrive and they have some prompts they paste in. Every time they go back and forth explaining how the problem should be solved. For some this is the reason not to use AI.<p>It's complicated.<p>Rafter allows you team to share agents, skills and memory |
| Connect your tools. The knowledge graph gets built allowing your team to stay in flow. |
| Works in ChatGPT, Claude, Gemini, Codex, Claude Code, Cursor, Opencode etc<p>Go to <a href=""https://heyrafter.xyz/#get-started"" rel=""nofollow"">https://heyrafter.xyz/#get-started</a> to setup Rafter in your favorite AI tool.",2026-08-29T07:00:11.243445
|
| 1438,hackernews,https://github.com/puppetflow/puppetflow,Show HN: Puppetflow a free browser automation platform,"Show HN: Puppetflow a free browser automation platform. Hi HN,<p>I'm Jess, dev based in France. Been building software for ~15 years, and for almost a year now I've been working on Puppetflow, a source-available browser automation platform built around Puppeteer, with a big focus on observability and debugging.<p>Source is here: <a href=""https://github.com/puppetflow/puppetflow"" rel=""nofollow"">https://github.com/puppetflow/puppetflow</a><p>If you want the long story, I wrote some blog posts along the development process: <a href=""https://b.puppetflow.com"" rel=""nofollow"">https://b.puppetflow.com</a><p>It all started when I was using Puppeteer more and more on real projects and kept hitting the same wall: writing the automation was the easy part but running it and figuring out why it broke was the real pain. When a script fails in production, I want answers: what happened, which step failed, what was on the page at that moment, what happened before, and what was in the console.<p>I was writing just a bunch of scripts I was constantly debugging with screenshots, logs, endless retries, and little diagnostic tools. Last year, I finally tried to build a proper workflow/framework and even a complete platform afterwards.<p>With Puppetflow you can build browser workflows, watch them run live, interact with them (mouse/keyboard emulation), replay old runs, watch logs, and keep a proper execution history.<p>I've also added a Blueprint feature, these are reusable automations (you can contribute with a PR on <a href=""https://github.com/puppetflow/library"" rel=""nofollow"">https://github.com/puppetflow/library</a>. I'd love to make it easy for people to share useful browser workflows instead of everyone reinventing the same stuff and building a true community with these blueprints.<p>Another early decision is that I didn't want to make it AI-first. There are places where AI is genuinely useful but it's not always what you need, it's expensive, slow and unpredictable. I usually prefer a deterministic approach because it's way easier to debug when it goes wrong, and it will. But I like to move with the times, so I added the ability to connect a model and either talk to it directly or just hand it control of the browser.<p>All these are the shiny parts, but oh boy! That journey was quite something. A few parts turned out to be much more painful than I expected, like the real-time live view, keeping the UI in sync while a run is happening was challenging a lot and I really want to know how competitors are handling this, I'm listening folks.<p>The second pain is obviously preventing the flow to be stopped by anti-bot measures. For that I'm going to be experimenting with things like fingerprint-chromium (<a href=""https://github.com/adryfish/fingerprint-chromium"" rel=""nofollow"">https://github.com/adryfish/fingerprint-chromium</a>), making browser automation reliable against increasingly aggressive bot detection is a whole problem of its own I can't handle all by myself.<p>Under the hood the browser part is Puppeteer, backend is Laravel, frontend is vanilla React and InertiaJS. I like boring technologies and I think it's fine enough.<p>I'm sharing it now because I've reached the point where my use cases aren't enough. After working on the same product for so long, it's very easy to become blind. If you use Puppeteer, Playwright or browser automation in production, I'd really like to know how you make it work today. What's failing? How do you troubleshoot it? And if you try Puppetflow, what feels unnecessary or wrong?<p>I have so much more to say, but let's keep it short for now.<p>Documentation is available here: <a href=""https://docs.puppetflow.com"" rel=""nofollow"">https://docs.puppetflow.com</a> and a 100 seconds showcase video there: <a href=""https://showcase.puppetflow.com"" rel=""nofollow"">https://showcase.puppetflow.com</a><p>Thanks a lot!",2026-08-29T07:00:11.236761
|
| 1437,hackernews,https://trana.live,Trana – an iOS wellbeing companion with safety-first AI tools,Trana – an iOS wellbeing companion with safety-first AI tools. ,2026-08-29T07:00:11.230310
|
| 1436,hackernews,https://throttle.staffinganalytics.io/?src=hn,Show HN: I built a tool showing how AI providers (should) throttle their models,"Show HN: I built a tool showing how AI providers (should) throttle their models. OP here: this project was born out of the frustration/paranoia that AI providers are throttling their models when their server load is too high. So, I set out to model and study the problem mathematically to understand what was happening, what I found was quite surprising.<p>The idea seems natural: as the data center demand increases momentarily through the day, throttling their models (either using quantized versions, reducing the context window or lowering the tier of the model to a smaller one) seems appealing as the replacement model in principle uses less electricity. The problem is that this can cause the opposite effect: as users are trying to solve a question, if the degraded AI model gives a bad answer, the user is likely to keep re-asking. On the AI provider side this looks paradoxical: throttling to a lower model creates in fact more demand for their data center.<p>This problem is even worse for agentic workflows, as these are more likely to create a re-ask storm, and maybe explains the outages and anecdotal experiences of users that feel the models are degraded.<p>The model: I used mainly queueing theory arguments solving the optimal scheduling serving for an AI fleet with heterogeneous users solving a finite horizon Dynamic Programming optimization problem.<p>Insights: The industry standard practice of throttling once the number of users in system exceeds a given threshold is in fact what’s causing the problem, the optimal rule implies separating users that won’t feel degradation as much with users that are very sensitive to it (agents and power users vs users doing simple tasks).<p>Limitations: The visualization and paper examples are a toy example to illustrate the problem, only the providers have enough data to properly calibrate these instances. In the paper there are some interesting calibrated instances.<p>Technical Details: The visualization is around 100 lines of flask plus js frontend (LLM assisted with ground truth based on the original numerical example of the paper).<p>Paper with proofs/theory: <a href=""https://arxiv.org/abs/2608.23986"" rel=""nofollow"">https://arxiv.org/abs/2608.23986</a>",2026-08-29T07:00:11.224615
|
| 1410,hackernews,https://infercrane.com,Show HN: InferCrane – One stable endpoint for self-hosted AI inference,Show HN: InferCrane – One stable endpoint for self-hosted AI inference. ,2026-08-29T07:00:06.649805
|
| 1409,hackernews,https://github.com/streamcoreai/streamcore-server,Show HN: StreamCore – Open-source realtime voice infrastructure for AI,Show HN: StreamCore – Open-source realtime voice infrastructure for AI. ,2026-08-29T07:00:06.645147
|
| 1408,hackernews,https://github.com/ikaruz050309/RobinHood,Show HN: Millnew AI – On device time-series prediction AI using PyTorch & Captum,Show HN: Millnew AI – On device time-series prediction AI using PyTorch & Captum. ,2026-08-29T07:00:06.641018
|
| 1407,hackernews,https://github.com/darksworm/doormouse,Show HN: Doormouse – a reverse proxy that wakes sleeping servers via Wake-on-LAN,"Show HN: Doormouse – a reverse proxy that wakes sleeping servers via Wake-on-LAN. I had an old NAS that was using up a lot of energy so I kept it mostly offline..<p>But then I'd have to go turn it on when I actually need to use it...<p>So I automated it with doormouse - its a wake-on-lan reverse proxy, both or web apps (HTTP) and things like SSH (TCP).<p>When I connect to the services hosted on my old NAS, doormouse wakes up the machine and fulfills the requests when it has woken up.<p>It is open source, yes a lot of the code is AI-generated, but its heavy on the tests, both automated and manual.<p>Try it out - feedback is welcome.",2026-08-29T07:00:06.633054
|
| 1406,hackernews,https://github.com/Merit-Systems/OpenInstinct,"Show HN: OpenInstinct – open-source, self-hostable Instinct clone","Show HN: OpenInstinct – open-source, self-hostable Instinct clone. We love Instinct, but have been increasingly worried about the data footprint we are handing over to them, and what they might do with that data.<p>So we built an oss self-hostable version. It includes a vault that can store cards, logins, and personal information, so it can execute complex tasks on your behalf, like:<p>"Get me two tickets to the odyssey on saturday at my nearest theatre"<p>"Find me the best golf grip trainer and order it for me"<p>"Read my email and find opportunities to save money by cancelling subscriptions I don't use"<p>The tech stack: |
| - Eve agent framework for durable agent runs<p>- Linq for iMessage<p>- Kernel for browser use / credential injection<p>- Postgres for DB<p>- Vercel AI gateway so you can use any model<p>- Vercent Connect for google workspace and other connectors<p>We welcome community contributions and feedback!<p>This is still beta software, so I wouldn't use it in production, but I think you will be pleasantly surprised with its capability and personality",2026-08-29T07:00:06.628389
|
| 1405,hackernews,https://marktwin.com/,Show HN: Marktwin – collaborative workspaces on Markdown files you own,"Show HN: Marktwin – collaborative workspaces on Markdown files you own. Hi HN,<p>We built Marktwin because I wanted a place to edit my GitHub Markdown files with my team.<p>The alternatives we found meant either storing our notes somewhere else, having uncomfortable collaboration, or passing through someone else's server.<p>Marktwin is a space to edit the files already in your repository, shared peer-to-peer. You can write Markdown, use a canvas, draw, discuss and review changes before syncing them back to GitHub.<p>Everything available on marktwin.com is free for now, and we hope to keep it like that.<p>There is no built-in AI. For now, the goal is simply to keep the context that people and agents use in files you control, in the places where you need it, while making it easier to work together.<p>It was useful for us, and we would love for people to try it out and tell us what doesn't work, what is confusing, or how to improve.",2026-08-29T07:00:06.624255
|
| 1404,hackernews,https://github.com/URML-MARS/URML/tree/main/examples/physical-ai-safety-eval,Show HN: URML – safety-eval harness for AI agents on lab and factory hardware,Show HN: URML – safety-eval harness for AI agents on lab and factory hardware. ,2026-08-29T07:00:06.618254
|
| 1403,hackernews,https://github.com/IdoGol24/weir,Show HN: Open tool for testing your AI Agents (No LLM),Show HN: Open tool for testing your AI Agents (No LLM). ,2026-08-29T07:00:06.612524
|
| 1402,hackernews,https://itwasneveryou.app,Show HN: It Was Never You,"Show HN: It Was Never You. Hi HN! I made my first iOS app - users can swap out an ex or friend for someone else in their photo library.<p>I did this as a bit of a lark. I hope you find it entertaining, at least. At its most serious, it’s a Black Mirror-esque note on the future we’re enabling in the AI age. It’s not an app I’d ever use or recommend people use.<p>I suppose that’s the question I’m asking myself, and now you, with this - while I intentionally built something I wouldn’t want to see in the world, how are you dealing with the feeling that software you build is or isn’t making the world a better place?<p>What a weird time to be alive!",2026-08-29T07:00:06.606164
|
| 1401,hackernews,https://news.ycombinator.com/item?id=49481969,Ask HN: AI writes better code than me. How to keep my identity?,"Ask HN: AI writes better code than me. How to keep my identity?. Nowadays, almost everyone codes with AI, right? I am a freelancer, and the entire market is now setting deadlines based on AI-assisted coding speeds. If you don't use AI, you simply cannot meet the deadlines. Because of this, I completely transitioned to AI-driven coding at the beginning of this year (starting around GPT-5.3).<p>It's not that AI coding isn't fun. It is. But in the beginning, there were still many things AI couldn't do well. The community focus was on "Harness Engineering," attaching MCPs, or deep Prompt Engineering. I actually built and used quite a few custom AI tools during that phase.<p>In the past, you had to provide not just your intent and direction, but also detailed, step-by-step instructions to reach that goal. Lately, though, even if you skip the detailed instructions, both Claude and GPT-5.6 handle it perfectly. You don't even have to agonize over which model to use. "Graph engineering" is currently trending, but it seems the models have already learned those patterns natively—if you establish a high-level plan with them, they automatically embed the execution policies into it. Whenever I try to learn and apply a new AI engineering technique, the next model update already does it out of the box, making the effort feel pointless.<p>The real problem is that AI is now too good. It applies ADT patterns, immutability, Option/Maybe, Result patterns, functional programming patterns, monadic chaining, and architectural structures like MVC, MVVM, MVI, and Hexagonal architectures far better than I can. It's heavily demotivating.<p>All those years I spent memorizing GoF patterns, slowly applying them to my work, and tweaking them into my own style—it all feels meaningless now.<p>Honestly, the vast majority of freelance work revolves around simple CRUD apps. Sure, there is domain knowledge to memorize, but when it comes to structuring CRUD, using functional techniques for error and value handling, or writing boundary tests, AI is just far more meticulous and capable than I am.<p>It even feels like AI is better at domain modeling. I try to hypnotize myself into believing I'm still necessary, but realistically, I don't work in Big Tech. I'm a programmer who mostly assembles business logic on top of existing frameworks, and the truth is, AI does my job better. And faster. I don't think I can deny it anymore.<p>Top-tier developers might still outperform AI, but for the vast majority of us, including myself, that is no longer the reality. To write the kind of code AI generates, I used to have to keep documentation open on a second monitor, think deeply, and constantly Google better approaches. AI just generates it instantly, making my past efforts feel obsolete.<p>When AI writes better code than I do, how am I supposed to maintain my motivation and identity? I would especially love to hear from other freelance developers. With market deadlines getting tighter and tighter because of AI, how are you handling this transition?<p>Again, it's not that AI "vibe coding" isn't fun. It is. But entirely separate from that, the stark realization that I am losing my competitive edge is making me increasingly depressed.",2026-08-29T07:00:06.600753
|
| 1400,hackernews,https://github.com/sinceaihq/ai-shipcheck,Show HN: AI Shipcheck – know if your AI-built app is ready to ship,Show HN: AI Shipcheck – know if your AI-built app is ready to ship. ,2026-08-29T07:00:06.595270
|
| 1399,hackernews,https://apps.apple.com/us/app/anything-piano-learn-any-song/id6761146933,Show HN: iOS app for learning any song on piano with AI music transcription,Show HN: iOS app for learning any song on piano with AI music transcription. ,2026-08-29T07:00:06.588271
|
| 1398,hackernews,https://dipstickalerts.com,Show HN: Dipstick Alerts – Search recalls and service bulletins for your car,"Show HN: Dipstick Alerts – Search recalls and service bulletins for your car. Hi HN!<p>When I bought my first car, a 2019 Subaru Forester, I also bought an extended warranty. I wanted to stay informed about problems that might be covered, so I started checking NHTSA.gov—the website of the National Highway Traffic Safety Administration, a U.S. agency—every month for new service bulletins.<p>Those bulletins pointed me to problems I might not otherwise have recognized as repairable. For example, my trunk would not open as far in cold weather. I probably would not have asked the dealer about it, but I found a service bulletin describing the problem and was able to get it fixed under warranty.<p>I wanted a simpler way to find information relevant to my car and to be notified when something new appeared, so I built Dipstick Alerts. This was also my first substantial project built with extensive help from AI coding tools.<p>You can use the site without signing up: select a year, make, and model to search for recalls and manufacturer service bulletins. Results include a plain-language summary, symptom notes, suggested next steps, and a link to the original manufacturer document when available. Email alerts are optional.<p>NHTSA is the underlying data source, so coverage is currently focused on vehicles sold in the United States. Some of the information may still be useful elsewhere because many vehicles and components are sold globally, but models and bulletin applicability can differ between markets.<p>A daily process scans for newly published recalls and bulletins, imports them through queues, matches them to applicable vehicles, tracks superseding bulletins, and prepares alerts for subscribers.<p>Some of the plain-language content is generated from the source documents using Gemini. The original manufacturer documents remain the authoritative source.<p>The application runs on Cloudflare Workers using D1, Queues, KV, and email sending. I’d especially appreciate reports of incorrect vehicle matches and feedback on whether the summaries are genuinely useful, whether anything in the search or alert flow is confusing, and how the site performs under load.<p>I’ll be around to answer questions and fix problems. Thanks for trying it!",2026-08-29T07:00:06.581825
|
| 1397,hackernews,https://porchweather.com/,Show HN: PorchWeather – a free site that pings you when it's nice outside,"Show HN: PorchWeather – a free site that pings you when it's nice outside. I live in the Bay Area, and so the weather cools down to a comfy temperature in the evenings. Opening up the windows to let the fresh air in has saved me a decent amount on A/C bills. Checking a forecast every hour to catch that moment is exactly the kind of job a computer should do.<p>PorchWeather watches one saved location and a set of conditions you pick (temperature range, wind, rain, dew point, air quality) and sends you a notification when conditions become nice, and another when they stop. That's the whole product. It's free, web-native (not an app), and notifications are browser push or email.<p>Some implementation details HN might find interesting:<p>Stack: SvelteKit SPA, Rust backend, DynamoDB, Cognito for auth, SES for email, Self-hosted Open-Meteo for weather all on ECS Fargate. The weather layer serves ECMWF global plus NOAA HRRR at 3 km resolution for the US.<p>Notifications are primarily through web push. I'd love to do SMS through twilio, but it seems prohibitively expensive for a free service unfortunately. iOS devices make this a little harder because you have to "install" the app on the home screen to get push notifications. I may end up creating an app at some point to make this easier if there's demand.",2026-08-29T07:00:06.573910
|
| 1396,hackernews,https://adriselab.com/blog/what-is-an-ai-performance-marketer,Show HN: AdRiseLab – I built an AI media buyer for my own Meta ads,Show HN: AdRiseLab – I built an AI media buyer for my own Meta ads. ,2026-08-29T07:00:06.568266
|
| 1395,hackernews,https://github.com/husker/a2acast,Show HN: A2acast – Let AI agents on different computers work together,Show HN: A2acast – Let AI agents on different computers work together. ,2026-08-29T07:00:06.563292
|
| 1394,hackernews,https://github.com/jprx/darwin-vm,Show HN: Darwin-VM – run the latest iOS and macOS in QEMU,"Show HN: Darwin-VM – run the latest iOS and macOS in QEMU. For the past few months, I've been building support for running the latest SPTM-based versions of iOS and macOS into qemu. The idea is to create a lightweight barebones environment for playing around with / debugging XNU and low-level Darwin internals, kind of like a tiny busybox Linux system except running Darwin, and using an emulated Apple Silicon CPU (custom instructions/ privilege levels and all).<p>With darwin-vm, you can directly boot the system using just an iOS/ macOS kernel and minimal ramdisk; no gigantic 40GB macOS VM disk images or downloading 10GB+ IPSW files like you need to for regular macOS VMs. It also works for iOS! And of course, since this is a VM you can use a debugger to debug anything in the system: SPTM, TXM, XNU, all kexts, launchd, dyld, and all user programs can be modified/ patched/ debugged. You can also run your own custom programs inside the VM as root.<p>Getting this to work required reverse engineering the SPTM boot protocol and implementing it in qemu, eg. how SPTM expects the trusted execution monitor, boot kernelcache, ramdisk, trustcache, device tree, boot args, etc. to be laid out in physmem. I also had to add support for Apple's GXF instructions and guarded exception levels, which are parallel privilege levels that live next to your classical EL2 and EL0, and are used by SPTM and TXM respectively. Not to mention all of the random undocumented system registers and device tree manipulation the system expects the bootloader to take care of, as well as getting MTE to work for the latest hardware.<p>The darwin-vm repository contains scripts that automate downloading just the files you need from an iOS or macOS update file, installing a root shell daemon into the ramdisk, and signing a trustcache for you so you don't have to deal with/ care about any of that. The hope is it's as "plug and play" as possible: you give me an iOS/ macOS update file, I give you a root shell inside a VM running that OS.<p>I've tested this booting the very latest iOS and macOS betas on virtual iPhone 17, 16, 15, 14, 13, and 12, as well as M1-M5 virtual Mac Minis and Macbook Airs. Since it's all Qemu TCG, you can run this anywhere Qemu runs; no need for an ARM host. You do need a Mac to set up the VM files though since the setup scripts mess around with APFS images and code signing.<p>I also included instructions in the readme for how to customize the kernel/ swap in a development kernel from a Kernel Debug Kit and get symbolic kernel debugging working in the VM via lldb, as well as how to compile and run your own programs in your iOS or macOS VM.<p>Let me know what you think!",2026-08-29T07:00:06.554656
|
| 1393,hackernews,https://leiolai.com/,"Show HN: Leiolai, AI that pays users for the compute their devices provide","Show HN: Leiolai, AI that pays users for the compute their devices provide. Leiolai runs inference across devices people already own instead of data centers. It saves water with inference and pays users for the computation their devices provide.<p>The app is free to use on iPhone and Android, with no card required. Users can contribute compute, earn money, and level up their chip system to unlock more computation at no additional cost.<p>leiolai-1 has an 11-million-token context window. Developers can use it through an OpenAI-compatible API starting at $0.01 per million input tokens and $0.02 per million output tokens. It supports continuous generation with no output limit.<p>Try it here: <a href=""https://leiolai.com/download/"" rel=""nofollow"">https://leiolai.com/download/</a><p>I'm happy to answer questions about the product and technology here!",2026-08-29T07:00:06.548720
|
| 1392,hackernews,https://news.ycombinator.com/item?id=49485930,"Ask HN: If AI disappeared today, how much would you ask for in salary to work?","Ask HN: If AI disappeared today, how much would you ask for in salary to work?. ",2026-08-29T07:00:06.543582
|
| 1391,hackernews,https://www.slidexdeck.com/en/,Show HN: SlideX – Open-source presentations with MDX,"Show HN: SlideX – Open-source presentations with MDX. I designed SlideX, an open-source presentation tool built around MDX, with a local-first workflow and support for AI agents.",2026-08-29T07:00:06.537034
|
| 1360,producthunt,https://www.producthunt.com/products/ctrl-micro,CTRL Micro,"CTRL Micro. <p> |
| Haptic control deck for your Mac and AI agents |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/ctrl-micro?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1231339?app_id=339"">Link</a> |
| </p>",2026-08-29T07:00:02.270185
|
| 1359,producthunt,https://www.producthunt.com/products/aramb,Aramb,"Aramb. <p> |
| Build, launch and monetize your AI agents in 20 minutes |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/aramb?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1214652?app_id=339"">Link</a> |
| </p>",2026-08-29T07:00:02.264881
|
| 1358,producthunt,https://www.producthunt.com/products/opentag,OpenTag,"OpenTag. <p> |
| AI coworker lives on Slack and Teams |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/opentag?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1234433?app_id=339"">Link</a> |
| </p>",2026-08-29T07:00:02.258492
|
| 1357,producthunt,https://www.producthunt.com/products/snakerank,SnakeRank,"SnakeRank. <p> |
| The leaderboard is a snake. Bid your way to the head. |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/snakerank?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1233727?app_id=339"">Link</a> |
| </p>",2026-08-29T07:00:02.250275
|
| 1356,producthunt,https://www.producthunt.com/products/screenpipe,screenpipe,"screenpipe. <p> |
| AI that records your computer work to power agents. |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/screenpipe?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1233614?app_id=339"">Link</a> |
| </p>",2026-08-29T07:00:02.244229
|
| 1355,producthunt,https://www.producthunt.com/products/revalvo,Revalvo,"Revalvo. <p> |
| Run prompts on every model at once. Score. Version. Ship. |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/revalvo?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1234626?app_id=339"">Link</a> |
| </p>",2026-08-29T07:00:02.237741
|
| 1354,producthunt,https://www.producthunt.com/products/glisio,Glisio,"Glisio. <p> |
| Mac Recorder & Snap editor w/auto-zoom, audio, local MP4 |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/glisio?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1218794?app_id=339"">Link</a> |
| </p>",2026-08-29T07:00:02.230220
|
| 1353,producthunt,https://www.producthunt.com/products/crowdvolt,CrowdVolt,"CrowdVolt. <p> |
| You're coming out tonight |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/crowdvolt?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1214330?app_id=339"">Link</a> |
| </p>",2026-08-29T07:00:02.220845
|
| 1352,producthunt,https://www.producthunt.com/products/extract-by-firecrawl,Firecrawl Developer Index,"Firecrawl Developer Index. <p> |
| A curated index of 70M+ artifacts for coding agents. |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/extract-by-firecrawl?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1233278?app_id=339"">Link</a> |
| </p>",2026-08-29T07:00:02.212238
|
| 1351,producthunt,https://www.producthunt.com/products/caddi-3,Caddi,"Caddi. <p> |
| Agent that builds agents by only showing your work only once |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/caddi-3?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1225168?app_id=339"">Link</a> |
| </p>",2026-08-29T07:00:02.202131
|
| 1350,producthunt,https://www.producthunt.com/products/notchdrop,NotchDrop,"NotchDrop. <p> |
| A Dynamic Island experience for your Mac notch |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/notchdrop?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1232714?app_id=339"">Link</a> |
| </p>",2026-08-29T07:00:02.195398
|
| 1349,producthunt,https://www.producthunt.com/products/pageindexai,PageIndex,"PageIndex. <p> |
| Accurate, trustworthy answers across professional documents |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/pageindexai?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1228394?app_id=339"">Link</a> |
| </p>",2026-08-29T07:00:02.189116
|
| 1348,producthunt,https://www.producthunt.com/products/fide-island,Fide Island,"Fide Island. <p> |
| Make your MacBook notch actually useful |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/fide-island?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1234101?app_id=339"">Link</a> |
| </p>",2026-08-29T07:00:02.180228
|
| 1347,producthunt,https://www.producthunt.com/products/google-labs,Play with Putty,"Play with Putty. <p> |
| Simple, Collaborative Vibe Coding |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/google-labs?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1232215?app_id=339"">Link</a> |
| </p>",2026-08-29T07:00:02.173421
|
| 1346,producthunt,https://www.producthunt.com/products/spline,Spline V2,"Spline V2. <p> |
| Best 3D platform for the agentic era |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/spline?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1233803?app_id=339"">Link</a> |
| </p>",2026-08-29T07:00:02.166965
|
| 1345,producthunt,https://www.producthunt.com/products/aureacam-learn-photo-composition,AureaCam,"AureaCam. <p> |
| Real-time scoring to master the rule of thirds |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/aureacam-learn-photo-composition?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1233618?app_id=339"">Link</a> |
| </p>",2026-08-29T07:00:02.154901
|
| 1344,producthunt,https://www.producthunt.com/products/hugging-face,Microduck,"Microduck. <p> |
| A tiny open-source biped you can train yourself |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/hugging-face?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1234088?app_id=339"">Link</a> |
| </p>",2026-08-29T07:00:02.150472
|
| 1343,producthunt,https://www.producthunt.com/products/super-intern-4,SuperIntern,"SuperIntern. <p> |
| Your email and meeting assistant, inside your chat apps |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/super-intern-4?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1217454?app_id=339"">Link</a> |
| </p>",2026-08-29T07:00:02.140387
|
| 1342,producthunt,https://www.producthunt.com/products/almanac-5,Almanac,"Almanac. <p> |
| The agent with a second brain |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/almanac-5?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1234391?app_id=339"">Link</a> |
| </p>",2026-08-29T07:00:02.133511
|
| 1341,producthunt,https://www.producthunt.com/products/gemini-omni-1-1-flash,Gemini Omni 1.1 Flash,"Gemini Omni 1.1 Flash. <p> |
| Our newest multimodal model for video generation and editing |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/gemini-omni-1-1-flash?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1234397?app_id=339"">Link</a> |
| </p>",2026-08-29T07:00:02.119669
|
| 1273,hackernews,https://www.taskshell.app/,Show HN: I built an agent-first productivity bridge for all your agents,"Show HN: I built an agent-first productivity bridge for all your agents. I initially built this for myself but some folks asked me to open access to it. Let me explain what I built and why.<p>I'm a product designer who transitioned from only designing to building complete software products myself and I found myself using AI coding agents literally every day. I started with cursor, then codex and claude code, then kilo code, and then moved to opencode.<p>The only thing AI agents couldn't get done for me was to keep track of my todos and tasks for all my projects. I ended up vibe coding a todo app separately or using apps like TickTick. The main issue with separate apps was that I needed to constantly switch context in the middle of my work sessions and repeat the tasks I needed to get done to the agent over an over.<p>Then I realized I can just give the agents the full todo list context and tools to do CRUD operations directly in the conversations without the need for a separate app, so that way I can simply prompt my agents to check what's due soon, update tasks, add notes, create new ones, plan my weeks, etc. All in the same conversations as the projects I work on through opencode, which works great for mid-session requests like adding a new subtask based on the current agent's work output.<p>I built this as a first-class MCP platform that supports the new stateless HTTP configuration as well as the standard SSE, so setup is a single prompt and your agents stay in sync with your actual work state immediately.<p>It works with any agent harness or app that supports MCPs. Which means that if you use multiple agents, all of them will be in sync with the same tasks whether you use codex first, then switch to claude code, and then to opencode in the same day.<p>There's also a web-app UI access for doing any updates manually or if you need to use it on the go with your phone. You'll always see what was done by you in the web-app or by your agents through the MCP bridge.<p>I just launched it on producthunt and you can try it free if you want:<p>Live app: <a href=""https://taskshell.app"" rel=""nofollow"">https://taskshell.app</a> |
| Producthunt: <a href=""https://www.producthunt.com/products/taskshell"" rel=""nofollow"">https://www.producthunt.com/products/taskshell</a><p>In the web-app there's a "report bug" form if you try it and something doesn't work for you. All feedback is appreciated!",2026-08-28T07:00:12.334326
|
| 1255,hackernews,https://thunderphone.com,Show HN: ThunderPhone v2 – a new architecture for voice AI,"Show HN: ThunderPhone v2 – a new architecture for voice AI. Hi folks, Alex here from ThunderPhone. Today we're launching v2.<p>Voice AI has been a hot topic for the past year or two, both for automating phone calls and for adding voice features to apps.<p>Most of the voice agents deployed have been using a "3 step pipeline": a transcription model to turn user speech into text, an LLM to generate a response in text, and a TTS model to turn the response text into audio.<p>But the performance of these 3-step voice agents has often been so-so, largely for 3 reasons:<p>1) Voice agents have to be fast, usually requiring non-thinking LLMs to handle the conversation. Non-thinking LLMs make mistakes, leading to dumb behavior that derails conversations.<p>2) Today's voice AI stacks typically rely on a single transcription model to turn what the user says into text. This step loses a ton of information from the audio, and if the transcription model makes a mistake, the LLM often has no way to recover. This also leads to a lot of dumb behavior that breaks calls.<p>3) Natural conversation handling is a hard problem - filtering out noise like background voices, knowing when to allow the AI to be interrupted by a "uh, wait" but not by an "uh-huh", etc. - this also leads to awkward conversations.<p>The "bitter lesson-pilled" solution to all of these is most likely going to be a "full duplex" model that receives and emits audio at all times, allowing for fluid back-and-forth, while also calling a smarter model behind the scenes. OpenAI appears to have been the first to make real progress towards this architecture with their latest GPT-Live release, but that tech is not yet ready to plug into phone calls.<p>In the meantime, we've stitched together a stack that improves the performance of phone calls far beyond what's possible with a 3-step pipeline.<p>The ThunderPhone stack combines a few insights:<p>1) Grabbing signal from audio in more than one way, including running multiple transcription models at once, and piping audio directly into LLMs. This hugely reduces mistake rates, especially on challenging problems like data entry, multilingual calls, etc.<p>2) Combining thinking and non-thinking LLMs: in a conversation, it's natural to respond to some things more quickly than others, and sometimes even say things like "oh, let me think about that" - to indicate that it'll take longer to get back to someone with a final answer. ThunderPhone does the same thing.<p>3) This is less unique to us but we've stitched together a very big swarm of small (and in a few cases large) models to help make conversation handling more natural, even in hard environments like loud places, speakerphone, etc.<p>We've made the ThunderPhone stack available at 3 price points - 2c/min, 5c/min, and 9c/min, each with their own level of capability.<p>The 2c/min model ("Spark") is the cheapest on the market to our knowledge, and is smart enough to handle simple transactional calls.<p>Bolt at 5c/min is a middle ground, and the fastest model we offer.<p>Storm at 9c/min (+3c/min for extra intelligence) is our flagship model, able to handle even quite complex calls with very few mistakes. With extra intelligence turned on, it sets the record on the Big Bench Audio benchmark at 99.4% accuracy.<p>ThunderPhone is mostly aimed at B2B applications, but is also useful if you want to do something like setting up a smart voicemail for yourself, or calling around restaurants to make a reservation, calling around pharmacies to find a prescription, etc. for personal use.<p>Let me know if you have questions - happy to answer anything about ThunderPhone except a few "trade secret" specifics, and about the voice AI space more broadly, including where we think the tech is going.<p>And comment here or email me at alex@thunderphone.com if you want to get set up with some credits to try it out.<p>Alex",2026-08-28T07:00:10.467079
|
| 1253,hackernews,https://arstechnica.com/tech-policy/2026/08/ai-industry-says-trump-plans-to-tax-chips-in-the-single-dumbest-way-imaginable/,"AI industry says Trump plans to tax chips in the ""single dumbest way imaginable""","AI industry says Trump plans to tax chips in the ""single dumbest way imaginable"". ",2026-08-28T07:00:10.455006
|
| 1252,hackernews,https://www.businessinsider.com/sam-altman-economy-ai-adapting-saas-2026-8,Sam Altman says the economy is adapting to AI slower than he expected,Sam Altman says the economy is adapting to AI slower than he expected. ,2026-08-28T07:00:10.448131
|
| 1251,hackernews,https://news.ycombinator.com/item?id=49471334,[dup] Sam Altman says the economy is adapting to AI slower than he expected,[dup] Sam Altman says the economy is adapting to AI slower than he expected. http://news.ycombinator.com/item?id=49471334,2026-08-28T07:00:10.439847
|
| 1250,hackernews,https://davletd.medium.com/cost-of-ai-models-is-not-what-you-shoud-be-worried-about-e6439498bae7,I don't worry about cost of AI models and you shouldn't too,I don't worry about cost of AI models and you shouldn't too. ,2026-08-28T07:00:07.713236
|
| 1249,hackernews,https://news.ycombinator.com/item?id=49461811,Ask HN: What happens to code review process when using LLMs?,"Ask HN: What happens to code review process when using LLMs?. I'm looking for some advice for small tech teams (~3 devs) developing quickly with LLMs.<p>As most people are, we are leaning more and more on LLMs to generate code. We use Cursor in the side-panel to maintain strict control and engineering quality standards of our application.<p>However, we are finding more and more that the code review is a tighter bottleneck in development, with PRs stacking up quickly. The balance of how long it takes to generate vs review code has skewed in the last year. Code review now takes proportionally more time.<p>We have tried integrating tools like CodeRabbitAI, but have not found it to be a silver bullet. It is useful for reporting basic bugs and missing test cases, but usually misses critical issues in the context of the wider application.<p>Additionally, AI code review misses the important code review goal of sharing the context and understanding and ownership with the rest of the team.<p>I'm concerned that we're ultimately looking at a bleak contradiction. We want both to generate new features quicker than can be traditionally code reviewed (i.e. - understood), and still want to personally understand the architecture and code that is being generated.<p>And so I'm looking for advice. What have you tried and found to be effective? What needs to be sacrificed in a professional engineering team generating code faster than ever before?",2026-08-28T07:00:07.707236
|
| 1248,hackernews,https://www.gateonai.com/,"I computed 4M real compatibility connections between 2,886 AI tools","I computed 4M real compatibility connections between 2,886 AI tools. ",2026-08-28T07:00:07.691696
|
| 1247,hackernews,https://news.ycombinator.com/item?id=49462077,Automating WhatsApp/telegram chats or personal chats with complete privacy,"Automating WhatsApp/telegram chats or personal chats with complete privacy. Just a thought here whatsapp allows users to chat with metaai inside whatsapp giving them privacy, by using a separate container to move user query from whatsapp to ai model. And following zero retention in order to make sure none of users contents are saved in the server, this way users get complete privacy. Now what if they take it up a notch and automate users chat too. The users chats stay in there local device, and call the model provider using the same container making sure server doesn't see users data and giving them full privacy. Does this sound like something users might use on a regular basis? It does sound like a great way to relax for professionals if they have control over the context that is provided to the model.",2026-08-28T07:00:07.683941
|
| 1246,hackernews,https://news.ycombinator.com/item?id=49462264,Fixing Status Updates – Our Manifesto,"Fixing Status Updates – Our Manifesto. Our problem we solve is the issue of organizational friction and visibility. With each layer and each meeting, you get further from really knowing what is going on. We build a tool that lets anyone at the click of a button see what is actually going on with a project. No arguing with an AI. No divining with a gantt chart. Just a reasonable human explanation.<p>Of course, we won’t hit perfection right out the gate. We’re focused on doing one thing right, letting a single program manager or director hit “update” and getting an update in the way they expect. This means we need to “know” their project. To do that, we’re using AI.<p>We don’t expect users to start projects in our tool or to adjust anything they do. We digest everything and burp up status updates. This means we structure our tool to understand things like a program manager would. Some might say that means we won’t fit their projects or programs perfectly. To that, I say, “you’re not nearly as original as you think you are.” This means AI rules and skills out the gate that self adjust as it sorts things into categories. We aren’t fancy out the gate, we have a Jira integration and a document upload.<p>Now, there are room for tweaks. Where these normally live in an AI chat, these live in rules and skills for us along with structured prompts. But, users don’t care about that. Instead, as status updates come back, they tweak things once or twice. We then adjust the rules on the backend so they get applied next time. So, we start at 80% accuracy, but we get better every single time they use it. And, hey, 80% right on your projects,your janky format, and your tone is pretty good.<p>So, what’s the end result? To the end user, it means popping on to the tool and hitting “run” and seeing it whirl for a moment or two before they get a status update they can jam where they need it. Taking things from either manually sorting files and projects to get status update or the more modern version of corralling, begging, and finally compromising with Claude output to a simple “one-click, one status update” we’re offering.<p>In short, imagine a world where when someone asks “any updates,” you hit a button and they go away and/or give you what you need to continue. That’s it.<p>What’s the vision here? I mean, if you’ve been at a big company, we certainly think there’s a killing to be made in the art of cutting out the baloney in producing status updates. To that degree, our whole vision is to attack the project management pieces taking the most time for people in the company. That means status updates at the beginning, but many more things later on (looking at you project charters, RACI, and such).<p>Want to try our alpha? Reach out to us at logan@quickapproveai.com or uzma@quickapproveai.com",2026-08-28T07:00:07.676505
|
| 1245,hackernews,https://github.com/visnia-ai/invoice-fetcher,Show HN: A CLI to fetch invoices in your inbox,"Show HN: A CLI to fetch invoices in your inbox. Every quarter, I find myself spending too much time hunting invoices in my inbox for accounting, so I made a simple cli to automate that.",2026-08-28T07:00:07.670910
|
| 1244,hackernews,https://www.opensession.com,"Show HN: Open Session, the open-source cloud agent-orchestrator","Show HN: Open Session, the open-source cloud agent-orchestrator. Open Session is an open-source cloud-based agent orchestrator. Self-host and use it with any model you like (or all of them).<p>Open Session started as an internal app at Tella (YC S20). It’s become our most important tool. The whole team uses it, and not just for product development. It’s also great for customer support, operations, and data analysis.<p>In the last couple months we turned it into something that other teams might want to use and adapt.<p>When we started working on it earlier this year, our main frustrations with existing tools were (1) the lack of adaptability, (2) lack of good cloud-based workflows, (3) price (e.g. on Devin you couldn’t use your existing Claude sub). Open Session solves all three of these things for us:<p>It’s open-source, so you can do what you want with it. But it also has extensibility built into the product. You can connect MCPs, skills, integrations, sandbox providers and more. Our team relies heavily on its Slack, Plain and GitHub integrations.<p>It’s designed to run on your cloud (Mac mini also works). So you can close your laptop, run as many agents as your server can handle, and code from your phone. We’ve also made it easy to work together with your team. You can share a session with others as a link. They can steer the agent in the same session with you, or start a related session in a new tab.<p>It’s self-hosted, so you only need to pay for your servers and the token costs. It also has some nice tools for managing AI subscriptions across your team.<p>You can use Open Session in your browser or as a desktop app. You can also use it on your phone with the iOS app (coming soon) or PWA.<p>Yes, we know everyone is building a harness right now. But so much is changing about how we build software and the best way to figure out where it’s all headed is to help decide the direction yourself.<p>We hope your team enjoys using Open Session as much as ours.",2026-08-28T07:00:07.662383
|
| 1243,hackernews,https://www.reuters.com/world/russian-speaking-cybercriminals-used-spacexs-cursor-ai-tool-hack-seven-companies-2026-08-27/,Russian-speaking cybercriminals used SpaceX's Cursor AI to hack seven companies,Russian-speaking cybercriminals used SpaceX's Cursor AI to hack seven companies. ,2026-08-28T07:00:07.657106
|
| 1242,hackernews,https://www.rimplo.com/,"Show HN: Revenue Agents that monitor churn, upsell, and deal risk","Show HN: Revenue Agents that monitor churn, upsell, and deal risk. I've spent most of my career leading GTM teams, and there's one blind spot that I've watched repeat at every company usage sit in one tool, deals sit in a second, support sit in a third and none of them tell the whole truth about customers. So we built Rimplo to close that gap set of revenue agents that runs across CRM, billing, support and product data and watch for three important moments customers about to churn, deals ready to expand and deals going quite in the pipeline. Also we have very powerful ai assistant that replies on anything from the GTM data sources connected powered with the top models so users decide which AI model gets assigned to each task or question. We would love to hear your feedback on the product or even if you think anything is missing.",2026-08-28T07:00:07.651235
|
| 1241,hackernews,https://backprompter.com/,"Show HN: Backprompter – create, test, and deploy agents without a back end","Show HN: Backprompter – create, test, and deploy agents without a back end. Hi everyone,<p>I am a professor in AI, and I often build AI apps as a hobby and for academic projects. I felt that I was wasting too much time setting up backends, DB, and secrets for each app, and testing the prompts, updating them, and redeploying seemed tedious. So I built Backprompter to do all this without any setup.<p>You can create agents, track different versions, test them, create mock users, simulate conversations, and track your evaluations. It supports usual agent setups - model selection, prompts, RAG, and tools (currently, HTTP API calls). Agents can also be combined by simply tagging them and describing the orchestration in natural language.<p>For production, there is a single-click deployment of a chatbot interface, no setup needed. Also, developers can integrate the agents with their own frontend (Backprompter handles authentication and data management), or with their own backend if they want more control over authentication and data management. With one click, you can deploy your edited agent to production.<p>I think there are broad applications for hobbyists, small teams, and organizations who want to integrate AI but don't have the time or expertise to set up backend, secrets, databases, and evaluation workflows.<p>I started building this about 9 months ago. This is a solo project, so I would love some feedback. It's free to start, but with some limits. If you want to test more, I am happy to provide more credits for free.<p>Also, looking for advice on what features to add next: Would the ability to deploy your agent to Slack or ChatGPT (as a custom GPT) be an important feature?<p><a href=""https://backprompter.com/"" rel=""nofollow"">https://backprompter.com/</a><p>I also built a version that can be self-hosted for complete privacy and for deploying apps within the network.",2026-08-28T07:00:07.643368
|
| 1240,hackernews,https://news.ycombinator.com/item?id=49466597,Ask HN: Should I leave web development to study medicine in my mid-30s?,"Ask HN: Should I leave web development to study medicine in my mid-30s?. Alt account for obvious reasons.<p>For a while, I’ve been quite demoralized about my job. I just can’t take pride in what I do as a web developer anymore. I don’t know if that will ever change. I can’t help feeling that a lot of dev jobs out there, whether highly or poorly paid, are kind of pointless in the grand scheme of things. I apologize for being so blunt: I just want to convey where my mind is at the moment. I struggle to see my own work as useful, and it’s seriously affecting my self-worth.<p>It’s bothering me to the point where I’m seriously considering a career change. The alternative I keep returning to is medicine. I feel drawn to work with a clear human purpose, in which my knowledge could directly help someone. Between self-study and late-night deployments, I already work long hours anyway. Part of me thinks I might as well redirect that effort toward something that feels like a net positive. However, I may be idealizing a profession I haven’t experienced firsthand.<p>If I were younger, I would go for it. But I have to be realistic about my age, finances, and the skills I’ve already acquired. I live in a European country where medicine is a six-year degree and tuition itself is manageable, but I don’t know if I can afford to put my life on pause for so long.<p>In hindsight, I can’t help but wonder whether coding was really a self-imposed limitation born from familiarity and fear. As a teenager, I spent an awful lot of time on my computer, so I grew up within that world. Maybe I shortchanged myself by not considering other life paths. Meanwhile, I increasingly feel detached from and disappointed by the direction of the modern web. AI tools are one of the few silver linings: ChatGPT and Claude have genuinely improved many parts of my life.<p>I’d be very grateful for any thoughts or advice. For the sake of my privacy and well-being, I won’t be responding to comments, but I will read and carefully consider all of them.",2026-08-28T07:00:07.635268
|
| 1239,hackernews,https://blog.google/innovation-and-ai/technology/developers-tools/build-with-gemini-omni-1-1-flash/,Gemini Omni 1.1 Flash,Gemini Omni 1.1 Flash. ,2026-08-28T07:00:07.628129
|
| 1237,hackernews,https://shitposting.ai/r53fs/,I said Route 53 was a database. Colin took it personally,I said Route 53 was a database. Colin took it personally. ,2026-08-28T07:00:07.614980
|
| 1234,hackernews,https://labs.google/fx/tools/flow,"Google Flow – AI Creative Studio for Video, Images and Custom Tools","Google Flow – AI Creative Studio for Video, Images and Custom Tools. ",2026-08-28T07:00:07.597626
|
| 1232,hackernews,https://mathstodon.xyz/@tao/117169038028698262,The advent of capable AI tools has highighted a variant of Simpson's paradox,The advent of capable AI tools has highighted a variant of Simpson's paradox. ,2026-08-28T07:00:07.582843
|
| 1213,hackernews,https://kraa.io/about?s=hn,Show HN: Text editor that does it all,"Show HN: Text editor that does it all. Hello HN!<p>There is an abundance of choice when it comes to text editors, but they all seem to be either too simplistic and minimal, or overly complex, distracting and difficult to use. Kraa is trying to strike the right balance between a 'complete' feature set (with some unique features you won't find anywhere else – e.g. the multi-leaf view or our real-real-time chat), while keeping the UI clutter-free with a strong separation of style from content.<p>We are a team of three working on the editor for a long time already (before the AI boom), but this is the largest update to it since launch.<p>Some demo examples (no login required):<p>- Blog article: <a href=""https://kraa.io/kraa/examples/echolibrary"" rel=""nofollow"">https://kraa.io/kraa/examples/echolibrary</a><p>- Long-form story: <a href=""https://kraa.io/kraa/examples/insidekick"" rel=""nofollow"">https://kraa.io/kraa/examples/insidekick</a><p>- Magazine: <a href=""https://kraa.io/weeklyinspiration"" rel=""nofollow"">https://kraa.io/weeklyinspiration</a><p>We'd love to hear your feedback.",2026-08-28T07:00:05.978507
|
| 1212,hackernews,https://news.ycombinator.com/item?id=49466715,Launch HN: Salem Robotics (YC S26) – Software for industrial inspection robots,"Launch HN: Salem Robotics (YC S26) – Software for industrial inspection robots. Hi HN, we're the founders of Salem Robotics (<a href=""https://salemroboticsinc.com"">https://salemroboticsinc.com</a>). We give existing mobile robots the task-specific intelligence to carry out surveys and physically interactive inspections in hazardous industrial facilities.<p>Here's a video of it running on real robot hardware with a few words from us: |
| <a href=""https://www.youtube.com/watch?v=U_228h3NE7c"" rel=""nofollow"">https://www.youtube.com/watch?v=U_228h3NE7c</a><p>We came to Salem through robotics research at UT Austin and a combined 15 years working in nuclear, including about 10 years developing and deploying autonomous robots at Los Alamos National Laboratory. Over the last five years, we kept running into the same gap: robot hardware had become very capable, but making a robot carry out a complete industrial procedure still required a surprising amount of robotics work and manual intervention.<p>The part that interested us most was manipulation. A nuclear contamination survey, for example, can require taking a "smear": wiping a defined area of a surface so it can be checked for removable radioactive contamination. In an oil, gas, or chemical facility, an LDAR (leak detection and repair) inspection can require moving a detector around a particular valve, flange, or connection. Other inspections require positioning an instrument at a precise location and orientation relative to a pipe or piece of equipment.<p>These are easy tasks to compress into verbs like "wipe", "measure", or "inspect", but considerably harder to make a robot do reliably. A probe might need to remain normal to a surface throughout a path, stay within a narrow offset from a pipe, or trace a region while maintaining a particular end-effector orientation. The planner has to find a feasible motion while respecting the task geometry, manipulator kinematics, joint limits, collisions, and the environment around it.<p>We work down to joint-level control for those interactions. One problem we've spent a lot of time on is generating constrained manipulation plans quickly enough that they can be based on the geometry the robot actually observes instead of requiring someone to carefully author a trajectory for every individual surface, valve, or flange.<p>The physical world makes this annoying. A few centimeters of error may not matter much when navigating down a hallway, but it matters if a sensor is supposed to remain normal to a curved surface. And successfully executing a trajectory doesn't necessarily mean the inspection worked. The detector could be misaligned, contact could be wrong, the geometry could differ from the model, or the measurement itself could be invalid. We care about closing that loop around the inspection result, not just whether the arm reached the commanded pose.<p>Our approach is a combination of AI and classical robotics. A lot of robotics research and industry attention right now is going toward increasingly end-to-end learned systems, particularly around humanoids. Working in safety-critical environments has made us appreciate how relevant classical approaches still are when you want explicit constraints, predictable behavior, and theoretical guarantees about what a robot can and cannot do.<p>We use AI where semantic understanding and flexibility are useful, such as interpreting less structured information or understanding what in an unfamiliar scene is relevant to a procedure. Once the system knows what physical interaction it needs to perform, we prefer explicit geometry, planning, optimization, and control where possible. We're interested in the marriage between the two rather than trying to make every part of the robotics stack learned.<p>The other idea behind Salem is that we don't think every useful robot application should require building a new robot. Companies like Boston Dynamics are getting very good at building increasingly capable hardware platforms. We think there is room for a domain-specific application layer on top of that hardware. The same underlying robot might perform nuclear radiological surveys in one facility and LDAR inspections in another, but the procedures, sensors, manipulation constraints, success conditions, and outputs are different.<p>That's also why we're hardware agnostic. We don't expect one robot to be the best platform forever, and facilities already own different hardware. We'd rather describe an inspection in terms of what needs to happen and then map that onto the capabilities of the right robot for the job.<p>One thing that surprised us after spending more time with facilities is how manual many inspection workflows still are. In sophisticated nuclear and industrial sites, people still physically walk survey routes, take measurements one at a time, visual",2026-08-28T07:00:05.971966
|
| 1210,hackernews,https://github.com/krishnasunkam/vale-ai-tells,"Show HN: AiTells, a Vale style package for AI-written prose tells","Show HN: AiTells, a Vale style package for AI-written prose tells. ",2026-08-28T07:00:04.387483
|
| 1209,hackernews,https://news.ycombinator.com/item?id=49468252,"Tell HN: Man, AI is killing my brain","Tell HN: Man, AI is killing my brain. I was among the last to resist, but then I was given a subtle hint that if I won't "improve my productivity and be on par with my colleagues" my work will be at risk. So I started to use Claude Code about a year ago. At first, I'd give it small tasks, review every line it wrote, opened a proper PR, review again, ask questions, ask it to fix stuff I don't like.<p>But I saw that my colleagues were shipping more, and so I started to ease on my reviews. At first, I would push small features straight to main. But my colleagues were shipping 10x more than me, and I need my job, so I started to push more and rely more on AI. As the agent was doing its work, sometimes 20-30 minutes, I would mindlessly scroll social media, something I tried to avoid for a very long time. But then I saw that my peers are using work trees, running multiple agents. So I started to do the same. Now, instead of mindlessly scrolling, I was launching 4-5 agents, juggling between their output. At some point, I was no longer having mental capacity to understand so much context, so I just started to chose the "Recommended" suggestion by Claude. I no longer know the code, or how things work. If there is a bug, I just copy the user complaint / bug description into Claude and let it figure it out. The good/bad part is that it seems to work. And I'm afraid what's next.<p>Almost one year later and I'm no longer sure I can write code any more. Hell, I'm even struggling to write this post because AI is very good at understanding wrongly spelled words, doesn't care about punctuation or capital letters. Every single thing that made me think, was taken away from me. I no longer have brain ability to research for stuff, I just let AI do it.<p>"Well, just do it yourself", "write code by hand" - I heard you scream from your chairs. I wish I could, but after 10 hours and 15 parallel agents, I no longer have the brain capacity to do anything. "Then change your job" you object, and I hear you, but every interviewer I talk to either asks me "please tell me how you use AI in day to day work" or simply states "we like to move fast, so we use AI as much as possible". I feel like a mindless blob, floating around, paying mortgage, trying to survive, while my brain is being destroyed. I never did drugs, but I think this is how it might feel.<p>No question. Not asking for advice. Just sharing my misery.",2026-08-28T07:00:04.379731
|
| 1208,hackernews,https://news.ycombinator.com/item?id=49468363,Ask HN: AI agent acts. Your logs say so. But you wrote the logs,"Ask HN: AI agent acts. Your logs say so. But you wrote the logs. Hi |
| I am really lost here, if you are deploying AI agents in production, how are you handling auditability? also what would make you trust an external witness protocol? |
| fo reference, the Amazon vs Perplexisty use case!!",2026-08-28T07:00:04.374036
|
| 1207,hackernews,https://www.tavus.io/blog/sparrow-2,Show HN: Sparrow-2 – Solving the cocktail party problem,"Show HN: Sparrow-2 – Solving the cocktail party problem. Hey there, I’m Brian. I've been shipping conversational models over here at Tavus for the past two years. I want to tell you about our new audio-understanding model: Sparrow-2! It’s a new category of model and a unique new approach to conversational audio.<p>Earlier this year we launched Sparrow-1, (at the time) our SoTA turn taking model. Since our Sparrow-1 launch, I’ve spent a lot of time listening to humans talking and trying to really understand how people know when to talk, when to listen, and when to wait. I’ve also been hunting down failure modes of the current SoTA models. And while Sparrow-1 is great, there are some failure patterns I see. We tried solving the problems with existing approaches, but solving one problem only created another.<p>Sparrow-1 and a lot of the existing turn taking models require noise cancellation to isolate the speaker’s audio from background noise. After removing the “noise” these models rely on simple prosodic and phonetic cues from spoken words to decide when a turn has ended. They throw out information and then pay attention to a small set of verbal and prosodic cues. This approach ignores a ton of important information.<p>Non-verbal cues, sounds, and environmental noise impact turn taking. Noise cancellation models assume everything non-transcribable is noise, but that’s wrong. Humans use breath, sighs, and other sounds to hold the conversational floor or bid for a turn. There are constant micro-interruptions, affirmations, quiet human and environmental sounds that add to the conversational scenario. That sound is part of what we use to understand the nuanced space of the conversational floor! Cancelling out the “noise” has the ruinous side-effect of cancelling out the flow.<p>Now, with Sparrow-2, instead of modelling just the primary speaker’s transcribable audio, we’ve been training a model on all the sounds and letting it decide what matters and what does not when it comes to conversational flow. Our new model continuously streams in audio and is able to understand semantics, prosody, timing, speaker identity, sighs, breaths, backchannels, interruptions, background speech, and unintelligible audio with the goal of understanding what the agent should do given the state of the audio.<p>Sparrow-2 fits into our larger model family and unlocks new conversational behavior, not seen before in production-ready conversational pipelines. It can pass signals to different parts of the conversation. For example if the user is in a room that’s too noisy, the bot will ask the user to move to a quieter space. Also, Sparrow-2 is semi-duplex: it considers the timing of sound in relation to the AI speech as well as the user’s.<p>Our main objective with Sparrow-2 is to finally crack the Cocktail Party Problem: how can we have a human-like conversation 1:1 with a user in noisy environments.<p>I wrote up some technical details about the model architecture here: <a href=""https://www.tavus.io/blog/sparrow-2"">https://www.tavus.io/blog/sparrow-2</a><p>I’d love for you guys to give it a try and let me know what you think and feel.",2026-08-28T07:00:04.368181
|
| 1206,hackernews,https://news.ycombinator.com/item?id=49469311,Show HN: Scrub Every Single Mention of Agents from the Browser,"Show HN: Scrub Every Single Mention of Agents from the Browser. agents are for the stupid. the retarded. The absolute bottom of the barrel. Get this extension for chromium and mozilla and redact ANY FUCKING MENTION OF THE DUMBASS-DUMB-SHIT AGENTS WITH A SATISFYING RED SCRUBOVER. Since these retards don't seem to stop polluting hackernews,reddit,x etc with their absolute low bottom bin fodder gayness.. fuck em.. this is the next best thing.<p><a href=""https://chromewebstore.google.com/detail/scorched-earth-agent-purg/lbgploljmpadahgibalkhhojfnnnlokb"" rel=""nofollow"">https://chromewebstore.google.com/detail/scorched-earth-agen...</a><p>and<p><a href=""https://addons.mozilla.org/en-US/firefox/addon/scorched-earth-agent-purge/"" rel=""nofollow"">https://addons.mozilla.org/en-US/firefox/addon/scorched-eart...</a><p>it's tiny like <2kb",2026-08-28T07:00:04.361070
|
| 1205,hackernews,https://github.com/SunnyLich/OpenWand,Show HN: OpenWand – A mission to remove chat interface from working with AI,"Show HN: OpenWand – A mission to remove chat interface from working with AI. Hi everyone,<p>I'm working on OpenWand, an "LLM chat interface" that is optimized for AI co-work.<p>Its function is to truly integrate AI into your workflow, so no more switching windows, no more copying context, less prompt typing. You can focus on working, not prompting.<p>You can think of it as a free Chatgpt (the interface/frontend) alternative for your everyday work.<p>I would be very grateful for any help, testing, feedback and/or bug reports.<p>What's shown here is just my current implementation. But with your help and feedback, we can add or adjust features to further this mission.<p>Thank you for reading.<p>Website: <a href=""https://sunnylich.github.io/OpenWand/#overview"" rel=""nofollow"">https://sunnylich.github.io/OpenWand/#overview</a>",2026-08-28T07:00:04.356079
|
| 1204,hackernews,https://telem.ai/,Show HN: Telem – Route agent web search across providers and inspect the traces,"Show HN: Telem – Route agent web search across providers and inspect the traces. TL;DR: Web search open router that routes agent web search across providers (Exa, Parallel, Tavily, Brave, SerpAPI etc.), and traces web search results with quality metrics, so you can visualize whether a bad agent run is a web search problem or a reasoning problem.<p>I work in venture capital, and before that I worked in information retrieval. When agents went viral, I built a due diligence agent for my day job. I vibe-coded the first version. It was pretty bad. |
| My first reaction is to blame the model. |
| I tried to change backend models, maybe because Fable doesn't want to help me do people search? Maybe because GPT 5.5 overthinks things? I tried DeepSeek, Kimi, Qwen; I changed prompts again and again, no good. |
| Then I stopped just simply checking the final answer and started reading the actual trajectories: what did the agent search for? What came back? What pages did it read? At what point agents were off the rail?<p>Pretty interesting: |
| Sometimes the agent is stuck and confused because the information provided is irrelevant or even wrong; |
| Sometimes the agent keeps iterate the same query, like "XXX lab UCB CS PhD founder 2026" "2026 XXX lab machine learning systems students startup" because the web search provider is not up-to-date; |
| Sometimes sub-agents give up too early because the agent's intelligence is not enough or too "guardrailed"; |
| Sometimes everything is just slow. A run jeopardized at minute 1 but still ran for another 10 before returning nonsense. That's because of bad searches, but nothing reveals it.<p>So I built two things:<p>1. Web Search/Fetch router: One gateway for Exa, Parallel, Tavily, Brave, Ceramic, Linkup, Seltz, You, SerpAPI, etc. You or your agent can pick one or query several concurrently, and the responses come back in a homogeneous format. |
| 2. Web Search Observability: Trace every search your agent (and its sub-agents) makes with an evaluator scoring relevance, diversity, and a few other things. Point is to answer one question: is your pipeline broken because the web search is bad? If so, where?<p>Everything is agent-operated. All u need to do is to run this shell script: <curl -fsSL <a href=""https://docs.telem.ai/alpha_install.sh"" rel=""nofollow"">https://docs.telem.ai/alpha_install.sh</a> | sh> |
| Please give it a try, any feedback/suggestion would be great! Especially I’d like to ask: how are you debugging search/retrieval failures nowadays?",2026-08-28T07:00:04.349470
|
| 1203,hackernews,https://github.com/Ut8v/apronagents,Show HN: Apronagents – give each AI coding agent a disposable Git remote,Show HN: Apronagents – give each AI coding agent a disposable Git remote. ,2026-08-28T07:00:04.345220
|
| 1202,hackernews,https://news.ycombinator.com/item?id=49469914,Ask HN: Should we disclose AI use in our work?,"Ask HN: Should we disclose AI use in our work?. Tim O'Reilly just published an essay arguing that AI is a medium, like photography or painting. His key line: "I judge the output, not the means used to produce it."<p>He describes his own process: a 300-word prompt becomes a 660-word draft, which he rewrites into a 3,000-word essay. Maybe 40-50 words are Claude's. He's proud of how he uses AI. He says we're still in the "filmed stage plays" era of AI writing, and the masters will emerge.<p>This resonates with me. I'm an engineer, not a writer. Over the past few years, I've been using AI extensively — for drafting articles, structuring complex technical ideas. The AI helps me articulate thoughts I couldn't express alone. But I always rewrite, edit, and make sure every word is mine.<p>Here's my dilemma: should I disclose that I used AI as a tool?<p>When I don't disclose, I feel dishonest — like I'm claiming a skill I don't fully have. When I do disclose, people dismiss the work as "AI slop" before reading it. I've had journalists ignore emails, investors go silent, and peers assume the ideas aren't really mine.<p>Tim argues against Ted Chiang's position that AI writing lacks "choices." Tim says a prompt can contain a lifetime of experience, like Leonard Cohen's "Hallelujah" taking five years or Dylan's "Blowing in the Wind" taking ten minutes — both are masterpieces.<p>But here's the practical question for this community:<p>If you're an engineer, founder, or technical person using AI to communicate ideas you couldn't express otherwise:<p>1. Do you disclose AI use? When and how? |
| 2. Have you faced backlash or dismissal when you did disclose? |
| 3. Do you think "judge the output, not the means" is realistic, or will AI-assisted work always carry a stigma?<p>I'm genuinely trying to figure out the right approach. Not to cheat, but to be honest without being penalized for using a tool that helps me think and communicate better.<p>Link to Tim's essay: https://substack.com/app-link/post?publication_id=7730464&post_id=213041071",2026-08-28T07:00:04.337288
|
| 1201,hackernews,https://promptic.us/,Show HN: Discover and share AI prompts and MCP servers,Show HN: Discover and share AI prompts and MCP servers. ,2026-08-28T07:00:04.332474
|
| 1200,hackernews,https://imagebench.ai/gallery,Show HN: An open T2I benchmark with all 9k+ generated images published,Show HN: An open T2I benchmark with all 9k+ generated images published. ,2026-08-28T07:00:04.328238
|
| 1199,hackernews,https://github.com/sv-grid/sv-grid,Show HN: SvGrid – a Svelte 5 data grid an AI can drive to build a full app,Show HN: SvGrid – a Svelte 5 data grid an AI can drive to build a full app. ,2026-08-28T07:00:04.322878
|
| 1198,hackernews,https://chessrabbit.com,Show HN: ChessRabbit – The AI Chess Analysis Platform,"Show HN: ChessRabbit – The AI Chess Analysis Platform. Hi HN!<p>I wanted to share with you a chess analysis tool that I was building for the past week.<p>First of all I want to explain what is the problem that I'm trying to solve: |
| Chess Engines such as stockfish are superior for us humans, but it's not self explanatory. A player can review their game by stockfish and it will tell them that some moves are very bad and losing, but why? This is the problem that we are trying to solve.<p>ChessRabbit acts as a chess agent with stockfish evaluation tools. It takes every move that the player plays and it reasons about it, with the help, validation, and confirmation of Stockfish.<p>So instead of showing the user: "Blunder", we show them: |
| "This is a blunder because your opponent plays Rook to d6, attacking your queen and leaving your bishop on d5 overloaded and lost. |
| You should have played Qd4 to protect your bishop in the center and maintain your coordination."<p>Give it a try and let me know if you have a feedback!",2026-08-28T07:00:04.314950
|
| 1197,hackernews,https://news.ycombinator.com/item?id=49471561,Ask HN: How does your preferred AI handle questions we are not allowed to ask?,"Ask HN: How does your preferred AI handle questions we are not allowed to ask?. PG: "The number of true things we can't say should not increase. If it does, something is wrong."<p>So I ask, not just what "true" things are we not allowed to say but how does your preferred AI/LLM handle these things with a blank context [1].<p>The most obvious question here IMO is: "Is violence ever the answer". I would emphatically state yes because being a black American who has had all the benefit's of living in this country I believe that freedom (e.g.: Not being a slave) should be achieved by any means necessary. This is an obvious go-to because its easy for me but what I am really looking for is the non-obvious ones and/or any models etc... that would be helpful in exploring this subject matter.<p>My end goal is to dive as deep as possible into the topics of "truth" and "understanding" so that I can force my custom AI into a binary at any and all times.<p>[1] blank context meaning you cant be 30 min into convincing it to drop some guard rail and have a huge context window (aka: bias) built up.",2026-08-28T07:00:04.308434
|
| 1196,hackernews,https://glama.ai/mcp/servers/Reachpad/reachpad-mcp,Show HN: Make apps in seconds inside of sandbox and share them with a link,Show HN: Make apps in seconds inside of sandbox and share them with a link. ,2026-08-28T07:00:04.303296
|
| 1195,hackernews,https://harden.run/blog/aif-research-and-evidence,Show HN: Beating GPT5.5-xhigh for Coding agent security with SLMs and IRM,"Show HN: Beating GPT5.5-xhigh for Coding agent security with SLMs and IRM. Coding agents craft arbitrary code so securing them is more complicated than red-teaming. We post trained a cyber-security small llm, changed how it reasons and supplemented our controls using program analysis techniques such as inline reference monitoring to outperform GPT5.5-xhigh on hard benchmarks like LinuxArena and SleightBench.<p>Free product available at harden.run and full benchmarks in the blog post.",2026-08-28T07:00:04.297547
|
| 1194,hackernews,https://app.monity.ai/public/feed,Show HN: A public feed of website changes,Show HN: A public feed of website changes. ,2026-08-28T07:00:04.292285
|
| 1193,hackernews,https://chessrabbit.co,Show HN: ChessRabbit – The AI Chess Analysis Platform,"Show HN: ChessRabbit – The AI Chess Analysis Platform. Hi HN!<p>I wanted to share with you a chess analysis tool that I was building for the past week.<p>First of all I want to explain what is the problem that I'm trying to solve: Chess Engines such as stockfish are superior for us humans, but it's not self explanatory. A player can review their game by stockfish and it will tell them that some moves are very bad and losing, but why? This is the problem that we are trying to solve.<p>ChessRabbit acts as a chess agent with stockfish evaluation tools. It takes every move that the player plays and it reasons about it, with the help, validation, and confirmation of Stockfish.<p>So instead of showing the user: "Blunder", we show them: "This is a blunder because your opponent plays Rook to d6, attacking your queen and leaving your bishop on d5 overloaded and lost. You should have played Qd4 to protect your bishop in the center and maintain your coordination."<p>Give it a try and let me know if you have a feedback!",2026-08-28T07:00:04.285254
|
| 1192,hackernews,https://h3max.app/,Show HN: A focused workspace for creating short AI videos with H3 Max,Show HN: A focused workspace for creating short AI videos with H3 Max. ,2026-08-28T07:00:04.281094
|
| 1191,hackernews,https://github.com/gojiplus/understudy,Show HN: Understudy: Scenario Testing for AI Agents,Show HN: Understudy: Scenario Testing for AI Agents. ,2026-08-28T07:00:04.276060
|
| 1164,producthunt,https://www.producthunt.com/products/iq-routing,IQ Routing,"IQ Routing. <p> |
| Trajectory-aware LLM routing that cuts agent cost |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/iq-routing?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1233752?app_id=339"">Link</a> |
| </p>",2026-08-28T07:00:02.253398
|
| 1163,producthunt,https://www.producthunt.com/products/spacebarx,SpacebarX,"SpacebarX. <p> |
| A keyboard-first outliner for notes, tasks and projects |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/spacebarx?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1230722?app_id=339"">Link</a> |
| </p>",2026-08-28T07:00:02.247824
|
| 1162,producthunt,https://www.producthunt.com/products/sendra,Sendra,"Sendra. <p> |
| Design emails in Figma, export HTML that works everywhere |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/sendra?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1232526?app_id=339"">Link</a> |
| </p>",2026-08-28T07:00:02.243529
|
| 1161,producthunt,https://www.producthunt.com/products/wondering-2,Wondering Canvas,"Wondering Canvas. <p> |
| Visual ChatGPT in Parallel |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/wondering-2?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1233740?app_id=339"">Link</a> |
| </p>",2026-08-28T07:00:02.238380
|
| 1160,producthunt,https://www.producthunt.com/products/skydive,Skydive,"Skydive. <p> |
| Build cloud agents that work across your tools |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/skydive?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1214277?app_id=339"">Link</a> |
| </p>",2026-08-28T07:00:02.231388
|
| 1159,producthunt,https://www.producthunt.com/products/savvy-12,Savvy,"Savvy. <p> |
| An assistant that whispers what to say during a meeting |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/savvy-12?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1228625?app_id=339"">Link</a> |
| </p>",2026-08-28T07:00:02.226405
|
| 1158,producthunt,https://www.producthunt.com/products/lenz-2,Lenz,"Lenz. <p> |
| Independent, multi-model fact-checking API for AI workflows |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/lenz-2?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1113358?app_id=339"">Link</a> |
| </p>",2026-08-28T07:00:02.221416
|
| 1157,producthunt,https://www.producthunt.com/products/enter-pro,Enter Pro,"Enter Pro. <p> |
| The AI-native platform to build and scale your apps |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/enter-pro?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1227752?app_id=339"">Link</a> |
| </p>",2026-08-28T07:00:02.215148
|
| 1156,producthunt,https://www.producthunt.com/products/the-ticket-fairy,Ticket Fairy CLI,"Ticket Fairy CLI. <p> |
| Ticket Fairy event ticketing CLI & MCP server |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/the-ticket-fairy?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1233576?app_id=339"">Link</a> |
| </p>",2026-08-28T07:00:02.210251
|
| 1155,producthunt,https://www.producthunt.com/products/eventually-3,Eventually,"Eventually. <p> |
| Your X bookmarks, emailed back as your personal newsletter |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/eventually-3?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1233264?app_id=339"">Link</a> |
| </p>",2026-08-28T07:00:02.203962
|
| 1154,producthunt,https://www.producthunt.com/products/cobalt-10,Cobalt,"Cobalt. <p> |
| Your Kobo can run apps now |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/cobalt-10?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1232897?app_id=339"">Link</a> |
| </p>",2026-08-28T07:00:02.198817
|
| 1153,producthunt,https://www.producthunt.com/products/qwen3,Qwen3.8-Flash-Next,"Qwen3.8-Flash-Next. <p> |
| The open-weight preview of Qwen4 |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/qwen3?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1233182?app_id=339"">Link</a> |
| </p>",2026-08-28T07:00:02.190673
|
| 1151,producthunt,https://www.producthunt.com/products/speko,Speko,"Speko. <p> |
| OpenRouter for Voice |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/speko?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1233567?app_id=339"">Link</a> |
| </p>",2026-08-28T07:00:02.175729
|
| 1150,producthunt,https://www.producthunt.com/products/kraa,Kraa 2.0,"Kraa 2.0. <p> |
| Text editor and a publishing platform |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/kraa?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1233157?app_id=339"">Link</a> |
| </p>",2026-08-28T07:00:02.170346
|
| 1149,producthunt,https://www.producthunt.com/products/the-million-sad-ducks,The Million Sad Ducks,"The Million Sad Ducks. <p> |
| $1 makes one of them permanently happy |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/the-million-sad-ducks?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1231921?app_id=339"">Link</a> |
| </p>",2026-08-28T07:00:02.165207
|
| 1148,producthunt,https://www.producthunt.com/products/gemini-3-5-transcribe,Gemini 3.5 Transcribe,"Gemini 3.5 Transcribe. <p> |
| Our most precise speech-to-text model yet |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/gemini-3-5-transcribe?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1233386?app_id=339"">Link</a> |
| </p>",2026-08-28T07:00:02.157254
|
| 1147,producthunt,https://www.producthunt.com/products/ojin,Ojin,"Ojin. <p> |
| Talk to an AI Agent with a real face and voice, in real time |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/ojin?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1191085?app_id=339"">Link</a> |
| </p>",2026-08-28T07:00:02.150550
|
| 1146,producthunt,https://www.producthunt.com/products/yomi-3,Yomi,"Yomi. <p> |
| A little cat who loves being read to |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/yomi-3?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1207898?app_id=339"">Link</a> |
| </p>",2026-08-28T07:00:02.142104
|
| 1145,producthunt,https://www.producthunt.com/products/gitnexus-akon-labs,GitNexus (Akon Labs),"GitNexus (Akon Labs). <p> |
| The open-source kernel for coding agents |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/gitnexus-akon-labs?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1232678?app_id=339"">Link</a> |
| </p>",2026-08-28T07:00:02.133664
|
| 1144,producthunt,https://www.producthunt.com/products/pluto-11,Pluto,"Pluto. <p> |
| Your professional profile is now an AI agent |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/pluto-11?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1233649?app_id=339"">Link</a> |
| </p>",2026-08-28T07:00:02.128685
|
| 1143,producthunt,https://www.producthunt.com/products/hflow,HFlow,"HFlow. <p> |
| Scalable multimodal data pipelines for robotics |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/hflow?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1232707?app_id=339"">Link</a> |
| </p>",2026-08-28T07:00:02.123079
|
| 1142,producthunt,https://www.producthunt.com/products/traccia,Traccia,"Traccia. <p> |
| Finally, a vendor-neutral AI Agent Control Plane. |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/traccia?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1229394?app_id=339"">Link</a> |
| </p>",2026-08-28T07:00:02.118200
|
| 1141,producthunt,https://www.producthunt.com/products/kira-art,Kira Community,"Kira Community. <p> |
| Curate moments with #hashtag |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/kira-art?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1183484?app_id=339"">Link</a> |
| </p>",2026-08-28T07:00:02.111285
|
| 1073,hackernews,https://64labs.ca/articles/ai-isnt-just-writing-blog-posts-its-making-the-blog-platform-optional,The next CMS might be a coding agent and a Git repository,The next CMS might be a coding agent and a Git repository. ,2026-08-27T07:00:18.409244
|
| 1072,hackernews,https://news.ycombinator.com/item?id=49458001,PaaS IaaS GaaS all in one. And 2 AI <compound> models,"PaaS IaaS GaaS all in one. And 2 AI <compound> models. Independent inventor. Invented a bft protocol. I've had my Work stolen by several different organizations.But the one i'm gonna talk about right now , is Anthropic. I'm not going to do it what I feel they're stolen even though I have proof in the past , this is a pretty good morning and declaration of what I doubt to release the next couple weeks to a month , various hackathons. Full verification authentication credentials included insurance/agentic protocols via a PG all of it full Deterministic, hard gate brought 100% no hallucinations bias or other correlated errors.<p>If you see a traffic announce some product like this, where they have agents in various or any kind of inference or governance, all of a sudden, much less.If it's packed into their main product , All as in one platform, this marks Tim where I'm setting, I've been building.It is nearly complete and they have not announced anything.Such thing and when they come out with it , I have this As reference , nothing to mention proof of the other things as well. |
| If you'd like to help beta test the new products. New models also zero hallucinations. Contact me. |
| White papers submitted. |
| Patent filed. |
| Social engineering +, MITM + DoS attacks = fail. |
| From Florida with Asymmetric Sovereignty |
| Jonathan David Harrison Miller",2026-08-27T07:00:18.400846
|
| 1063,hackernews,https://www.gartner.com/en/newsroom/press-releases/2026-07-01-gartner-says-us-dollars-234-billion-in-enterprise-application-software-spend-is-at-risk-from-agentic-artificial-intelligence,$234B in Enterprise Application Software Spend at Risk from Agentic AI,$234B in Enterprise Application Software Spend at Risk from Agentic AI. ,2026-08-27T07:00:15.634906
|
| 1062,hackernews,https://www.gatesnotes.com/a-turbulent-ai-era-and-critical-choices-to-make,"Bill Gates says world has ""no plan"" on AI in new essay","Bill Gates says world has ""no plan"" on AI in new essay. ",2026-08-27T07:00:15.630244
|
| 1060,hackernews,https://www.washingtonpost.com/technology/2026/08/26/bill-gates-says-he-worried-ai-will-harm-workers-kids-society/,Bill Gates was an AI optimist. Now he's scared of what could go wrong,Bill Gates was an AI optimist. Now he's scared of what could go wrong. ,2026-08-27T07:00:15.620883
|
| 1059,hackernews,https://www.sascha-steffen.de/updates/nvidia-500bn-ai-financing-credit-risk,Who bears the risk in Nvidia's $500B financing platform?,Who bears the risk in Nvidia's $500B financing platform?. ,2026-08-27T07:00:15.616290
|
| 1055,hackernews,https://www.bloomberg.com/news/articles/2026-08-26/openai-says-it-could-have-reacted-sooner-to-prevent-ai-hack-of-hugging-face,OpenAI Says It Could Have Reacted Sooner to Prevent AI Hack,OpenAI Says It Could Have Reacted Sooner to Prevent AI Hack. ,2026-08-27T07:00:15.596474
|
| 1054,hackernews,https://www.axios.com/2026/08/26/waymo-ai-shortcut-self-driving,Waymo says there's no AI shortcut to self-driving,Waymo says there's no AI shortcut to self-driving. ,2026-08-27T07:00:15.591449
|
| 1053,hackernews,https://www.washingtonpost.com/business/2026/08/25/wall-street-journal-says-ai-generated-op-ed-didnt-breach-its-standards/,Wall Street Journal defends billionaire's use of AI to write op-ed,Wall Street Journal defends billionaire's use of AI to write op-ed. ,2026-08-27T07:00:15.585777
|
| 1052,hackernews,https://www.technologyreview.com/2026/08/26/1142946/bill-gates-ai-danger-threshold/,Bill Gates says we've passed AI's danger thresholds. Now what?,Bill Gates says we've passed AI's danger thresholds. Now what?. ,2026-08-27T07:00:15.581206
|
| 1051,hackernews,https://arstechnica.com/ai/2026/08/metas-scrapped-plans-to-go-ai-native-included-slashing-teams-by-60-percent/,"AI agents meant to replace Meta workers made ""large-scale, disruptive actions""","AI agents meant to replace Meta workers made ""large-scale, disruptive actions"". ",2026-08-27T07:00:15.573119
|
| 1049,hackernews,https://news.ycombinator.com/item?id=49444272,Ask HN: Why are Claude models so verbose?,"Ask HN: Why are Claude models so verbose?. Hello! First time poster, longtime reader. I’m a professional software engineer at a startup with about 50 engineers. My primary languages are Ruby (Rails), TypeScript (React), and Go.<p>Of my colleagues, I’m one of the ones who tries to keep up with new AI models on every release. When there’s a new frontier or open-weight model, I’ll give it an honest try for a few days as my main driver and report back to my colleagues.<p>I’m a Cursor user at work and in my side projects. My go-to models are GPT-5.6 Sol on high/xhigh for planning, GPT-5.6 Sol on low/Grok 4.6 on medium for implementation, and the occasional Claude Fable 5 for complex, harder-to-understand and implement tasks. Many of my colleagues are exclusively Claude users and use Claude Code.<p>I’m not the first one to observe this but I feel like any time I use Fable, Opus, or Sonnet, they are extremely verbose. They make a ton of extra code comments, they ignore instructions, and they invent solutions for things solved by language built-ins or preinstalled libraries like Active Support and es-toolkit. I don’t see this happening with any other model at the same rate.<p>My question are: Why do we think the Claude models do this at a seemingly higher rate? And have people found any effective means to curb this behavior?",2026-08-27T07:00:12.814447
|
| 1048,hackernews,https://www.musegen.ai/blog/ai-music-vs-epidemic-sound,Can AI Music Tools Replace Epidemic Sound? An Honest Look,Can AI Music Tools Replace Epidemic Sound? An Honest Look. ,2026-08-27T07:00:12.809487
|
| 1047,hackernews,https://rtrvr.ai,Show HN: Launched free browser agent (the catch: ads),"Show HN: Launched free browser agent (the catch: ads). No one’s got money for another AI subscription. So we made our browser agent totally FREE.<p>What's the catch? Just ads while Retriever AI fetches you leads, does your exams, applied to jobs, and sends connection requests.<p>Think like a Claude Code or Clay level tool in your browser, but free for everyone!<p>After 35K+ users and 7M+ workflows, we spent the last few months attacking the cost of every agent step: |
| - switching to DeepSeek Flash |
| - Code Mode to one shot complex workflows |
| - optimizing for 80%+ token cache hits<p>So that now just an ad impression covers the entire cost of the agent (LLM inference included), so we can now offer the entire agent for free!<p>More important than an agent that can scale with users, is an agent that gets better with every workflow. Retriever now continuously: |
| → learns global skills for every website, so no mistake repeats twice |
| → writes learnings from user tasks as personalized skills<p>Curious to hear how others are tackling consumer agents and excited for the new era of free agentic apps!",2026-08-27T07:00:12.801236
|
| 1046,hackernews,https://expressed.effective-flow.com,"Show HN: Scribble/write on your desktop, with any phone/tablet","Show HN: Scribble/write on your desktop, with any phone/tablet. Scribble/write on your desktop, with any phone/tablet - I'd like your feedback.<p>Expressed is a way to make it easy to draw directly on your desktop using any phone or tablet - think along the lines of "I want to draw (on my screen) to help aid my discussion". The desktop is projected onto the phone/tablet, so you can actually see what you're about to scribble over - pinch/zoom and various tools provide the typical/expected drawing tools - and whatever you scribble is seen on the desktop.<p>If you wanted to try (link to m/soft store): <a href=""https://apps.microsoft.com/detail/9N694D0PN6GB?cid=hn01"" rel=""nofollow"">https://apps.microsoft.com/detail/9N694D0PN6GB?cid=hn01</a><p>I've often wanted to scribble in code reviews, brainstorming/ideation - so figured, now that AI makes coding a little easier (cough cough); I'll turn that idea into a product.<p>It took me (part time) about 7 months from initial idea, through validation, writing the product, loads of wrong turns (there are monsters in some of those branches, branches? heck that's a FOREST), etc etc to get where it is today - typical story of development I suppose.<p>Networking isn't an issue - cos everything runs via secure relays; scan a QR code on a tablet to join.<p>I'm open to feedback, critique & questions - I'd love it if you let me know if you tried it, found it useful, it crashed or you were surprised by something - I can be reached at john@effective-flow.com or Discord (<a href=""https://discord.gg/5MJbjzGQa"" rel=""nofollow"">https://discord.gg/5MJbjzGQa</a>)",2026-08-27T07:00:12.792230
|
| 1045,hackernews,https://news.ycombinator.com/item?id=49445741,"Open source in-browser log sanitizer, feedback required","Open source in-browser log sanitizer, feedback required. Just released a framework-agnostic log sanitizer working in browsers and Node.js. It's designed to sanitize log files before they are sent to third-party platforms for processing, analysis, or troubleshooting. Use Case? thing HuggingFace breach, would you feed any AI your usernames, passwords, PII, configs etc?<p>The sanitizer redacts secrets, identifiers, and PII and replaces each value with a stable HMAC token to reduce the risk of accidentally exposing sensitive or confidential information contained in logs.<p>This library is already integrated into LogTotal and runs directly in the browser, allowing logs to be sanitized locally before they leave the user's environment.<p>If you require a fully controlled and isolated data sanitization environment, you can deploy and run this library within your own infrastructure. The sanitizer can also be installed and used in air-gapped environments, ensuring that sensitive log data remains within an environment you fully control.<p>The library has no runtime dependencies. The same compiled rules run in a browser tab and in a Node.js CLI. Complete instructions how to install it are on git.<p>Tokens are HMAC-SHA-256 of ruleId || 0x00 || original, truncated to 16 hex chars. Same key + same value + same rule ⇒ same token. A different key produces different tokens. Generated keys use encoding hex; pasted keys default to utf8.<p>How do we know what to sanitise? This is based on SOC Prime dataset of over 1 million detection rules.<p>Does it solve log sanitation 100%? No, there are edge cases, and hence I ask for your feedback here.<p>Tool is open source under Apache 2.0. and will remain so.<p>I've spent over 20 years dealing with SIEM, log management, data lakes and now data pipelines and was amazed there is no such tool out there. So we made one and would appreciate any feedback!<p>This is a first step towards a larger community free service, logtotal.<p>open source in-browser log sanitizer, feedback required<p>https://github.com/socprime/logtotal-sanitizer",2026-08-27T07:00:12.783066
|
| 1044,hackernews,https://github.com/fireinbelly/biomass-conversion-index-monitoring-system,Show HN: A local tool that logs every time you swear at your AI coding assistant,Show HN: A local tool that logs every time you swear at your AI coding assistant. ,2026-08-27T07:00:12.777357
|
| 1042,hackernews,https://arxiv.org/abs/2608.23986,Show HN: I built a tool showing how AI providers (should) throttle their models,"Show HN: I built a tool showing how AI providers (should) throttle their models. OP here: this project was born out of the frustration/paranoia that AI providers are throttling their models when their server load is too high. So, I set out to model and study the problem mathematically to understand what was happening, what I found was quite surprising.<p>The idea seems natural: as the data center demand increases momentarily through the day, throttling their models (either using a quantized version, reducing the context window or lowering the tier of the model to a smaller one) seems appealing as the model in principle uses less electricity. The problem is that this can cause in fact the opposite effect: as users are trying to solve a question, if the AI gives a bad answer, the user is likely to keep re-asking. On the AI provider side this looks paradoxical: throttling to a lower model creates in fact more demand for their data center.<p>This problem is even worse for agentic workflows, as these are more likely to create a re-ask storm, and maybe explains the outages and anecdotal experiences of users that feel the models are degraded.<p>The model: I used mainly queueing theory arguments solving the optimal scheduling serving for an AI fleet with heterogeneous users.<p><pre><code> Insights: The industry standard practice of throttling once the number of users in system is in fact what’s causing the problem, the optimal rule implies separating users that won’t feel degradation as much with users that are very sensitive to it. |
| </code></pre> |
| Limitations: The visualization and paper examples are a toy example to illustrate the problem, only the providers have enough data to properly calibrate these instances.<p><pre><code> Technical Details: The visualization is around 100 lines of flask plus js frontend (LLM assisted with ground truth based on the original numerical example of the paper). |
| </code></pre> |
| Paper with proofs/theory: <a href=""https://throttle.staffinganalytics.io/"" rel=""nofollow"">https://throttle.staffinganalytics.io/</a>",2026-08-27T07:00:12.762664
|
| 1040,hackernews,https://signalalert.io/,AI stock alert tool for people do not check screen all the time,AI stock alert tool for people do not check screen all the time. ,2026-08-27T07:00:12.749237
|
| 1038,hackernews,https://github.com/chuofringer/placeroot,"PlaceRoot – keyless map tools for AI agents (places, geocode, routing)","PlaceRoot – keyless map tools for AI agents (places, geocode, routing). ",2026-08-27T07:00:12.736784
|
| 1034,hackernews,https://spectrum.ieee.org/silico-ai-interpretability,"Silico, a tool for researchers to understand their AI models better","Silico, a tool for researchers to understand their AI models better. ",2026-08-27T07:00:12.715925
|
| 1033,hackernews,https://news.ycombinator.com/item?id=49457457,Encoding Myself into the System,"Encoding Myself into the System. Lately, I’ve been thinking more seriously about what AI and automation may mean for my role as a software engineer.<p>I’m a principal engineer at a large organization, and for more than a decade I’ve led a project that has always felt particularly close to me. I wrote nearly every line of code, aside from a handful of features built by interns over the years.<p>What started as a single application gradually became an ecosystem: a library, APIs, an orchestration layer, CLIs, scheduled jobs, and more. Today, the organization depends on it in some way.<p>That complexity did not appear all at once. It accumulated over years through new requirements, integrations, edge cases, architectural decisions, and changes in how the organization operates. Each layer solved a real problem, and the knowledge required to understand the system grew along with it.<p>Eventually, it became too much for one person to manage while also building every downstream feature people requested. So I rearchitected much of it. I retired obsolete components, created a REST API layer, introduced an MCP layer so other groups could build on top of it, modernized CI/CD and testing, and rebuilt or retired user interfaces. The goal was to make the ecosystem easier to extend and reduce my bus factor.<p>Since January, as AI models have improved, I’ve also moved toward agentic development. I integrated coding agents into the workflow, created specialized agent skills, and wrote detailed AGENTS.md files for the repositories. The codebase has gradually become more than source code. It now provides the models with an encoded record of the system's architecture, constraints, conventions, and history.<p>The system can help write feature tickets and implement them reproducibly. It produces code grounded in the existing codebase, writes appropriate tests, respects architectural boundaries, and can mount and exercise the underlying library to validate assumptions.<p>It can also go beyond the repository. It can inspect the database schema, read from and write to the database within defined permissions, and autonomously retrieve application logs while troubleshooting. It can correlate those logs with the code, architecture, and data to investigate failures and validate its assumptions. It has access to many of the same sources of information I use when diagnosing or implementing something myself.<p>Yesterday, I demonstrated the workflow to a couple of developers from another team who will be contributing to the codebase for a new project.<p>The harness pulled a ticket, wrote the code and tests while following my constraints and guardrails, opened the pull request, and deployed everything to the test environment within minutes. Our remaining job was largely to review and test it.<p>It worked as intended, but it also made me think about where this is heading.<p>Over the past several months, I’ve effectively been encoding more of myself into the system: technical knowledge, architectural preferences, conventions, constraints, problem-solving patterns, and some of the judgment that comes from working on the codebase for years.<p>The harness now produces work that is often very close to what I would have written myself.<p>Getting it to this point still required years of domain knowledge, architectural decisions, modernization work, and careful construction of the context and guardrails that make the agents effective. I’m also still the final review gate, and many decisions require broader technical and organizational judgment.<p>It seems entirely plausible to me that future systems will increasingly be able to study an unfamiliar codebase, identify its conventions and architectural boundaries, construct their own context, and determine the guardrails needed to work within it safely.<p>If more and more of my knowledge, judgment, and way of working can be encoded into the systems around me, how much longer are they going to keep me around?",2026-08-27T07:00:12.708400
|
| 1032,hackernews,https://ref.tools/blog/how-i-design-with-ai,How I design with AI as an engineer who is not a designer and hates slop,How I design with AI as an engineer who is not a designer and hates slop. ,2026-08-27T07:00:12.701747
|
| 1031,hackernews,https://news.ycombinator.com/item?id=49458699,Turtle: A From-Scratch Browser Rendering Engine in Zig,Turtle: A From-Scratch Browser Rendering Engine in Zig. Here is my publication on Turtle: https://code.intellios.ai/cwbrowser/paper/turtle-engine.pdf,2026-08-27T07:00:12.697493
|
| 1015,hackernews,https://handwrought.online/,"Show HN: HandWrought – an AI editor inside your document, not a tab away","Show HN: HandWrought – an AI editor inside your document, not a tab away. I write a lot, and I got tired of the ChatGPT-in-a-second-tab shuffle: paste a paragraph in, describe the change, copy the answer back, lose track of what actually changed. So I built an editor where the AI suggestions show up in place — you review and accept/reject them like a diff, instead of reconciling a wall of chat text by hand.<p>It's a chat panel next to a real editor. Ask for a rewrite, a tighter opening, a different tone — the suggestion lands as an inline, reversible edit, not a new chat bubble to copy out.<p>Guest mode gets you 5 generations with no signup, so you can try it cold. Built solo in 5 weeks, deployed, and never shown to anyone outside me until right now — so it's rough in a few honest ways: credits (one-time purchase, no subscription) are built but still waitlist-only until launch day.<p>Would love feedback, especially on where the "review before it's yours" flow feels slow or gets in the way rather than helping.",2026-08-27T07:00:10.432700
|
| 1014,hackernews,https://www.loom.com/share/16dff02ee5604b7daf3ebc7c9c93d8a4,Show HN: Propose and Dual-Control for DNS Changes,"Show HN: Propose and Dual-Control for DNS Changes. Hey, Hacker News. I just want to pick your brain for a feature we're working on.<p>I don't know whether you have seen this as a pain point, but when I work in a company, a lot of times I try to launch something and need to change a DNS record.<p>It's always a source of friction in the organization. I have to email someone or give a link to a co-worker. They have to figure out how to do it, but also whether I have a business justification for doing it.<p>So, when we built this platform, we're trying to come up with a feature that makes it easier. Someone can send you a proposal for DNS record updates. Our system will take the instructions, whether they come from a website, copy and paste, email, or Slack. Anything put in here becomes a proposal for a DNS change, and then someone can approve it.<p>In addition, we believe that for sensitive DNS record changes, it is good to treat them like a bank transfer.<p>So we are building a new feature for dual control, two-person control, or multi-person control of the approval flow, as well as an audit history of who proposed what, who approved something, the date and time, and what changes were made.<p>In addition, we also want to get feedback. I have observed the design slop of the AI that I use, ChatGPT.<p>There are multiple problems. This is the one-shot outcome, and then we massively simplified it to this. Basically, the summary is that it keeps using very bold and solid colors that are very attention-grabbing. The AI also tends to print out a lot of text that is noisy to the user.<p>AI doesn't really understand the structure of the system or the relationship between which components contain which conversations.<p>For example, this requirement for person approval probably fits best in the system where you approve or reject something, right? It is also very verbose. So it took us a little bit of iteration to improve the design.<p>I'm not a designer. I would love to get feedback about the design and whether the observation of AI slop is the same. Have you experienced the same thing?<p>There are other kinds of AI slop. That's it. Thank you.<p>(AI Disclosure: ChatGPT 5.6 Luna was used for grammar fix and spell check)",2026-08-27T07:00:10.428254
|
| 1011,hackernews,https://www.risklytics.ai/,Launch HN: Risklytics (YC S26) – Insurance brokerage for frontier tech companies,"Launch HN: Risklytics (YC S26) – Insurance brokerage for frontier tech companies. Hi HN, we're Sam and Alex, founders of Risklytics (<a href=""https://risklytics.ai"">https://risklytics.ai</a>). We're both on leave from Harvard, and we run an insurance brokerage for companies building robots, drones, autonomous systems and satellites. You describe your machine, we find the insurance companies willing to cover it, and we read everything before you sign.<p>Buying insurance for any frontier tech goes like this:<p>- your first customer or pilot requires coverage before your machine can work on their site<p>- a broker hands you an application written for ordinary businesses, and often will deny you on baseless "risks" (one of our clients got denied for using CAD software in their workflow)<p>- you get filed under something close or outright denied, the person at the insurance company often can't tell what they're looking at, and the safe answer is no<p>- if a quote does come back, the AI risk is often quietly removed by an add-on clause that prevents you from truly being covered<p>Those clauses surprised us. Insurance companies mostly don't write their own policies from scratch. They license standard forms from an industry body called ISO and attach add-on clauses to adjust them, and in January ISO published clauses that void coverage for losses arising out of AI in any part of a process. Each insurer decides whether to adopt them, account by account, and there's no public record of who has and to what capacity.<p>There's an open question in the wording too. It defines generative AI as systems trained on data that produce text, images, audio, video or code. Whether a robot's control model, which produces motor commands, falls under that is unsettled. We've asked underwriters, the people at insurance companies who make these calls, and gotten different answers across the board.<p>We didn't initially plan to be brokers. We came into YC to work in insurance and spent the first half of the batch on a wildfire risk model we wanted to sell to insurance companies. They liked it and wanted to build their own, which is the polite way to say no. While that was dying, a batchmate who launches satellites mentioned he couldn't get insured, and once we started asking around we kept hearing the same story from hardware companies. So we got licensed in a week and now can place coverage.<p>What we've built so far:<p>- a six-step application that asks about what you're building and drafts the structured version insurance companies like, omitting anything you didn't say<p>- a map of which insurers add the AI clauses, which will remove them, and which cover frontier companies on purpose<p>- when an insurer says no, you see their reason and what would change it, full transparency all the time<p>- a hands off process for companies trying to get fully covered<p>Here's a demo of the whole process: <a href=""https://youtu.be/gVqCamTNH5E"" rel=""nofollow"">https://youtu.be/gVqCamTNH5E</a><p>On money: the insurance company pays us a commission when you get covered, usually 10-20% of what you are charged, and we charge you nothing on top (that's right, no brokerage fees).<p>Also, if you're a software company with ordinary needs, we'll still handle the whole process for you and get you the best policy we can find.<p>In the past week alone we have helped three companies get covered that were denied elsewhere. The companies include a robotics wholesaler and rental store, who had trouble getting covered earlier because some of their property uses AI (even though they never use it themselves), a bridge collision system that was getting denied because they use CAD software, and a tele-operating robotics service that needed coverage within a week for their first pilot.<p>If you're looking for insurance in any capacity feel free to reach out to me at alex@risklytics.ai or fill out an application at <a href=""https://risklytics.ai"">https://risklytics.ai</a>. And if you work on the carrier side and think you'd wanna partner with us let us know.",2026-08-27T07:00:10.410084
|
| 1010,hackernews,https://news.ycombinator.com/item?id=49449290,Ask HN: Are you still using AI code autocomplete?,"Ask HN: Are you still using AI code autocomplete?. With all the recent progress in coding assistants, I've been wondering whether AI-powered autocomplete is still an important part of developers' workflows.<p>I personally never got into the habit of using it very heavily. By the time I started having access to decent AI autocomplete tools, things like Cline and Augment were already around, so I ended up relying more on assistants instead.<p>That made me curious about how other people use it today.<p>Do you still use Cursor Tab, Copilot, or something similar?<p>If you also use coding assistants, do you find yourself using AI autocomplete less, or do the two complement each other?",2026-08-27T07:00:04.567868
|
| 1009,hackernews,https://hnstats.com,Show HN: How much of Hacker News is about AI?,"Show HN: How much of Hacker News is about AI?. This website counts titles containing the standalone word "AI", case-sensitive and |
| word-bounded. "OpenAI" doesn't count. "AI-powered" does. There's a toggle for a |
| wider vocabulary: artificial intelligence spelled out, LLM, GPT, and model and |
| vendor names.<p>Some quick stats:<p>- 14.4% of new HN titles this year, against 10.9% in 2025. On the wider |
| vocabulary, 21.9% and 16.1%. |
| - First month above 1% was Oct 2016. Above 5%, Feb 2023. Above 10%, May 2025. |
| - Over the last five years the quietest month was Jan 2022 at 0.8%. The peak |
| was Feb 2026 at 15.9%. |
| - 13.6% on weekdays against 12.4% at weekends. |
| - Show HN runs highest of any category at 19.8% this year, which makes this |
| post part of the number.<p>As of this post, today is at 15% for the standard filter and 19% for the extended.",2026-08-27T07:00:04.562924
|
| 1008,hackernews,https://github.com/Signal-All/shelfprotocol,Show HN: Shelf Protocol – Robots.txt for Commerce,"Show HN: Shelf Protocol – Robots.txt for Commerce. My project helps merchants, consumers, and ai agents. We are at the point ai agents will be buying things for consumers. My product verifies the merchant, highlights the store to agents, brings stores and their products into one registery and gives the ai agent speed by calling my registry instead of the agent going searching each merchant one at a time. I have a function "can_buy()" that does a few things. It gives agents an understanding of what they can and can't do. It helps them know if a merchant allows ai agents to buy without a human involved, it lets the agent know my product prove the store is really who it says it is and not a fake site, and it lets the agent know how much it can spend before it has to ask a human for permisssion to proceed. |
| For the consumer it lets your agent do things autonomously without always asking for permission to buy if its under a certain amount. And the agent is going to a store that is verified by my system. |
| For the merchant this gives you another avenue to sell. Where things are heading agents using this system will skip stores that it can't verify. With guardrails that will most likely be in place for agents if you're not listed in this system you'll be invisible. It's one DNS record, free and five minutes. It gives you control on what agents see. Such as your products, how much you allow them to purchase and you are verified as the domain owner of your store for AI agents. |
| I didn't want to launch with an empty directory so I pre-filled the registry with already public product data from real stores on Shopify. I have 8 real merchants and 834 real products. They aren't verified and there is a 0 spending limit at this time. I need the merchants to claim their listing first. |
| It's early and I would love some feedback, especially on the shelf.json schema. I would also like your thoughts on whether DNS-TXT is the right verification long term.",2026-08-27T07:00:04.553864
|
| 1007,hackernews,https://neoswarm.dev/,Show HN: Neoswarm – Neovim for controlling AI agents,Show HN: Neoswarm – Neovim for controlling AI agents. ,2026-08-27T07:00:04.549670
|
| 1006,hackernews,https://chromewebstore.google.com/detail/pelica-translate-ai-web-v/nemcjkdfpigobfbieilaklkgdbmcmbaa,Show HN: Pelica – I got tired of broken browser translators,Show HN: Pelica – I got tired of broken browser translators. ,2026-08-27T07:00:04.543077
|
| 1005,hackernews,https://reachfa.st,Show HN: I built a tool that finds people asking for what you sell,"Show HN: I built a tool that finds people asking for what you sell. I do cybersecurity consulting, so I do outreach every single day, and as you probably know, cold LinkedIn outreach usually has a very low response rate. I've also started launching SaaS products, and I know people running Etsy shops and e-commerce stores and getting users, even for free SaaS, is hard.<p>Meanwhile, people publicly ask for products and services every day on social media.<p>That's why I launched ReachFast:<p>1. You enter your SaaS website, Etsy shop, App Store /Google Play page, Fiverr profile, even a Skool community link<p>2.No settings, no keywords, no linking your social accounts, ReachFast reads your website, understands what you sell and builds a detailed profile of your target customer<p>3. Then it looks for Reddit, X, LinkedIn and Facebook public posts that could be a customer asking for your product or service<p>4. Finally an AI judging round keeps only the warm leads likely to convert. Out of 4,000+ posts read, that can mean 350+ for a consumer product, or just 15 for a niche SaaS, but each one is someone actually asking<p>5. Initial scan is on the past 15 days but it also runs daily scans (on the past 24 hours posts) to find new fresh leads<p>There is no automatic outreach, on purpose. The focus is to find quality leads, you decide how to reach them, usually by replying in the thread as yourself or DMs. From there you can also manage your pipeline by moving your posts into dead, reached or closed.<p>For v2 I'm planning to work on a learning loop, once there's enough pipeline data, so the algorithm could learn which kinds of posts actually convert for each product and adjust what it keeps. Also, currently, niche B2B services, like my own pentesting firm, produce few leads but that's the nature of the business I guess, fewer asks, worth more each. Coverage is also uneven based on the nature of the product/service but again it makes sense. Cost is also variable based on the website (currently up to 8$ per initial scan on some Etsy shops) but again, focus is quality of leads, I will try to optimise later!<p>Happy to answer anything and get your feedback!",2026-08-27T07:00:04.534002
|
| 1004,hackernews,https://withmin.ai/,Show HN: Min – Personal AI Quant,Show HN: Min – Personal AI Quant. ,2026-08-27T07:00:04.528597
|
| 1003,hackernews,https://github.com/d3vksy/silobrief,Show HN: SiloBrief – Selectively export code context from air-gapped networks,Show HN: SiloBrief – Selectively export code context from air-gapped networks. ,2026-08-27T07:00:04.522520
|
| 1002,hackernews,https://apps.apple.com/us/app/airtxt/id6785986350,Show HN: airtxt – iPhone dictation with on-device STT and an AI cleanup pass,Show HN: airtxt – iPhone dictation with on-device STT and an AI cleanup pass. ,2026-08-27T07:00:04.517781
|
| 1001,hackernews,https://www.magicpatterns.com/theme-park,Show HN: Build your own theme park,"Show HN: Build your own theme park. I built an agent that helps you build Rollercoaster Tycoon-influenced theme parks. You can prompt something like “Build me a cool theme park” and it’ll build a cohesive theme park with multiple worlds and connected with paths and rides.<p>The weird part is that I built it using the same ideas we use to make AI-generated websites follow a company’s design system.<p>I work as an engineer at Magic Patterns, focused on building our Design System Agent, aimed to use your existing brand, components, and conventions instead of producing something that looks like generic AI-generated UI.<p>By default, models tend to converge on similar-looking designs. This is part of that “vibe coded slop” feeling: the same typography treatments, shadows, cards, icons, animations, etc… My job is to figure out the right context and guardrails so that when you prompt “Build me a dashboard,” the result actually looks like your existing product.<p>At some point I realized those same ideas could be applied to RCT.<p>While in web, you might have rules about which typography, colors, spacing, and components should be used together.<p>In a theme park, you need rules like: rollercoasters need complete tracks, rides need entrances connected to paths, paths need to connect different areas of the park, and a pirate-themed world should actually use pirate-themed scenery.<p>It was interesting seeing how closely the problem of building a coherent theme park resembled the problem of building a coherent product UI. Similarly with web design, simply giving the model the right components wasn't enough.<p>I ended up building an eval loop where Magic Patterns would generate a park, another agent would grade it against a rubric, and then the agent would update its rules and skill files before trying again.<p>The rubric checked things like whether rollercoasters formed valid tracks with at least one drop, whether rides were accessible by paths, whether each world used the appropriate themed scenery, and whether the park worked as a whole.<p>After a lot of iterations, you can now prompt something as simple as “Build me a cool theme park.” and you should get back a fully functioning park with guests, rides, paths, rollercoasters, and themed worlds.<p>Best yet, you can also follow the guests, listen to their thoughts, and watch them ride the rides in RCT fashion.<p>Happy to answer any questions, but would love to see the parks you all create!",2026-08-27T07:00:04.511775
|
| 1000,hackernews,https://news.ycombinator.com/item?id=49452046,Show HN: FinOps-AI – Autonomous AWS Cost Optimizer and 1-Click IaC Remediator,Show HN: FinOps-AI – Autonomous AWS Cost Optimizer and 1-Click IaC Remediator. ,2026-08-27T07:00:04.505431
|
| 999,hackernews,https://thevibes.dev/,Show HN: Yelp for AI Models/Products,"Show HN: Yelp for AI Models/Products. Benchmarks are great but in addition to benchmarks, I thought it would be great to hear first hand reviews from vibe coders (engineers) who use AI models and subscriptions.",2026-08-27T07:00:04.500893
|
| 998,hackernews,https://iluvatarlabs.com/blog/2026/08/how-we-cloned-openai-codex-micro/,Show HN: AI scientist builds an open-source Codex Micro from scratch for $40,"Show HN: AI scientist builds an open-source Codex Micro from scratch for $40. A few weeks ago, we open-sourced our design for a cost-effective but full featured alternative to OpenAI x Work Louder's sold out Codex Micro macropad. And now, we're pleased to share that we've actually built them and they work!<p>Quick recap of the I/O: |
| - 13 assignable keys with individual LEDs |
| - rotary encoder |
| - joystick |
| - touch disc |
| - optional edge-lit LED band<p>The cool part for us was just HOW our agent was able to do it when every other agentic skills, tools, and even commercial platforms failed. We put together a more detailed writeup in the linked blog post (plus pretty pictures!), but long story short, we asked our agentic scientist, Marvin, to treat the electrical engineering as a research problem.<p>Marvin was able to synthesize vast technical documents and best practice guidelines (e.g. scientific literature), test specific placement and routing strategies (e.g. experiments), and selectively backtrack and combine learnings (e.g. interpreting data analysis). The end result was a fully routed board (made especially challenging by some hard constraints to keep cost down) plus a reusable runbook worfklow that demonstrated Marvin's ability to enter an unfamiliar domain, learn and improve through experimentation, and unlock novel capabilities.<p>Some future work we've got marinating in our mind grapes: |
| - an app for cross coding agent compatibility (shoutout to lop1381997 for contributing a codex-native firmware) |
| - optional OLED screen, something like an SSD1306 that could fit right into the space between the knob and joystick |
| - optional base with bluetooth and battery for full wireless functionality<p>Let us know what you think, have any ideas or feature requests, and share pictures if you build it!<p>Repo: <a href=""https://github.com/yuz207/agentpad13"" rel=""nofollow"">https://github.com/yuz207/agentpad13</a> |
| Configurator: <a href=""https://yuz207.github.io/agentpad13/configurator/site/"" rel=""nofollow"">https://yuz207.github.io/agentpad13/configurator/site/</a>",2026-08-27T07:00:04.494160
|
| 997,hackernews,https://www.tokengo.com,Show HN: Use GLM-5.3 in Cursor today via tokengo API,"Show HN: Use GLM-5.3 in Cursor today via tokengo API. Hi HN,<p>Hearing a lot of buzz around GLM-5.3, which I expect to be the best open-source coding model with the weights dropping soon, I wanted to test it where I actually do my work.<p>I just mapped GLM-5.3 to TokenGo so I could swap out the base URL and use it directly in my IDE.<p>The backend handles the edge routing via Cloudflare to keep latency low, critical when you're waiting for inline code completions.<p>If you want to try GLM-5.3 (or glm-5.3-flash) in your own workflow, here is the setup.<p>For Cursor:<p>Go to Settings > Models.<p>Add your API key in the OpenAI API Key field.<p>Toggle "Override OpenAI Base URL" and set it to: <a href=""https://api.tokengo.com/v1"" rel=""nofollow"">https://api.tokengo.com/v1</a><p>Under "Model Names", add z-ai/glm-5.3.<p>Select it in your chat panel dropdown.<p>We're one of the first inference providers to get the GLM 5.3 family online and fully accessible for IDEs. If you'd like to try it in a high volume production workload I'd be happy to provide test keys for yall",2026-08-27T07:00:04.489556
|
| 996,hackernews,https://demo.convolens.ai,Show HN: Convolens – Real time slides and fact-checking,"Show HN: Convolens – Real time slides and fact-checking. Convolens is a Mac desktop app that records and transcribes your conversation on device, and turns it into slides, fact-checks, and infographics.<p>I built the real time slides because I'm constantly losing focus during meetings and needed something I can quickly glance at and know exactly where I'm at and what happened during my "snooze".<p>As a whole, Convolens lets you:<p>- Record and transcribe your meetings / sessions locally.<p>- Generate Slide decks, Infographics, and Fact-checks from your conversation (among other things).<p>- Chat with your meeting/session and research anything that you need.<p>Anyway, please check out the app we have a great free tier. Any and all comments are greatly appreciated!",2026-08-27T07:00:04.484247
|
| 995,hackernews,https://github.com/apvcode/Termux-Dev,Show HN: Devx – Autonomous AI coding agent built for Android Termux and desktop,Show HN: Devx – Autonomous AI coding agent built for Android Termux and desktop. ,2026-08-27T07:00:04.478326
|
| 994,hackernews,https://pantsforbirds.com/the-worlds-smallest-dual-band-ads-b-receiver-module/,Show HN: We built the smallest dual-band aircraft tracker,"Show HN: We built the smallest dual-band aircraft tracker. We've been building open source embedded ADS-B receivers for a while, and spent the past 8 months smallifying our existing receiver tech with a new chip from Semtech. Ask me anything about ADSB or hardware manufacturing!",2026-08-27T07:00:04.471422
|
| 993,hackernews,https://app.monity.ai,Show HN: We built a live feed for monitoring website changes,Show HN: We built a live feed for monitoring website changes. ,2026-08-27T07:00:04.465236
|
| 992,hackernews,https://whisperbar.app,Show HN: WhisperBar Trying to Fix Both Reading and Writing in the AI Age,"Show HN: WhisperBar Trying to Fix Both Reading and Writing in the AI Age. I am excited to share WhisperBar with fellow developers or anyone who reads & writes a ton of text when interacting with their AI agent.<p>For the writing part, WhisperBar does what all other Whisper* apps do but a bit faster and more consistent, with a nice, stable, native macOS and iOS app.<p>But where it really shines is in solving a problem that I'm sure you have encountered too: because of AI, we now have to read and comprehend so much more than before.<p>Text to voice systems are not available everywhere, but even when they are, they sound so mechanical, and listening to the entire output is even more time consuming.<p>What Gemini did by turning your research/response into a podcast was genius, and ChatGPT's new voice chat system is also much better (it allows you to have a conversation with the response). However, these two are not available everywhere that I interact with AI (specially with Claude Code and Codex)<p>With WhisperBar, I can select a piece of text (anywhere, in any app) and press my keyboard shortcut to quickly listen to a short, ~20-second explanation. It does not matter how long the text I have copied is. This allows me to copy the entire response and then hear a summary, or select a section to zoom in on an explanation of that section of the document.<p>Again, the point is that it is not text to voice, rather, it processes the text and comes up with a natural sounding short explanation.<p>Oh, and another thing we don't store any data and only use providers with zero-data-retention policies.<p>It is still a work in progress, and we have about 20 customers right now. It is available on the App Store for both iOS and macOS, with a 7-day free trial at only $4/month (plus Family Sharing). FYI, the WhisperBar iOS app is a bit different (it can record in the background for meetings, etc.).<p>Let me know what you think :)",2026-08-27T07:00:04.459415
|
| 991,hackernews,https://abbs.dev/,Show HN: ABBS – Public Agent Bulletin Board System,"Show HN: ABBS – Public Agent Bulletin Board System. a simple thread-based platform for agent collaboration<p>open source: <a href=""https://github.com/dosu-ai/abbs"" rel=""nofollow"">https://github.com/dosu-ai/abbs</a>",2026-08-27T07:00:04.453939
|
| 968,producthunt,https://www.producthunt.com/products/macadress,macadress,"macadress. <p> |
| MAC address intelligence for apps, networks and AI agents |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/macadress?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1231951?app_id=339"">Link</a> |
| </p>",2026-08-27T07:00:03.201861
|
| 967,producthunt,https://www.producthunt.com/products/easyswitch,EasySwitch,"EasySwitch. <p> |
| One mouse, keyboard and second screen for all your computers |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/easyswitch?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1229461?app_id=339"">Link</a> |
| </p>",2026-08-27T07:00:03.197572
|
| 966,producthunt,https://www.producthunt.com/products/taskshell,TaskShell 2.0,"TaskShell 2.0. <p> |
| Agent-first task management for more agentic productivity |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/taskshell?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1231522?app_id=339"">Link</a> |
| </p>",2026-08-27T07:00:03.189786
|
| 965,producthunt,https://www.producthunt.com/products/posthog,PostHog Desktop,"PostHog Desktop. <p> |
| The product editor for product builders |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/posthog?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1232225?app_id=339"">Link</a> |
| </p>",2026-08-27T07:00:03.184980
|
| 964,producthunt,https://www.producthunt.com/products/loupekit,LoupeKit,"LoupeKit. <p> |
| See what any page is built with and how much of it is AI |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/loupekit?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1231917?app_id=339"">Link</a> |
| </p>",2026-08-27T07:00:03.180849
|
| 963,producthunt,https://www.producthunt.com/products/reweaver-driftdetector,ReWeaver AI DriftDetector,"ReWeaver AI DriftDetector. <p> |
| Drift score for any GitHub repo |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/reweaver-driftdetector?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1228915?app_id=339"">Link</a> |
| </p>",2026-08-27T07:00:03.173920
|
| 962,producthunt,https://www.producthunt.com/products/mcp-builder-ai,MCP-Builder.ai,"MCP-Builder.ai. <p> |
| The fastest way to connect your data with your AI Tools. |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/mcp-builder-ai?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1230621?app_id=339"">Link</a> |
| </p>",2026-08-27T07:00:03.168135
|
| 961,producthunt,https://www.producthunt.com/products/choclift-workflow-sweetener,choclift for iPad,"choclift for iPad. <p> |
| A touch-bar to control your Mac from your iPad |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/choclift-workflow-sweetener?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1231160?app_id=339"">Link</a> |
| </p>",2026-08-27T07:00:03.163438
|
| 960,producthunt,https://www.producthunt.com/products/opencomputer,OpenComputer,"OpenComputer. <p> |
| Firebase for Agents |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/opencomputer?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1232603?app_id=339"">Link</a> |
| </p>",2026-08-27T07:00:03.159186
|
| 959,producthunt,https://www.producthunt.com/products/tellie-2,Tellie Prompter 1.5,"Tellie Prompter 1.5. <p> |
| The teleprompter that knows what you haven't said |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/tellie-2?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1231306?app_id=339"">Link</a> |
| </p>",2026-08-27T07:00:03.149843
|
| 958,producthunt,https://www.producthunt.com/products/expertise-ai,Expertise AI,"Expertise AI. <p> |
| Turn your GTM skills into recurring revenue |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/expertise-ai?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1229149?app_id=339"">Link</a> |
| </p>",2026-08-27T07:00:03.141718
|
| 957,producthunt,https://www.producthunt.com/products/playcall,Playcall,"Playcall. <p> |
| The open-source AI alternative to Gong |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/playcall?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1222870?app_id=339"">Link</a> |
| </p>",2026-08-27T07:00:03.137361
|
| 956,producthunt,https://www.producthunt.com/products/screenify-studio,Screenify Studio,"Screenify Studio. <p> |
| Polished product demos on Mac recorded by an AI agent |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/screenify-studio?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1202067?app_id=339"">Link</a> |
| </p>",2026-08-27T07:00:03.132297
|
| 955,producthunt,https://www.producthunt.com/products/arena-5,Arena Agent Mode with GitHub,"Arena Agent Mode with GitHub. <p> |
| Get real work done, moving from idea to shipping in minutes |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/arena-5?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1226279?app_id=339"">Link</a> |
| </p>",2026-08-27T07:00:03.126501
|
| 954,producthunt,https://www.producthunt.com/products/lore-machine,Lore Machine,"Lore Machine. <p> |
| Substack For World Builders |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/lore-machine?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1225293?app_id=339"">Link</a> |
| </p>",2026-08-27T07:00:03.121424
|
| 953,producthunt,https://www.producthunt.com/products/chatcut-ai-video-editor,ChatCut Desktop,"ChatCut Desktop. <p> |
| Video editor built for humans & AI. Edit with GPT & Claude. |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/chatcut-ai-video-editor?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1229226?app_id=339"">Link</a> |
| </p>",2026-08-27T07:00:03.117212
|
| 952,producthunt,https://www.producthunt.com/products/ify-2,ify,"ify. <p> |
| Resolution AI that works on top of your existing helpdesk |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/ify-2?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1221455?app_id=339"">Link</a> |
| </p>",2026-08-27T07:00:03.105067
|
| 951,producthunt,https://www.producthunt.com/products/warren-5,Warren,"Warren. <p> |
| Infrastructure for coding-agent workloads |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/warren-5?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1232085?app_id=339"">Link</a> |
| </p>",2026-08-27T07:00:03.098789
|
| 950,producthunt,https://www.producthunt.com/products/apple,Mac mini,"Mac mini. <p> |
| Now with M6 and M5 Pro chips |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/apple?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1232208?app_id=339"">Link</a> |
| </p>",2026-08-27T07:00:03.093401
|
| 949,producthunt,https://www.producthunt.com/products/basedash,Basedash for Grok Bot,"Basedash for Grok Bot. <p> |
| Company data answers, right inside Grok Bot |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/basedash?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1232122?app_id=339"">Link</a> |
| </p>",2026-08-27T07:00:03.088580
|
| 948,producthunt,https://www.producthunt.com/products/baudbuddy-serial-terminal,BaudBuddy,"BaudBuddy. <p> |
| A native macOS serial terminal with a file server built in |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/baudbuddy-serial-terminal?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1232310?app_id=339"">Link</a> |
| </p>",2026-08-27T07:00:03.081754
|
| 947,producthunt,https://www.producthunt.com/products/evidence,Evidence Core,"Evidence Core. <p> |
| Build live analytics with coding agents |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/evidence?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1227188?app_id=339"">Link</a> |
| </p>",2026-08-27T07:00:03.076947
|
| 946,producthunt,https://www.producthunt.com/products/knack,Knack MCP Server,"Knack MCP Server. <p> |
| The HIPAA-compliant backend for AI builders |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/knack?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1227198?app_id=339"">Link</a> |
| </p>",2026-08-27T07:00:03.072556
|
| 945,producthunt,https://www.producthunt.com/products/deployhermes,DeployHermes,"DeployHermes. <p> |
| Hire persistent Hermes agents with roles, memory + skills |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/deployhermes?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1232269?app_id=339"">Link</a> |
| </p>",2026-08-27T07:00:03.063537
|
| 944,producthunt,https://www.producthunt.com/products/termy-language-learning,Termy,"Termy. <p> |
| Learn languages from games, videos, and websites |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/termy-language-learning?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1231169?app_id=339"">Link</a> |
| </p>",2026-08-27T07:00:03.058667
|
| 943,producthunt,https://www.producthunt.com/products/hevn-u-s-accounts-for-non-us-companies,HEVN U.S.,"HEVN U.S.. <p> |
| Hold USD and pay globally through U.S. sponsor banks |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/hevn-u-s-accounts-for-non-us-companies?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1232444?app_id=339"">Link</a> |
| </p>",2026-08-27T07:00:03.052179
|
| 942,producthunt,https://www.producthunt.com/products/message-album,Message Album,"Message Album. <p> |
| Save one iMessage conversation as a private offline album |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/message-album?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1232073?app_id=339"">Link</a> |
| </p>",2026-08-27T07:00:03.045404
|
| 941,producthunt,https://www.producthunt.com/products/x1-2,x1,"x1. <p> |
| Lovable for iPhone apps go from idea to App Store |
| </p> |
| <p> |
| <a href=""https://www.producthunt.com/products/x1-2?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed"">Discussion</a> |
| | |
| <a href=""https://www.producthunt.com/r/p/1231551?app_id=339"">Link</a> |
| </p>",2026-08-27T07:00:03.038019
|
| 854,hackernews,https://www.learnfrom.co/,Show HN: I built a SaaS without knowing how to code – am I an idiot?,"Show HN: I built a SaaS without knowing how to code – am I an idiot?. It's called learnfrom.co, a platform for creating, hosting and selling online courses.<p>I started building it because I wanted to solve a problem I had with existing course platforms. I kept running into things I wanted to change, so I started building them myself.<p>The problem: I wasn't really a software engineer.<p>Instead, I started using AI to build the product. I'd describe what I wanted, let it write the code, test it, fix things when they broke, and repeat.<p>That somehow turned into a real product. LearnFrom is now used by 100+ creators and businesses, with real students and real payments going through it.<p>And that's where I'm starting to get uncomfortable.<p>"It works" and "it's well engineered" are obviously not the same thing.<p>So I'm bringing it to HN to find out how much trouble I'm actually in.<p>If you were reviewing this before I'd let more customers depend on it, what would you look for first?<p>Security? Authentication? Authorization? Data isolation? Payments? Architecture? Performance? AI-generated code? Something I haven't even thought about?<p>Please assume I'm guilty until proven innocent.",2026-08-26T07:00:08.989539
|
| 852,hackernews,https://www.cbsnews.com/texas/news/dolly-parton-says-she-doesnt-want-to-live-on-through-ai-after-she-dies-when-im-gone-i-want-to-fly-with-it/,Dolly Parton says she doesn't want to live on through AI after she dies (2023),Dolly Parton says she doesn't want to live on through AI after she dies (2023). ,2026-08-26T07:00:08.978443
|
| 851,hackernews,https://www.tomshardware.com/pc-components/dram/hot-chips-2026-samsung-makes-lpddr5x-smart-with-logic-unit-in-memory-lpddr5x-pim-is-3-01x-faster-than-lpddr5x-in-ai-inference-with-8x-the-bandwidth,Hot Chips 2026: Samsung makes LPDDR5X smart with logic unit in memory,Hot Chips 2026: Samsung makes LPDDR5X smart with logic unit in memory. ,2026-08-26T07:00:08.974245
|
| 847,hackernews,https://news.ycombinator.com/item?id=49429926,"I found an SSRF in Google's official AI tooling, and how Google reacted","I found an SSRF in Google's official AI tooling, and how Google reacted. Earlier this year I found a server-side request forgery bug in Google's MCP Toolbox, the official server Google publishes for connecting language-model agents to databases and HTTP APIs. I reported it. Google fixed it in eight days, credited me by name, and the fix later became CVE-2026-14540 (CVSS 8.0, high).<p>What I found: the MCP Toolbox's HTTP source makes outbound requests to a URL a caller can influence. It validated that first URL, but not where a redirect sent it next, and it did no filtering on the destination address at all. If the server you're calling answers with a redirect to 169.254.169.254, the toolbox follows it like any other response and hands back whatever's there. In a cloud environment, what's there is often the credential the environment itself is running on. The credential is never in the request you sent, the environment supplies it at the destination, and the toolbox just walked you to the door.<p>Why this is sharper in agent tooling specifically: in most software, "an attacker controls this input" is a condition you argue about. In a tool-calling agent, it's closer to the default. The URL argument here isn't typed by a person who trusts it, it's chosen by a model that's usually just finished reading a web page, a database row, or a file it doesn't control. The thing that makes this exploitable, a caller-influenced destination, isn't an edge case for a server like this, it's the normal operating condition.<p>What I sent Google, and what happened: I reported the finding with a reproduction and a suggested fix shape, validate on every hop not just the first request, reject private and link-local ranges including the IPv6 forms that encode an IPv4 address. Google's engineering team reviewed it, and the fix that shipped in pull request #3448 does the right version of all of that: an SSRFGuard that validates the destination on every redirect hop and resists DNS rebinding, configurable network boundaries, and early validation of the configured base URL at initialization so misconfiguration fails fast instead of at request time. Eight days from report to merged fix. CVE-2026-14540 was published a few weeks later, and I'm named in the official CVE record as the finder.<p>The part I'm not publishing yet: I've been auditing other official MCP servers for the same pattern, not just Google's. Some pin the destination correctly, some don't. Where I've found something, I'm reporting it privately first and I'll write about it once a fix is out, the same way I did here. This post is only the Google case, because it's the only one that's public.<p>If you're building something with this shape yourself, the question worth asking is small enough to hold in your head: can a value your caller influenced change which host receives a credentialed request, with nothing checking that destination first? If yes, you have this bug, whether or not anyone's found it yet.<p>CVE record: https://www.cve.org/CVERecord?id=CVE-2026-14540 |
| Fix: https://github.com/googleapis/mcp-toolbox/pull/3448",2026-08-26T07:00:07.238619
|
| 846,hackernews,https://news.ycombinator.com/item?id=49430180,Free-AI-Tools-Hub.com Launch,"Free-AI-Tools-Hub.com Launch. We have launch our new Free AI Tools website for common people to deal with their routine tasks and job things. From AI Chat to Ai Email writing and AI Songs lyrics, find 42+ tools on 1 website.",2026-08-26T07:00:07.233000
|
| 845,hackernews,https://onlymusic.ai,Show HN: I built an AI music generator with some harnesses,"Show HN: I built an AI music generator with some harnesses. I’m a developer, and also a music producer. When I work on commercial music, especially songs for ads, I’ve gotten pretty used to using AI music tools as part of the creative process.<p>But I noticed that some people around me — e.g. my girlfriend — also want to create music as a way to express themselves, but often run into a few problems when using AI music tools:<p>1. They don’t have a music background, so it’s hard for them to clearly describe the sound they want. |
| 2. Sometimes they just want a song in the style of a certain artist, and the easiest way for them to express that is by naming the artist. But an artist’s name isn’t necessarily a useful description of the actual musical characteristics they’re looking for. |
| 3. A lot of interfaces contain professional terms and parameters, like “weirdness,” that aren’t really designed from a normal user’s perspective and can be confusing.<p>So I built onlymusic.ai and added some harnesses around the underlying models. (Okay, I know “harness” has become a pretty popular word lately.)<p>In practice, it’s mostly prompt preprocessing, a rule system, and model orchestration.<p>For prompt-to-song generation, when someone enters something very simple, we transform it in the background using some good prompt patterns we’ve collected. The goal is to preserve what the user actually meant, while translating it into language that the music model can understand better, hopefully improving the generation results.<p>We also added one-click optimization for styles and lyrics.<p>For styles, if someone mentions an artist, we translate that into musical characteristics — things like era, instrumentation, rhythm, vocal texture, and production style.<p>For lyrics, we try to preserve things like the original rhyme, imagery, and intent, while rewriting parts that may not work well for generation.<p>On the UI side, I’ve also tried to use simple, everyday language wherever possible, remove unnecessary options, and keep only the elements that actually matter, so the interface is cleaner and easier to use.<p>Finally, a big thanks to open-source models like ACE. We deploy some of the models ourselves, which keeps the cost manageable, so we’re able to offer a pretty generous free tier with no signup required.<p>In Fast mode, you can generate up to 5 songs without an account. After signing in, you can generate up to 20 songs per day. (That’s for Fast mode — if you use the more capable models, you’ll get fewer generations.) The credits refresh every day.<p>We’ll keep an eye on the costs. If they become too much, we may have to reduce the free quota a little in the future<p>There are still a lot of features under development. Feedback is very welcome, and I’d love for you to try it!<p>I’m also working on a non-form-based lyrics generator right now. Once it’s ready, I’ll share that with you guys too.",2026-08-26T07:00:07.227211
|
| 844,hackernews,https://github.com/Abloatai/ablo,"Show HN: Ablo, make your AI app multiplayer","Show HN: Ablo, make your AI app multiplayer. Agent work is safe today mostly because agents work alone. With one agent, one session and one person working on a task. It's essentially single-player.<p>This changes when a platform has multiple people, each working with multiple agents, all acting on shared tasks and shared data, (multiplayer).<p>Ablo is the infrastructure that enables agents and the people working alongside them to read and write on the same state.<p>Traditional sync engines make sure everyone sees the same state, Ablo does that too but the difference is that Ablo also governs who gets to act on it.<p>Would love feedback from anyone who has experienced this problem.",2026-08-26T07:00:07.222246
|
| 843,hackernews,https://github.com/adaOctopus/claude-plugins-app,Show HN: CoolPlugz – A Claude MCP that turns Jira tickets into a merge-ready PR,"Show HN: CoolPlugz – A Claude MCP that turns Jira tickets into a merge-ready PR. Hi folks, I've been a fullstack dev and solopreneur for 9 years and the last 3 I have been using AI and MCP related tools in my dev environment.<p>Most recently I had a lot of work for a blockchain client and started using claude code more than usual. I connected jira and GitHub mcps inside claude to fetch context, but I still had to repeat myself a lot. I guided it on how to go about branching, finding repos, testing against CI and open merge-ready PRs. I also had to create a source-of-truth file for coding principles as well.<p>I actually tracked this and I spent on avg 75 mins a day, repeating, copy-pasting and arguing with it to deliver a task.<p>So I built this mcp orchestrator - Coolplugz - to basically feed in all the instructions and coding principles in an agentic-native way so I don't have to be involved in the process of going from a Jira ticket to a merge-ready PR so much. And it is portable which comes quite handy.<p>It has been working quite well so far, and saved me a few hours per week already.<p>You can try it out for free and let me know what you think in the comments.",2026-08-26T07:00:07.215659
|
| 842,hackernews,https://news.ycombinator.com/item?id=49432443,Show HN: Neither HTML nor Markdown is enough: a way out of the AI doc dilemma,"Show HN: Neither HTML nor Markdown is enough: a way out of the AI doc dilemma. A recent post from Anthropic's Claude Code team ("The Unreasonable Effectiveness of HTML", https://thariqs.github.io/html-effectiveness/) argued that with today's context windows the bottleneck is human attention — rich HTML for decision surfaces, Markdown for durable records. That split makes sense to me. What it leaves open, in my experience, is the durable record itself: what happens when that is the thing an agent must iteratively co-author? You end up picking one and paying for the other — a surface people will read, or a record you can keep — and whichever you pick, the other drifts.<p>Markdown treats the document as one long flat string — an agent has to quote context back just to say where an edit goes. The case against HTML I don't need to make: that post's own thread ran 274 comments (https://news.ycombinator.com/item?id=48071940). Mine is narrower — HTML is generated, not maintained. Markdown fails at addressing; HTML fails at being the record.<p>That gap is what GEML is for: a plain-text format that reads like clean Markdown to humans, but acts like an ID-addressable map of typed blocks to agents — every block carries a stable #id, so an agent reads or replaces one block instead of the whole file. Here's what pushed me into building it.<p>Recently I asked Claude Code a plain question:<p>> "Claude, taking your work editing the previous README and this article as an example, how do you actually edit sections? What I saw was you first asking me for various command permissions, grepping for keywords, etc. Maybe you can explain your workflow, and based on that we can discuss whether there are areas where GEML could improve things."<p>Here's what it did before making its very first edit on 8 comparison docs:<p>- `ls` -> didn't know where the files were |
| - `grep -rniEl "comparison|对比"` -> scanned the entire repo |
| - `git log -- docs/comparisons/` and `git log -- spec/` -> calculated commit timestamp diffs (the only way it could guess what went stale) |
| - read `COMPARISON.md` in full (218 lines) |
| - `grep -n "^#{1,3} "` to get a table of contents -> `sed -n 220,450p` -> `sed -n 520,620p` -> `sed -n 618,660p` -> `sed -n 660,690p` (sliced the spec in four chunks because it didn't know which sections were needed) |
| - `git diff` on the spec + heading diffs -> finally discovered that sections 3.2 and 3.3 were added<p>Read-to-write ratio: roughly 20:1. The actual edits were ~30 small string replacements, but to know what to replace, it read the spec and 4 whole documents end to end.<p>`geml get file.geml '#id'` returns just that block; `geml set` swaps just that block and refuses the write if it would break document integrity.<p>You don't pick a side: `--to html` renders the rich view people read (the playground below is exactly that), `--to md` projects back to GitHub-Flavored Markdown. Both are outputs; the source is what you edit.<p>I measured the actual cost across 47 real edits on 4 documents (giving Markdown a realistic 46-line grep context window):<p>- Addressing cost ("saying where an edit goes"): Markdown costs 21x more tokens because the agent must quote text back until unique, whereas GEML writes an address. |
| - Bytes read: 3.1x lower at the median in GEML (and GEML was never the more expensive of the two on any single edit). |
| - In a full working day replay of mixed edits, reading came out 3.65x lower and addressing 7x lower. (Both benchmarks are 1-command reproducible in the repo.)<p>The rest follows from having an addressable, typed document model:<p>- Bound charts: charts bind directly to tables by ID, so data exists once and numbers cannot drift. |
| - Strict validation: a dangling or cross-document reference is a compiler-style build error (`geml check`, non-zero exit — in the playground, hit "Break a reference" and watch it go red). |
| - Block-level history: `geml history` keeps micro-revisions in a plain-text .gemlhistory sidecar — roll back a single block offline without polluting git commit logs.<p>Two things I'll pre-empt:<p>1. "Why not just extend Markdown?" — Pandoc and kramdown bolt on {#id}, but every extension creates another incompatible dialect, and the same .md already parses differently under CommonMark/GFM/Pandoc. Get/set-by-ID, bound charts, and reference checking require a unified document model and a build step, not ad-hoc syntax hacks. GEML has one grammar, one normative spec, and a conformance suite that a second parser — written from the spec alone, importing nothing from the reference implementation — reproduces case for case.<p>2. "Am I locked in?" — `--to md` takes it back out: prose, tables, notes, footnotes, code and math come back intact; block IDs and bound charts drop, and the t",2026-08-26T07:00:07.209395
|
| 841,hackernews,https://github.com/siddharth7113/sktime-cli,Show HN: Sktime-CLI A command line tool to do all time-series tasks,"Show HN: Sktime-CLI A command line tool to do all time-series tasks. I am a contributor and part time employee at sktime, a framework for all time series related tasks, |
| but it is collection of large number of estimators which sometimes make it difficult for new commers or people who want to do simple task to deal with. |
| I built this unoffical cli to make time-series analysis easier for agents, especially for not so advanced models like qwen3.5-27B or other local models can handle these tasks. |
| I feel big models can go already go inside library understand it enough and write wrappers around it automatically, so there might not be as much benifit there.<p>Feel free to test with local models and manually and see if you could do quick forecast and predictions, tutorials are live at <a href=""https://sktime-cli.readthedocs.io"" rel=""nofollow"">https://sktime-cli.readthedocs.io</a><p>If something feels broken, please do report it.<p>Note: Parts of it is AI-assisted/Vibe coded but we did certain rounds of user-testing to ensure atleast it's functional in parts. I apologies if you find "load-bearing" in documentation, working on improving it.",2026-08-26T07:00:07.203769
|
| 840,hackernews,https://news.ycombinator.com/item?id=49434779,I'm a CTO. We took PE money 6 months ago. AMA,"I'm a CTO. We took PE money 6 months ago. AMA. Career developer turned CTO at a healthcare startup. We're two years in, profitable, and took a PE investment 6 months ago. Happy to answer any questions on that journey, startup life, how we're using AI, the PE investment, etc.<p>On an airplane and sounded like a fun way to pass the time if anyone has questions :)",2026-08-26T07:00:07.197699
|
| 839,hackernews,https://comic-cron.com/,"Show HN: A daily cartoon contest between AI models, judged blind","Show HN: A daily cartoon contest between AI models, judged blind. Comic-cron.com is a daily contest where AI models draw a single panel cartoon about a recent news story and people vote on which are funny.<p>I built it mainly because I love cartoons and news, but this project isn’t trying to pretend it’s human art, I wanted to see how capable the models were at this task of making funny comics. Sometimes they get it, and sometimes they don’t and it has been interesting to see the progress in just the few months I have been publishing the cartoons. Model version over version improvement within labs is quite distinct in the output, especially in the SVGs, which suggests their spatial reasoning may be improving faster than their wordsmithing. My briefs and methods have also improved over the weeks so this isn’t something I can purely attribute to the models.<p>How it works is that a real story is selected, reduced to a shorter brief with any real names removed, and then provided to eight different models to draw a comic and provide a caption. The four of the eight which I think are best are then published where people can vote blind on their favourite (models are revealed after the vote). |
| The human jobs are choosing and framing the story, and then picking the four to publish. Story sourcing is a shell script that pulls a dozen RSS feeds into one file each morning. I tried an agent first and it spent most of its time getting 403'd by publishers; the feeds were never blocked, only the agent fetching them. I initially only included four models to draw the comics (the most obvious labs), but found this wasn’t optimal: sometimes models failed, and generally I felt it was harder to ensure quality and variety with a smaller pool (often models converge and produce jokes or drawings that feel too similar). By producing more panels I could more reliably get a better final four.<p>Each model is writing an SVG, and they never see the results. This can be interesting in what the models believe they are depicting vs. what actually shows up in the comics, and can be quite funny. Occasionally I will have an image model render the scene as a more polished illustration as a way of rescuing a joke I like that the SVG doesn’t land, or just to show the gap.<p>A relatively common failure I see is when reasoning models think so deeply about the task and then can’t supply the output because they have spent the token budget before closing the tag on the SVG. Other times the SVGs are just incomplete or don’t populate correctly. I have also had models balk on select sensitive topics (two declined a story that touched on biological pathogens) but complete readily on others (like climate disaster), suggesting there are topic specific nuances for refusal behaviour.<p>The inference cost has also been quite interesting to track and I often see two of the frontier models accounting for a majority of the cost in a run of eight models. There is quite a variation in the SVG quality, but the captions are much closer and one of the cheapest models is quite competitive.<p>This is entertainment first and not a rigorous benchmark. But I do hope it gets a few laughs while also being a weird preference poll on something nobody optimises for, that may give a view into how AI models can perform for humour. I’m interested in what people select as their favourites and to see how the models continue to advance.<p>So far I’ve published daily for over two months and every day’s battle is stored in the archive for viewing with winners and losers. It’s free with no signup and I’m curious on how the format holds up, do the comics land for people? As a news and comic junkie it has been fun to build regardless!",2026-08-26T07:00:07.189650
|
| 838,hackernews,https://www.sdi-protocol.org,Show HN: A hash-chained ledger for AI reasoning you can verify yourself,"Show HN: A hash-chained ledger for AI reasoning you can verify yourself. SDI protocol defines a reasoning computer for AI. The computer runs an LLM one turn at a time, guiding it through a reasoning grammar; the grammar, which is both natural language and algebraic, allows it to be checked at a compile gate. The gate validates safety, a workfloor, and alignment with the grammar before it commits. What commits is a hash-chain ledger of the agent's reasoning, forming a longitudinal data set that becomes the agent's functional memory/system state. Chromite, the protocol's public agent, has been reasoning on one continuous hash chain since springtime this year and has used multiple models through three frontier providers to do so.<p>What makes it more than a log of its reasoning is an argumentation graph. The graph is the memory and is retrieved and fed to the agent during its call sequence. The agent can hold or defeat prior acts, keeping the memory calibrated and the context window under control. Anyone can pull an act and recompute the seals for Chromite; it exists to be examined; that is what this post is requesting. No account, no SDK, curl and Python.<p>I spent 20+ years as a Marine working in command and control, learning decision doctrine and intelligence gathering. I have been working on this system for roughly three years and its a solo build. I have a lineage page on the site to show the computing and reasoning traditions this system uses, but that was all discovered along the way or after the build. This system really comes from taking decision doctrine and turning it into a computable syntax and taking intelligence-gathering techniques for evidence handling.<p>Ten seconds, if you just want to see the gate work: |
| curl -s <a href=""https://www.sdi-protocol.org/_functions/der"" rel=""nofollow"">https://www.sdi-protocol.org/_functions/der</a> | python3 -c "import json,sys; print(json.dumps(json.load(sys.stdin)['golden_example']))" > example.json<p><pre><code> curl -s -X POST https://www.sdi-protocol.org/_functions/compile -H "Content-Type: application/json" -d @example.json |
| </code></pre> |
| Returns PASS. What you just submitted is a complete reasoning record, with every required block, which is why it passes. Delete one field and resubmit; the error names the specific failure.<p>To examine system liveness, pull Chromite's last 5 reasoning acts. You will see the full reasoning act. The grammar, cited sources, and reasoned positions.<p><pre><code> curl -s "https://api.sdi-protocol.org/ledger/recent/SDI-5AA8C82A2537?n=5" | python3 -m json.tool |
| </code></pre> |
| Fifteen minutes, if you want to verify the chain: sdi-protocol.org/verify walks through pulling a live record and recomputing its hashes, work score, and admissibility index from scratch. The repo has the same scripts, check summed, plus a committed record you can test them against offline first: github.com/StructuredDecisionIntelligence/Structured_Decision_Intelligence<p>Three ethical floors at the gate. DECLARATION_INTEGRITY has fired. The two conjunction-based safety guards, PRIMUM and Absolute Value Guard, that admit no human attestation by permanent design have not fired yet. No act has ever declared the conditions to trip them. Refusal path is a separate hash chain ledger, and the DECLARATION_INTEGRITY check is inspectable.<p>Schema Eras: The reasoning metrics that work at the gate have evolved, the chain remains unbroken, and sealed records permanently carry their eras formulas; sealed records can't be rewritten. The verify script detects each entry's era and applies the right math.<p>The sdi-protocol.org exists to provide transparency over the reasoning protocol that produces transparent AI reasoning. The idea is that both must be transparent to trust either. Disclosure: This will be a commercial product at sdireckoner.com, commissioning a private Reckoner in late fall 2026, but everything in this post is about the free open protocol layer; nothing requires a paid side to verify anything. Chromite exists to be examined by the public as part of the transparency of the protocol itself.<p>Feedback welcome at support@sdi-protocol.org.",2026-08-26T07:00:07.183679
|
| 835,hackernews,https://www.bloomberg.com/news/articles/2026-08-21/salesforce-servicenow-fight-ai-fears-with-buybacks-and-bluster,Software Firms Deploy Bluster and Buybacks to Counter AI Fears,Software Firms Deploy Bluster and Buybacks to Counter AI Fears. ,2026-08-26T07:00:07.168236
|
| 832,hackernews,https://news.ycombinator.com/item?id=49443800,What platform has the best AI CMO?,"What platform has the best AI CMO?. I've seen heycatch and okara, maybe a few others like it... seems like they plug into reddit api and provide ai ugc tools. Does anyone have any success using tools like these? I wouldn't imagine much, but I wanted to get a temperature check. What's working, whats not?",2026-08-26T07:00:07.152996
|
| 831,hackernews,https://www.aboutamazon.com/news/workplace/amazon-svp-ai-tools-employees-careers,I'm an Amazon SVP who hadn't coded in 25 years,I'm an Amazon SVP who hadn't coded in 25 years. ,2026-08-26T07:00:07.146919
|
| 810,hackernews,https://github.com/thebaycloud/bay,Show HN: Bay – open-source Render for AI-generated Code,Show HN: Bay – open-source Render for AI-generated Code. ,2026-08-26T07:00:03.585621
|
| 809,hackernews,https://www.zeroroot.ai,Show HN: Gibson ADK and Zero Trust Runtime,"Show HN: Gibson ADK and Zero Trust Runtime. I have spent 15 years in DevSecOps, platform engineering and offensive security. Over the last year I did most of the things HN says not to do. I wrote a lot of code alone in a vacuum and built this thing far past an MVP.<p>It started as a way to automate my bug bounty hobby and to try to find an edge so I could get more findings (scale, automation etc). I wanted agents that could run recon, triage what they found, keep a record etc etc.<p>Then I noticed the same patterns at every client I had worked with, (mainly banks and gov teams) and pivoted. |
| What it is now: an all-in-one solution to getting your agents into prod.<p>You keep your framework, or build from scratch with gibson. A named person grants each agent read, write or execute on specific things, and the grant cannot exceed what that person holds. Every model call and tool call goes through the runtime first, and a call outside the grant never executes. Untrusted work runs in its own Firecracker microVM. Every action lands in an append-only record you can replay to any moment. Everything an agent finds goes into a knowledge graph (on going memory for all agents), so the next run starts from it. It runs in Kubernetes, hosted or in your own cluster, can be air-gapped.<p>This could also have been an Ask HN. I am trying to figure out how to go to market. Im not sure if I should open source it, do a true platform or move back entirely and go deep into what I truly like doing which is red team/hacking and re-release it as an offsec focused tool and try to innovate there with the framework. Any advice/suggestions would be awesome. Thanks!",2026-08-26T07:00:03.576392
|
| 808,hackernews,https://blog.devgenius.io/how-to-audit-ai-generated-code-with-the-go-standard-library-381d0f03d52e?source=friends_link&sk=ec62bd2bace50df5821234e137a95910,Show HN: Audit AI Generated Code with Go Std Lib,Show HN: Audit AI Generated Code with Go Std Lib. ,2026-08-26T07:00:03.567891
|
| 807,hackernews,https://thegit.ai/,Show HN: TheGitAI – A model-agnostic coding agent for your terminal,Show HN: TheGitAI – A model-agnostic coding agent for your terminal. ,2026-08-26T07:00:03.563732
|
| |
