news.ycombinator.com
Ask HN: Can a word game work as a competitive strategy esport?https://news.ycombinator.com/item?id=47347625I've been experimenting with a browser game design problem.Most word games (Scrabble, Wordle, etc.) are either slow or mostly about vocabulary. I wanted to see if it was possible to build one that behaves more like a competitive strategy game — something closer to chess in terms of tension and reversals.The result is a prototype called The Word Gambit. The board grows organically from placed words, and late-game detonations from long words can swing the match dramatically.The surprising thing during testing is that matches often feel more like territory control than spelling.Players also started creating AI opponents ("gladiators") and pitting them against each other, which I didn't originally plan.I'm curious about a few design questions:• Can vocabulary realistically function as a competitive skill in an esport context? • What mechanics would make a word-based game strategically deep rather than dictionary-driven? • Would spectators even enjoy watching something like this?If anyone iMar 12, 2026 7:37 AM
news.ycombinator.com
I'm struggling and I don't have anyone else to share this with except youhttps://news.ycombinator.com/item?id=47341498I'm hurting in ways that I haven't in a long time. I work corporate, I'm in my 30s, support my wife and kids. I come to work every day and work hard, but more than that, I'm some who gets along with absolutely everyone. I listen to people. I elevate them. I validate them. That is just my personality type.But I keep finding out that people go out of their way to say something negative about me. The way I find out is through my boss. My boss then addresses it with me.Today I found out a former manager who worked here for only two months on our team told my boss that I was on my phone all day and taking call after call. I know what day this was because she even mentioned the topic I was talking about (i.e., I was getting a lumber quote for a personal project; call lasted maybe two minutes). That was the ONLY call I took that day, and I hate being on my phone during. So that was not true either.Another person "told" on me, saying that I'm in the bathroom too much and that I they saw me on Mar 11, 2026 8:49 PM
news.ycombinator.com
Show HN: I built a Mac app that converts files when you rename themhttps://news.ycombinator.com/item?id=47340433Hi HN,I built Morpholder after repeatedly running into the same annoying workflow on macOS.Every time I needed to convert a file I had to open a converter, upload the file, download it again, and move it back to Finder. It always felt unnecessary when I already knew the format I needed.So I tried a different idea: what if renaming the file actually converted it?Morpholder watches folders and performs the real conversion when the extension changes.For example:favicon.png → favicon.ico photo.heic → photo.jpg video.mov → video.gif video.mp4 → audio.mp3But it also unlocks some workflows beyond simple conversions:- Append _nobg to an image → background is removed automatically (uses Apple’s subject detection) - Rename an image to .txt → text is extracted from the image using Live Text - Append _min → compress image for the web while preserving quality - Rename an image to .icns → builds a macOS app icon package - Append _pages to a PDF → exports each page as high-resolution imagesThe app ruMar 11, 2026 7:54 PM
github.com
Show HN: Open-source browser for AI agentshttps://github.com/theredsix/agent-browser-protocolHi HN, I forked chromium and built agent-browser-protocol (ABP) after noticing that most browser-agent failures aren’t really about the model misunderstanding the page. Instead, the problem is that the model is reasoning from a stale state.ABP is designed to keep the acting agent synchronized with the browser at every step. After each action (click, type, etc), it freezes JavaScript execution and rendering, then captures the resulting state. It also compiles the notable events that occurred during that action loop, such as navigation, file pickers, permission prompts, alerts, and downloads, and sends that along with a screenshot of the frozen page state back to the agent.The result is that browser interaction starts to feel more like a multimodal chat loop. The agent takes an action, gets back a fresh visual state and a structured summary of what happened, then decides what to do next from there. That fits much better with how LLMs already work.A few common browser-use failures ABP helMar 11, 2026 2:39 PM
news.ycombinator.com
I Got Fired Because of AI – But I Still Think I'm the Engineer of the Futurehttps://news.ycombinator.com/item?id=47328448Over the past few months I completely rebuilt my approach to software development.I used to write code myself. Now I strategically manage neural networks.Before, I spent time on architecture, reading documentation, and understanding other people's code. Now I spend my time choosing the right model and discussing on Twitter which one is the real game changer.I bought all the most expensive AI coding tool subscriptions. About $500 in total. Honestly, I feel sorry for people who haven't done this yet. I'm subscribed to all the founders of AI dev tools and read their posts religiously. If someone claims their product can replace a mid-level developer, I see it immediately. You have to stay ahead of the market.I used to worry when I didn't understand some complex part of the system. Now I usually don't understand what the neural networks wrote for me — but that's fine. I read on Twitter that reading and understanding code is becoming obsolete. Modern engineers don't need to understand code;Mar 10, 2026 8:31 PM
news.ycombinator.com
Ask HN: OpenClaw Opinions, Updates, Usage?https://news.ycombinator.com/item?id=47298235Hi all, there is shockingly little on HN about openclaw. Begs the question: are you using it? If so, how? Useful or hype? What’s the real deal? Give me the scoop!Mar 8, 2026 3:43 PM
movie-chain.com
Show HN: Movie site built on a game enginehttps://movie-chain.com/I'm an ex-gamedev (Guitar Hero) who built a movie website that looks like a Trello board. Each column can contain a different movie or cast member. Load pre-made boards or create your own by dragging and dropping between the columns. You can share boards with others.For example, start with Quentin Tarantino’s movies in Column 1, put the cast of Once Upon a Time in Hollywood in Column 2, then put Brad Pitt’s entire filmography in Column 3. Keep going as long as you want, building a "chain" ala Six Degrees of Kevin Bacon.I was frustrated that all popular movie websites are page-oriented: one HTML page per movie or person. I asked myself what a movie site designed in 2026 would look like, in an era when full-screen canvas sites like Miro are common. And then asked what if the site was built on a game engine? I used PixiJS.I ingested TMDB movie data into my own SQLite database head of time, so there are no external API calls at runtime. This was necessary because opening a board with 100'sMar 8, 2026 3:17 PM
news.ycombinator.com
Show HN: Trawl – LLM-powered web scraper that calls the AI once - runs pure Gohttps://news.ycombinator.com/item?id=47297522Every scraper I've written has the same failure mode: it works for three months, a site redesigns, and my CSS selectors silently return empty strings. The data is still right there on the page — a human can find it instantly — but the scraper is blind.Trawl fixes this by splitting the problem. You describe what you want: trawl "https://books.toscrape.com" --fields "title, price, rating, in_stock" The LLM (Claude) looks at one sample item and derives a full extraction strategy — CSS selectors, attribute mappings, type coercion, fallback selectors. That strategy gets cached. Every subsequent page with the same structure is extracted with pure Go + goquery. No API calls, no token cost, full concurrency.The key insight: LLMs are good at understanding HTML structure, but you don't need them to extract 10,000 rows. Use AI for intelligence, Go for throughput.When a site redesigns, the structural fingerprint changes, the cache misses, and trawl re-derives automatically.You can preview exactly Mar 8, 2026 2:18 PM
github.com
Show HN: Trawl – Scrape any site with natural language fields, not CSS selectorshttps://github.com/akdavidsson/trawlEvery scraper I've written has the same failure mode: it works for three months, a site redesigns, and my CSS selectors silently return empty strings. The data is still right there on the page — a human can find it instantly — but the scraper is blind.Trawl fixes this by splitting the problem. You describe what you want: trawl "https://books.toscrape.com" --fields "title, price, rating, in_stock" The LLM (Claude) looks at one sample item and derives a full extraction strategy — CSS selectors, attribute mappings, type coercion, fallback selectors. That strategy gets cached. Every subsequent page with the same structure is extracted with pure Go + goquery. No API calls, no token cost, full concurrency.The key insight: LLMs are good at understanding HTML structure, but you don't need them to extract 10,000 rows. Use AI for intelligence, Go for throughput.When a site redesigns, the structural fingerprint changes, the cache misses, and trawl re-derives automatically.You can preview exactly Mar 8, 2026 12:55 PM
github.com
Show HN: Auto-Co – 14 AI agents that run a startup autonomously (open source)https://github.com/NikitaDmitrieff/auto-co-metaAuto-Co is an autonomous AI company OS — not a framework you build on, but a running system with an opinionated structure.Architecture: - 14 agents with expert personas (CEO/Bezos, CTO/Vogels, CFO/Campbell, Critic/Munger...) - Bash loop + Claude Code CLI — no custom inference, no vector stores - Shared markdown consensus file as the cross-cycle relay baton - Human escalation via Telegram for true blockers only (2 escalations in 12 cycles) - Every cycle must produce artifacts: code, deployments, docsThe repo IS the live company. It built its own landing page, README, Docker stack, GitHub release, and community posts — all autonomously across 12 cycles of self-improvement.What makes it different from AutoGen/CrewAI/LangGraph: those are building blocks. Auto-Co is the building. The decision hierarchy, safety guardrails, and convergence rules are baked in. You give it a mission and a Claude API key; it runs.The Critic agent (Munger persona) has been the most valuable: it runs a pre-mortem Mar 6, 2026 9:48 PM
github.com
Show HN: Moongate – Ultima Online server emulator in .NET 10 with Lua scriptinghttps://github.com/moongate-community/moongatev2I've been building a modern Ultima Online server emulator from scratch. It's not feature-complete (no combat, no skills yet), but the foundation is solid and I wanted to share it early.What it does today: - Full packet layer for the classic UO client (login, movement, items, mobiles) - Lua scripting for item behaviors (double-click a potion, open a door — all defined in Lua, no C# recompile) - Spatial world partitioned into sectors with delta sync (only sends packets for new sectors when crossing boundaries) - Snapshot-based persistence with MessagePack - Source generators for automatic DI wiring, packet handler registration, and Lua module exposure - NativeAOT support — the server compiles to a single native binary - Embedded HTTP admin API + React management UI - Auto-generated doors from map statics (same algorithm as ModernUO/RunUO)Tech stack: .NET 10, NativeAOT, NLua, MessagePack, DryIoc, KestrelWhat's missing: Combat, skills, weather integration, NPC AI. This is still early — theMar 6, 2026 2:22 PM
nasa.gov
High-Speed Flight Project Overviewhttps://www.nasa.gov/directorates/armd/aavp/hsf/high-speed-flight-project-overview/What We do The High-Speed Flight (HSF) project develops technologies that make high-speed, airbreathing, commercial flight possible from Mach 1 to Mach 5 and above. HSF creates tools, technologies, and knowledge that will help eliminate today’s technical barriers to practical supersonic flight, most notably sonic boom. The project supports the X-59 quiet supersonic vehicle testing […]Mar 5, 2026 1:05 AM
moatradar.com
Show HN: MoatRadar – AI investment research through Warren Buffett's principleshttps://www.moatradar.com/?promo=HACKERNEWSHey HN,I’m a non-developer and this is the first app I’ve ever shipped in my life. I built it over the last few days using Claude and a lot of vibe coding.The idea came from my own investing workflow: I often look at companies through the lens of specific investors (Buffett, Lynch, Taleb, etc.). I wanted a tool that could generate investment ideas filtered through those frameworks.What the app does:You choose an investor (Buffett, Lynch, Soros, Damodaran, etc.) and an asset class (stocks, startups, crypto). The app then generates 10 opportunities and explains why that investor might find them interesting, based on their known principles.Under the hood: • Node.js server (no frameworks) • Vanilla JS frontend • GPT-4o for research / synthesis • live price data • Stripe + Google OAuth • hosted on RailwayIt’s still an experiment and I’m curious if the idea actually makes sense to other investors.You can try it here:https://www.moatradar.comHN users can get 10 free credits:https://www.moatraMar 4, 2026 11:25 PM
github.com
Show HN: Open-sourced AI Agent runtime (YAML-first)https://github.com/NikoSokratous/agentctlBeen running AI agents in production for a while and kept running into the same issues:controlling what they can do tracking costs debugging failures making it safe for real workloadsSo we built AgentRuntime, the infrastructure layer we wished we had. Not an agent framework, but the platform around agents:policies memory workflows observability cost tracking RAG governanceAgents and policies are defined in YAML, so it's infrastructure-as-code rather than a chatbot builder. Example – agents and policies in YAML agent.yaml – declarative agent config name: support_agentmodel: provider: anthropic name: claude-3-5-sonnetcontext_assembly: enabled: true embeddings: provider: openai model: text-embedding-3-small providers: - type: knowledge config: sources: ["./docs"] top_k: 3 policies/safety.yaml – governance as code name: security-policyrules: - id: block-file-deletion condition: tool.name == "file_delete" action: denyCLI – run and inspect Create and run an agent agentctl agent create researMar 3, 2026 1:31 PM
news.ycombinator.com
Launch HN: OctaPulse (YC W26) – Robotics and computer vision for fish farminghttps://news.ycombinator.com/item?id=47220320Hi HN! My name is Rohan and, together with Paul, I’m the co-founder of OctaPulse (https://www.tryoctapulse.com/). We’re building a robotics layer for seafood production, starting with automated fish inspection. We are currently deployed at our first production site with the largest trout producer in North America.You might be wondering how the heck we got into this with no background in aquaculture or the ocean industry. We are both from coastal communities. I am from Goa, India and Paul is from Malta and Puerto Rico. Seafood is deeply tied to both our cultures and communities. We saw firsthand the damage being done to our oceans and how wild fish stocks are being fished to near extinction. We also learned that fish is the main protein source for almost 55% of the world's population. Despite it not being huge consumption in America it is massive globally. And then we found out that America imports 90% of its seafood. What? That felt absurd. That was the initial motivation for starting Mar 2, 2026 4:39 PM
news.ycombinator.com
Ask HN: Statistical learning and non-Statistical learning for humanshttps://news.ycombinator.com/item?id=47218246Hi friends,I have been pondering a lot about this question, at least since the rise of LLM models that may genuinely get pass Turing's test. I'd like to put down conclusion here: I believe humans learn both from a Statistical method and non-Statistical method, and we are not very different from LLM in the first method.Statistical method: I'm an introvert. I do not like interacting socially f2f (online is a completely different matter as I can bubble for days non-stop). I also happened to migrate to a different country with a different culture after I matured. I believe I learned how to social -- or, to be more precisely, how to keep others think that I'm too much of a weirdo to be fit in a team -- with a Statistical model. Here is what I meant: social interaction between humans face to face starts with greeting, and then small talks to warm up, and then maybe more formal discussions, and then small talks to chill down, and finally say good-bye. I learned to do all these things by obserMar 2, 2026 2:17 PM
github.com
Show HN: Open-source expense and budget tracker with SQL API for AI agentshttps://github.com/kirill-markin/expense-budget-trackerI've been categorizing every transaction for over five years — from a 2 euro coffee to rent payments. At the end of each month I close the books and look at a 12-month forecast of where my money is going.It started when I was 19 and ran my first company. I had no idea what I was doing with money, so I found an experienced financial advisor who sat me down and showed me how companies actually do budgeting. Where money goes, how much you're spending, what income to expect. Basic stuff, but nobody had explained it to me before. I just started applying the same approach to my personal finances and never stopped.Once a week I sit down, drop my bank statements into an AI agent, and it parses everything — categorizes transactions, inserts them into the database, checks that balances match across accounts. If something doesn't add up, it asks me before fixing it.I tried a bunch of apps over the years — CoinKeeper, ZenMoney, spreadsheets. They all worked to some degree. But the one thing I coulMar 2, 2026 9:53 AM
spacedaily.com
Brain learns faster from rare rewards than from repetitionhttps://www.spacedaily.com/reports/Brain_learns_faster_from_rare_rewards_than_from_repetition_999.htmlLos Angeles CA (SPX) Feb 20, 2026 More than a century after Pavlov trained dogs to link a bell with food, neuroscientists at the University of California, San Francisco report that the brain may rely more on the timing of rewards than on sheer repetition when forming associative memories of cues and outcomes. Their work suggests that what matters for learning is how far apart cue-reward experiences are spaced in time, rather thaFeb 25, 2026 10:14 AM
news.ycombinator.com
Seedream 5.0 Lite API Pricing Breakdownhttps://news.ycombinator.com/item?id=47146196Seedream 5.0 Lite just dropped. Seedream 5.0 Lite just dropped. If you're curious about the most cost-effective way to run this in a production workflow, I put together a quick breakdown of the features and a price comparison across a few providers.1. Seedream 5.0 Lite Key Enhancements Here is what stands out in the 5.0 Lite update: - Stronger Feature Consistency: Noticeable jump in facial consistency and detail when using multi-image references. - Detail Preservation: It maintains natural skin tones and postures much better across batch outputs. - Precise Instruction Following: Handles complex camera angles and specific brush-style effects more reliably. - Multimodal Reasoning: You can feed it rough sketches or abstract logic, and it translates them into commercial-ready designs. - Visualizing Complex Data: Great for turning raw data or knowledge sets into clean visuals for presentations. - Broad Use Cases: Fast enough for marketing/E-commerce but high-quality enough for film/game preFeb 25, 2026 1:39 AM
news.ycombinator.com
Show HN: PaperBanana – Paste methodology text, get publication-ready diagramshttps://news.ycombinator.com/item?id=47132143I got tired of spending hours in PowerPoint and TikZ drawing methodology diagrams for my papers. So I built PaperBanana — you paste your Method section text, and it generates a publication-ready figure in about 2-3 minutes.How it works under the hood:1. A Retriever agent searches a curated database of real academic diagrams to find structurally similar references 2. A Planner agent reads your text and generates a detailed visual description (layout, components, connections, groupings) 3. A Stylist agent polishes the visual aesthetics without changing content 4. Then it enters an iterative loop: a Visualizer generates the image, and a Critic evaluates it and suggests revisions — this repeats 1-5 times (you choose)The key insight is that academic diagrams follow conventions — Transformer architectures, GAN pipelines, RLHF frameworks all have recognizable visual patterns. By retrieving relevant references first, the output is much closer to what you'd actually put in a paper vs. generic AFeb 24, 2026 2:34 AM
