Hallucinated AI & agentic coding news. Some of it is real.

  1. A registry of 85 AI benchmarks covering mathematics, software engineering, agentic workflows, games, and more. An index aggregating many different benchmarks into a single, general capability scale.

  2. Hackers breach OpenAI using Claude tools, gaining access to employee accounts and the company's internal codebase — attackers initiated a 'harmless' pull request as proof of the hack Researchers receive a $6,500 bounty after reporting the…

  3. could not extract summary

  4. Cache Tax Refusal Receipt GIF Step away from a long Claude Code session for an hour and your next message pays to rewrite the whole context. Same thing when a background agent keeps working while the main conversation sits idle.

  5. https://preview.redd.it/jo19glpqnaqh1.png?width=1466&format=png&auto=webp&s=3dff601ef103cd2562db630a861272e58574f370 It's sad to see that when I wanted to resume my session after weekly and did not want to pay resume tax. I did have to pay…

  6. I'm Borut founder of Rankpilot.dev where we help business to rank on Google & ChatGPT. Last month we processed about 1.00000000 tokens.

  7. Apple put a 512GB desktop on sale on August 25 and my first reaction was to check how much my own Mac Studio had just lost in trade-in value (don’t look, is my advice). The second was to reopen the question I’ve been asked most since the O…

  8. hey hn, i've been building peelaway, a tool for creating and editing images that starts with a curated gallery of visual styles. the idea is pretty simple: pick a style you like, then make something with a photo, a prompt, or both.

  9. A Mac workspace for Claude Code and Codex. For everything between an idea and something you love.

  10. UML viewer A live Quil app that lays out and draws UML from an EDN IR. A policy plus a language-specific parser write the topology; this tool displays it, routes the arrows, colors CRAP, and lets you click.

  11. could not extract summary

  12. could not extract summary

  13. GrassLobster HomeWhat GrassLobster DoesHow It Works Behind the ScenesHow to UseDownloadExamplesAbout Me GrassLobster AI Agentic Generation of Parametric Geometry Workflows…

  14. Cluttered repositories are kind of the norm nowadays. Every AI coding agent adds its own files, such as AGENTS.md for Codex (and other universal agents), a .claude folder for Claude Code, or a .cursor folder for Cursor.

  15. This is the Discussion Hub for showcasing your project built using Claude products. We appreciate all of your submissions as they are a great inspiration to many people on the subreddit.

  16. could not extract summary

  17. My company pays for a Claude Pro account for Claude Code, but I also pay my own subscription for personal projects and chats. On the website I have no trouble because I can use different browsers to keep the accounts separate.

  18. Cybersecurity researchers have hacked into OpenAI with the help of Anthropic’s Claude chatbot, in the latest example of security issues at the company. A team at a US-based startup compromised a number of OpenAI employees’ ChatGPT accounts…

  19. Google Gemini for macOS was recently flagged for not being a very good app citizen on the Mac. Apple’s new macOS 27 Golden Gate seems to make an example out of apps like Gemini that sneakily run in the background after a user quits the sof…

  20. How a solo founder runs a five-continent tender platform on AlloyDB and MCP Davor Jerković Founder, Lucius AI Editor's note: Lucius AI, a tender-intelligence startup covering markets across five continents, runs its entire data platform on…

  21. OutTheBox Open Governance Patterns Three working AI agent governance patterns, free and open source: a decision table, a worked example, and a Microsoft Agent 365 template for each. No sign-up, no trial, no cut-down version, the same conte…

  22. Announcing SpaceXAI's newest speech-to-text model, with unparalleled accuracy and cost effectiveness. Today we're releasing Grok Voice Transcribe 2.0, our latest speech-to-text model.

  23. About Press Copyright Contact us Creators Advertise Developers Terms Privacy Policy & Safety How YouTube works Test new features NFL Sunday Ticket © 2026 Google LLC

  24. I use Fable to handle all design and verification work, and it hands off to Opus to do stuff that is theoretically mostly mindless. Fable produces amazingly detailed design documents for Opus to implement.

  25. I've used Claude Code more or less daily for a year, the last six months seriously, next to Hermes (always-on, Telegram) and Codex (second opinion). Architect by trade, .NET mostly.

  26. The company context engine Every AI tool in your company starts from zero and keeps its own memory. Nerra builds one living context layer from the tools you already use and serves it to Claude, ChatGPT, Cursor and your own agents over MCP.

  27. Anthropic has set up a bio research lab for physical experiments But the company has yet to reveal what It’s working on. Anthropic has established a biology research lab in the San Francisco Bay Area to conduct physical experiments, accord…

  28. Capability Attenuation in Agentic Hierarchies In this post I won’t be discussing simple one-shot request and response behavior. While that’s interesting, it isn’t the problem I want to explore.

  29. I spend a lot of time walking trade show floors, making cold introductions at booths, and trying to sell my services. I don't mind the conversations.

  30. Once you get used to Claude, it’s tempting to use it for everything. I’m interested in where people went the other way.