1. could not extract summary

  2. What do LLMs think about vision models?

  3. Last week, a lot of people reported that Opus 5.5 was nerfed [1], right before the release of Fable 5.2. Unfortunately, benchmarks like livenerf [2] didn’t detect it.

  4. I run about 15 scheduled tasks in Claude Desktop (Code tab). Each task lets me pick a model, a schedule and a permission mode.

  5. CUGA Policy Playground Two live CUGA agents, with and without policies There are nearly 200 agent harnesses to choose from. We gave three of them the same open-weight model and 90 AppWorld tasks, and changing the harness moved task complet…

  6. I already have $20 claude pro. what should i buy next ?

    +74 more
  7. 31st October 2026 - Link Blog How We Took Over ChatGPT Accounts (via) Parameter Labs describe an exploit they found against ChatGPT's implementation of the artifacts pattern (OpenAI used to call this "Canvas" but it looks like the latest r…

  8. I noticed that, before carrying out a task, coding agents run find, rg --files, or glob to locate files and the classes and modules they need. Then they search file contents with rg or grep.

  9. could not extract summary

  10. Linux Foundation Welcomes Agent Name Service to Bring Verifiable Identity and Discovery to AI Agents The Linux Foundation | 10 November 2026 Open standard builds on DNS and the certificates websites already use to verify AI agent identitie…

  11. Legora can’t read a single trace of what its agent does for a lawyer. The data belongs to the law firm’s clients.

  12. I use Claude Code in VS Code with Bypass permissions on. Still every time Claude wants to use e.g.

  13. I'm building Grip, a free & open source screen recorder for Mac. I wanted a loupe to blow up one spot without zooming the whole screen, like a regular zoom does.

  14. Agentic cheminformatics with Google Antigravity Designing new medicines requires curating complex data from charts, figures and literature, training predictive models, and proposing candidate molecules to synthesize — a workflow that typic…

  15. Really curious if there is actually a market for companies trying to sell a memory layer for AI agents. Look at Claude code's "automemory" for example, they don't use a vector database or anything complex.

  16. Boris Cherny loves to code. So much, in fact, that he once said coding is the reason most people choose to become engineers in the first place.

  17. could not extract summary

  18. What I built: Foreman, a Claude Code skill that runs a whole plan for you. Free to try, MIT, nothing to sign up for beyond the Claude plan you already have: github.com/MarioMagdy/foreman What it does: I agree a plan with Opus once.

  19. We read every open job posting that names an AI model. We kept only the roles where the job is to build software on that model.

  20. could not extract summary

  21. Been a Claude user since January. Love it.

  22. could not extract summary

  23. I built a 3D Gaussian splat of an ibuprofen liquid softgel from 189 turntable photos, using Claude Code (Opus 5.5 on medium). It's free and open source (MIT): my github repo has the raw photos, every script, the exact settings, and an offl…

  24. claudette Run Claude Code with Mistral models. claudette is a drop-in wrapper around claude.

  25. could not extract summary

  26. I’ve been trying to use voice for longer prompts because typing out context for a bug takes forever. In theory it’s perfect, I talk for a minute and Claude gets way more detail than I’d ever type.

  27. 8 items

    Qwen3.8-27B-Coder390-EfficientThink-Opus5.5-GPT6Astra-Grok4.7-DSV4Pro-K3-SFT-RLOO-MTP-DFlash2 is a post-trained model with 27 billion parameters, known for reducing truncations and maintaining performance across various benchmarks. Community feedback highlights issues like unexpected behavior in models (e.g., Grok 4.7's performance decline) and billing disputes related to usage limits and charges.

  28. Amid a deepening bear market for AI/tech (worthy of dark humor) with Anthropic bonds receiving a “junk” rating and OpenAI fundraise headlines, we continue to see very surprising capability finds in the community, such as today. The best wa…