Hacker News

New stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. RuneBench: Agent Benchmark on RuneScape Gameplay Tasks (maxbittker.github.io)
    1 pointsby 1 minute ago
    —discuss
  2. Towards safety cases for frontier AI training (openai.com)
    1 pointsby 2 minutes ago
    —discuss
  3. As AI models go rogue, do you still trust OpenAI and Anthropic to stop them? (theguardian.com)
    1 pointsby 3 minutes ago
    —discuss
  4. Trump asked Xi Jinping if China would like to buy US weapons (theguardian.com)
    2 pointsby 3 minutes ago
    —discuss
  5. Mathematical Origami (mathigon.org)
    1 pointsby 4 minutes ago
    —discuss
  6. The Quantum Atlas (quantum physics for non-experts) (umd.edu)
    1 pointsby 4 minutes ago
    —discuss
  7. Britain overtaxes well-paid workers (economist.com)
    2 pointsby 5 minutes ago
    —discuss
  8. Education Department formally rescinds Title IX protections for LGBTQ students (thehill.com)
    1 pointsby 9 minutes ago
    —discuss
  9. Automating eval design and hillclimbing with Claude (claude.dev)
    2 pointsby 10 minutes ago
    —discuss
  10. Humans Are Reading Copilot Prompts – and They're Horrified (404media.co)
    1 pointsby 11 minutes ago
    —discuss
  11. An Open-Source Terminal UI for Azure DevOps
    1 pointsby 19 minutes ago
    —discuss
  12. What's Behind the AI Panic (bombthrower.com)
    2 pointsby 21 minutes ago
    1comments
  13. The dataset where fine-tuning turned actively harmful – and how we now catch it (cacheverifier.com)
    1 pointsby 22 minutes ago
    —discuss
  14. Hackers Jailbreak AI Chatbots (twitter.com/hamzaonchain)
    1 pointsby 23 minutes ago
    —discuss
  15. CVE-2026-85706 (CVSS 10.0) (twitter.com/eqstlab)
    1 pointsby 24 minutes ago
    1comments
  16. Suicides in 5 Weeks. What's Going on at Cyber Command at Fort Meade? (thebanner.com)
    2 pointsby 26 minutes ago
    —discuss
  17. What 1,350 Runs Taught Us About Prompt Guardrails (humanbound.ai)
    1 pointsby 31 minutes ago
    —discuss
  18. CSV Lint update (Notepad++) text data syntax highlighting and quality validation (github.com/bdr76)
    1 pointsby 31 minutes ago
    1comments
  19. Uncensored and Offensive Security AI Models Benchmark (github.com/joasasantos)
    1 pointsby 34 minutes ago
    —discuss
  20. The Glimmer in Our Eye (With Grant Sanderson) Podcast (youtube.com)
    1 pointsby 34 minutes ago
    —discuss
  21. Cue – Your Personal Agents, with Their Own Identity (cue.im)
    1 pointsby 34 minutes ago
    1comments
  22. Show HN: I wired a fruit fly connectome to play tic-tac-toe (github.com/50rishu)
    1 pointsby 34 minutes ago
    1comments
  23. Show HN: Open-source skills for end-to-end video ad campaigns (github.com/supercmohq)
    2 pointsby 38 minutes ago
    —discuss
  24. Aw1-breaker: Sub-millisecond execution circuit breaker for AI agents (github.com/mvikmanners)
    1 pointsby 41 minutes ago
    —discuss
  25. Show HN: PocketUni – A mobile app for discovering universities (play.google.com)
    1 pointsby 41 minutes ago
    —discuss
  26. OpenAI halts frontier training over rogue models (stork.ai)
    2 pointsby 46 minutes ago
    1comments
  27. Terminal Velocity for Claude Code. Tickets, Loops, Focus. All in One (github.com/quazardous)
    1 pointsby 47 minutes ago
    —discuss
  28. A Rich Neighbor Named Xerox (folklore.org)
    1 pointsby 49 minutes ago
    —discuss
  29. Decryp Code U534 P1030680
    1 pointsby 51 minutes ago
    2comments
  30. A poem for the state of development in 2026
    1 pointsby 57 minutes ago
    —discuss

Automating eval design and hillclimbing with Claude

2 pointsby 12m agoclaude.dev
Open original View on HN
0 comments
A quiet thread, for now.Start the conversation on HN ↗