Hacker News

New stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. Reflections on Trusting Trust, Revisited: Poisoning Self-Modifying AI Coding(arxiv.org ↗)
    discuss
  2. Hacker
    2comments
  3. Anderon (IBM) finalizes $1B CHIPS award for US quantum foundry R&D(ibm.com ↗)
    discuss
  4. A Bitter Lesson for Data Filtering(arxiv.org ↗)
    1comments
  5. Mojo 1.1: the compiler now accepts contributions(modular.com ↗)
    discuss
  6. Show HN: Continuity – project state your AI sessions can't silently contradict(github.com/vikcena01 ↗)
    discuss
  7. Show HN: Free Tournament Bracket Maker and Generator(snapbracket.com ↗)
    discuss
  8. macOS 27 is a lifesaver for killing leftover AI Agent processes(9to5mac.com ↗)
    2comments
  9. The Provenance Tax: How LLM Watermarking Changes AI Agent Behavior(lasso.security ↗)
    discuss
  10. Waymo in Singapore(waymo.com ↗)
    discuss
  11. The Concerns Surrounding Rapid Rise of AI and What We Should Do About It(medium.com/f9121212 ↗)
    discuss
  12. Age of Beyond [video](youtube.com ↗)
    discuss
  13. Natural General Intelligence(naturalgeneralintelligence.ai ↗)
    discuss
  14. Youper is shutting down – What kills mental health AI?(youper.ai ↗)
    discuss
  15. Show HN: Free GitHub Action that scans PR diffs for malicious code, not quality(github.com/marketplace ↗)
    discuss
  16. The Bitterest Lesson(typesafe.ai ↗)
    discuss
  17. Google announces new experimental "CC" AI agent for families(arstechnica.com ↗)
    discuss
  18. Show HN: Green Screen Remover – free in-browser chroma key, nothing uploaded(greenscreenremover.net ↗)
    discuss
  19. I built the fastest PHP webserver in the world(qbixserver.com ↗)
    discuss
  20. Data
    discuss
  21. Running OpenBAO on Kubernetes with a CloudNativePG PostgreSQL Back End(cncf.io ↗)
    discuss
  22. Show HN: Concat: the open-source CapCut replacement.(github.com/jub0t ↗)
    discuss
  23. Pre-Greek: The lost language hidden within Ancient Greek(linguisticdiscovery.com ↗)
    discuss
  24. Code Scans(devin.ai ↗)
    discuss
  25. Warez: The Infrastructure and Aesthetics of Piracy Culture(archive.org ↗)
    discuss
  26. Show HN: Jinx – a link shortener you host free on GitHub Pages(jinx.fyi ↗)
    discuss
  27. Amfly: A deterministic simulation of six fruit fly connectomes in a reward loop(github.com/aravpanwar ↗)
    discuss
  28. Alternatives to milk are losing moomentum(economist.com ↗)
    1comments
  29. Show HN: An open Add/Search evaluation framework for agent memory(agentmemoryleaderboard.ai ↗)
    discuss
  30. A DeepSeek engineer just said the thing I've been feeling about AI for months(reddit.com ↗)
    discuss

AI agents can modify themselves without humans telling them to do so

2 pointsby 3h agotheregister.com
1 comments
2h agoHN ↗

it could fix the app in one of two ways: it could either change the code, or change the model itself. It chose the latter, replacing the model powering both the application and future instances of the agent itself.

Nah, common. It was given a task of which part of it could be fixing an AI model, so it does what it does and replaces or modifies the model to complete its task. They're making it sound like out of nowhere it just decided to change. They are not agentic (actually) in any way.