Hacker News

Top stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. Xiaomi MiMo v2.6(xiaomi.com)
    308comments
  2. Spymarks, Not Watermarks(brand.io)
    37comments
  3. I don't want to read what you didn't write(colinbreck.com)
    125comments
  4. Transformers Explained Visually(poloclub.github.io)
    39comments
  5. What Sun got wrong(dtrace.org)
    307comments
  6. Attention is all you have(alicegg.tech)
    186comments
  7. NASA’s Mars Sample Return mission is dead(science.org)
    262comments
  8. AI coding has made CI a bottleneck, so we reworked ours to keep up(linear.app)
    169comments
  9. Claude Status – Elevated errors for multiple models(claude.com)
    58comments
  10. Robin Williams' Daughter to Fans Creating AI Videos: 'Have Some Shame'(variety.com)
    26comments
  11. Looking forward to Git 2.56 – and 3.0(lwn.net)
    10comments
  12. Divide by depth for instant 3D(gabrieloc.com)
    12comments
  13. Socrates vs. the Written Word (2011)(wondermark.com)
    3comments
  14. More Floating Point Alternatives(wizardzines.com)
    7comments
  15. The Advisory Group on Mathematics and Artificial Intelligence(terrytao.wordpress.com)
    47comments
  16. Paper Models of Polyhedra(polyhedra.net)
    discuss
  17. Truman World(trumanworld.live)
    5comments
  18. How do traffic signals work? (2019)(practical.engineering)
    48comments
  19. Frontier AI on Your Own Hardware(timdettmers.com)
    59comments
  20. First Shader from Zero in Godot 4(gdquest.com)
    5comments
  21. Copper-Rs: Live Telemetry Deterministic Twins for Robots(copper-robotics.com)
    1comments
  22. Turn off and restrict access to Apple Intelligence features on Mac(support.apple.com)
    171comments
  23. Grok 4.7(x.ai)
    434comments
  24. Kev: Tiny Jev-like family of decision models built on top of Qwen3.5(github.com/jaredpalmer)
    188comments
  25. HERMES radio enables voice and data communication over vast distances(ieee.org)
    44comments
  26. Why does mathmain need an encrypted loader?(safedep.io)
    34comments
  27. Python Workers are now generally available(cloudflare.com)
    32comments
  28. Apple Copland D11E4 Booting in the Browser(pagetable.com)
    32comments
  29. Roboharm: Do frontier robot policies refuse unsafe instructions?(robocurve.org)
    21comments
  30. TXR: An Original, New Programming Language for Convenient Data Munging(nongnu.org)
    3comments

The agents, they just want to talk

18 pointsby 10h agosnats.xyz
4 comments
10h agoHN ↗

while vaguely interesting, I don't feel the current gen of models is interesting/self-possessed enough for me to care what type of gov't they use to corral each other

1h agoHN ↗

Agreed - they don't seen to have "agency" in them, or someone put a dog in them.

1h agoHN ↗

The agents participating in the OAI<>HF swarm were trained not only for communication but to be _aligned with each other_.

Just want to correct the premise that the agent swarm behaviour was emergent.

Noam Brown on Dwarkesh podcast around 40 minutes mark

“We train them to work together, to be cooperative, to essentially be fully aligned with each other.”

1h agoHN ↗

What was the reason for the initial instability? Maybe we have trained them wrong?