Hacker News

New stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. Hacker puts 'full redundancy' code-hosting firm out of business (2014)(pcworld.com ↗)
    discuss
  2. OpenAI Discloses Six New Incidents of ‘Concerning’ A.I. Behavior(nytimes.com ↗)
    discuss
  3. Math challenges a 2k-year-old story about Parthenon's optical illusions(phys.org ↗)
    discuss
  4. Could Your Chatbot Be Stealing Ideas from You?(nytimes.com ↗)
    1comments
  5. Weeping whales: Stillborn humpback whale grieving documented(phys.org ↗)
    discuss
  6. I gave my agents a heartbeat(mlsystemsri.com ↗)
    discuss
  7. Will AI Replace Your Doctor? Dr. Zeke Emanuel and AMA President Dr. John Whyte(youtube.com ↗)
    1comments
  8. A Large Database Does Not Mean Large Shared_buffers(keithf4.com ↗)
    discuss
  9. Size-Specialized Memory Allocation(go.dev ↗)
    discuss
  10. What's Scarier Than Agents Taking over Internet? CEO Cartel Trying Take over AI(fractalsofchange.substack.com ↗)
    discuss
  11. AI Cheating Is on the Rise(vals.ai ↗)
    discuss
  12. Unplug America(engageq.notion.site ↗)
    discuss
  13. Suzanne Ciani's Buchla Cookbook(echo.orpheusinstituut.be ↗)
    discuss
  14. OpenAI discloses six new AI safety incidents(axios.com ↗)
    discuss
  15. The Return of Sail Power: Cargo Ships Are Turning Back to the Wind(gcaptain.com ↗)
    discuss
  16. How the Strategic Petroleum Reserve Works(construction-physics.com ↗)
    discuss
  17. What Would Canada's Associate Membership of the EU Look Like?(wsj.com ↗)
    discuss
  18. Cert/CC covers up Google's data retention fraud(medium.com/istokovicsgyorgy79 ↗)
    discuss
  19. A dead man's switch with client-side crypto and an offline decryptor(etergis.com ↗)
    discuss
  20. Show HN: Hacker News, a little more focused(beehivesoftware.xyz ↗)
    discuss
  21. Recursive Meta-Intelligence(twitter.com/profbuehlermit ↗)
    discuss
  22. Unlocking the Brain's Potential, with Gül Dölen [video](youtube.com ↗)
    discuss
  23. Where Yesterday's Decisions Run Today(medium.com/gurvinder372 ↗)
    discuss
  24. AgentLane: Git-native coordination for coding agents(github.com/thetrueshags ↗)
    1comments
  25. Ask HN: How long before we will have a major AI driven cyberincident?
    discuss
  26. Bringing Correctly Rounded Math to Production with LLVM-Libc(devblogs.microsoft.com/cppblog ↗)
    discuss
  27. When the fractional part of a float fixes your shader(crocidb.com ↗)
    discuss
  28. Otis – a minimal AI agent that runs local models out of the box(triangllabs.ai ↗)
    1comments
  29. Researchers have created a mouse with a partly human brain(economist.com ↗)
    discuss
  30. Duotone icon library with 1k icons and 250 flags for React with RTL support(github.com/andrsrxn ↗)
    discuss

Reverse-engineered Jev-like model

71 pointsby 6h agogithub.com
9 comments
2h agoHN ↗

No question OSS is amazing, but this video is a satire at best. It doesn't take much attention to see the results on right vs. left side are significantly different.

Jev is not interesting if it's not "smart", a 1B param model is most definitely not smart.

1h agoHN ↗

I like it! I suspect Jev may have more going on under the hood, but I like the idea of efficient universal transformers

1h agoHN ↗

Any diffusion model is potentially a Jev in disguise: https://github.com/vllm-project/vllm/pull/57250

Runs ~0.2s per decision on my DGX Spark.

  10/10 programming language detection
  9/10 human language detection
  10/12 unit magnitude comparison

All incorrect answers are marked with low-P.

It (DiffusionGemma with the Jev mode) can also solve an ASCII maze.

1m agoHN ↗

Out of curiosity and semi unrelated — why do so many of these projects with customized encoder-decoder setups use earlier Qwen versions like 2.5 and 3 and not the smallest 3.5? Purely the few 100m params, or something else in the latter’s arch or pretraining?