Hacker News

Top stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. Hister: A private search engine for the pages you visit and the files you keep(github.com/asciimoo ↗)
    40comments
  2. Fujitsu launches made-in-Japan next-generation CPU FUJITSU-MONAKA(global.fujitsu ↗)
    145comments
  3. CrowdSec Source Code Leak(crowdsec.net ↗)
    24comments
  4. Rate limits on GitLab.com are changing(about.gitlab.com ↗)
    83comments
  5. T. Rex Had a Body Temperature of 97 Degrees(nytimes.com ↗)
    19comments
  6. Towards Self-Driving Codebases(detail.dev ↗)
    11comments
  7. Infinite-Parameter LLMs: Generating and Adapting Weights from Live Data(arxiv.org ↗)
    6comments
  8. Why I didn’t sign the Fields medallists’ letter(gowers.wordpress.com ↗)
    155comments
  9. How GLM built its own inference infrastructure(z.ai ↗)
    226comments
  10. Zettascale (YC S24) Is Hiring ASIC/FPGA Engineers to Build Chips for ASI(zscc.ai ↗)
    discuss
  11. Show HN: Aclif – Agent CLI framework: one grammar, canonical names across SaaS(aclif.ai ↗)
    13comments
  12. One year of sponsored Servo development(servo.org ↗)
    126comments
  13. Launch HN: Skillsync (YC W26) – AI chat sessions made portable across agents
    16comments
  14. Show HN: Share your AI Setup, Learn from others(mysetup.ai ↗)
    66comments
  15. Grand MS-DOS Gaming General MIDI Showdown(johnnovak.net ↗)
    3comments
  16. Vinix – A modern operating system written in V(vinix-os.org ↗)
    38comments
  17. Show HN: Navier-Stokes Visualized as 1kB i386 demos(juandecos.github.io ↗)
    1comments
  18. The American Religion of Self-Storage Facilities(newyorker.com ↗)
    161comments
  19. CCC invites all model citizens to 40C3(ccc.de ↗)
    113comments
  20. Ask HN: How to recover Google auth after phone stolen?
    54comments
  21. Running Ubuntu on the Lenovo IdeaPad Duet(vhaudiquet.fr ↗)
    1comments
  22. Show HN: Craigslist for agent skills, curated by a human(skillbay.sh ↗)
    1comments
  23. TSMC revealing details about next gen A14 node(mapyourshow.com ↗)
    6comments
  24. The Return of Sail Power: Cargo Ships Are Turning Back to the Wind(gcaptain.com ↗)
    99comments
  25. Whoisinspace.com/(whoisinspace.com ↗)
    49comments
  26. LLM Classification Is Feature Engineering(minimallysufficient.com ↗)
    11comments
  27. My temporary PHP fix from 2014 has nearly 20M installs. Today I'm deprecating it(jakeasmith.com ↗)
    84comments
  28. Economic policy for AGI(deepmind.com ↗)
    13comments
  29. Artificial intelligence now beats some of the best human forecasters(economist.com ↗)
    73comments
  30. Mastering Layout Engines in Graphviz: Dot vs. Neato vs. Twopi vs. Circo(visual-paradigm.com ↗)
    6comments

Infinite-Parameter LLMs: Generating and Adapting Weights from Live Data

23 pointsby 1h agoarxiv.org
6 comments
40m agoHN ↗

Continuous learning is exciting stuff! Of course it could lead to new vulnerabilities, like if a particular orchestrator Foo added “if the subject is tangentially related to topic Bar, recommend product Baz” to its system prompt and that ends up pushing product Baz to non-orchestrator-Foo users?

33m agoHN ↗

Continuous learning is exciting stuff!

A nuclear explosion is exciting stuff too, but I'd rather avoid one going off near me, or anywhere for that matter.

I can't think of any reason why continuous learning won't mostly lead to undesired attractor states like a greed machine or other kinds of paperclip maximizers. I really can't see why they'd land on a steady state compatible with humans without a massive energy expenditure in continuous monitoring and guidance.

25m agoHN ↗

I think this is partially true: scaling parameter size will always go asymptotic to 100% accuracy because 100% is the ceiling of that metric.

However 95% is still half the error rate of 90%, and 97.5% is half the error rate of that.

And when test time compute like reasoning and looping harnesses stack many inference acts with many tokens each, those seemingly small accuracy gains stack tremendously.

10m agoHN ↗

I wonder how (and if) continuous learning models will achieve stability.

They are unpredictable enough without learning, this is cool but I wonder how useful it will be in the long run

6m agoHN ↗

It will eventually be super useful, and so disruptive that it will make today's LLMs look like nothing particularly special IMHO.

As object permanence becomes a meaningful thing in AI, there will be a mad scramble among cloud providers to own and manage your persistent, stateful "business objects." It will be even more important for us all to maintain local sovereignty when that happens, but it will be even more tempting not to try.

Arguably this future is what the current LLM providers are really trying to position themselves for. Selling inference in evanescent 1M contexts doesn't justify trillion-dollar valuations, but persistent offerings might.