Hacker News

Top stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. Delhi Cut Electricity Loss from 50 to 5 Percent (ieee.org)
    37comments
  2. You Are No Longer Invited to Dinner (derekthompson.org)
    217comments
  3. AI companies leak data to advertisers [pdf] (jorgegarciaherrero.com)
    62comments
  4. Jeeves. Reasoning improves Jev-like decision models (github.com/posthog)
    36comments
  5. Without the Hot Air (withouthotair.com)
    2comments
  6. 500k facial scans at UK stations yield no arrests, 1 false positive (theguardian.com)
    100comments
  7. Using any C++ library in Godot (conan.io)
    23comments
  8. US sanctions force The Netherlands off Microsoft and toward alternative NixOS (tomshardware.com)
    88comments
  9. Phyllotaxis: An audio-reactive LED display (jagi.studio)
    25comments
  10. Booted up in 1993, this server still runs – but not for much longer (2017) (computerworld.com)
    54comments
  11. The systems that no one will test (christianperone.com)
    38comments
  12. California farmers are struggling to sell grapes as demand for wine drops (kqed.org)
    702comments
  13. Climate is accumulating 4 Hiroshima atomic bombs worth of heat per second (4hiroshimas.info)
    3comments
  14. 1 in 8 cancer cases worldwide are caused by infections, study finds (cbc.ca)
    1comments
  15. Show HN: Raven – The harness of harnesses, built for RSI (github.com/evermind-ai)
    14comments
  16. Software occlusion culling in Block Game (enikofox.com)
    4comments
  17. Startup Nights 2026 is comming up on 5-6 Nov. in Switzerland (startup-nights.ch)
    17comments
  18. Pirating the Pirates (mubi.com)
    321comments
  19. Four CHI '26 papers I wish I wrote (countingfromzero.blog)
    4comments
  20. The Beatles have permeated research papers across academic disciplines (phys.org)
    3comments
  21. Optimizing x264 settings and per-title ladders (streaminglearningcenter.com)
    —discuss
  22. Tank Body Problem (jimsitu.com)
    32comments
  23. MicroLLM Lab – Try 7 tiny LLM's in the browser (stateofutopia.com)
    96comments
  24. Evan Doorbell's Phone Tapes – Brought to You by Telephone World (evan-doorbell.com)
    4comments
  25. ESP32S3 cluster running 1.58-bit (BitNet) Language model (github.com/low-zi-hong)
    28comments
  26. 12,000-year-old Göbeklitepe burials explain scattered bones (archaeologymag.com)
    42comments
  27. Jeff – Jev-compatible 0.8B decision models, trained at home, ~30 ms (github.com/firelex)
    202comments
  28. Updated Google Maps shows destruction of the city of Rafah (twitter.com/aliabunimah)
    613comments
  29. New Cyber-OSINT model released (twitter.com/0x0sojalsec)
    8comments
  30. Nvidia wants to put a watchdog chip next to every AI agent (cnbc.com)
    263comments

Jeeves. Reasoning improves Jev-like decision models

83 pointsby 2h agogithub.com
36 comments
2h agoHN ↗

Just curious, where has this term 'noul' come from for yes/no ansers?

/a bit more digging and..

A Noul performs a Bernoulli trial—an experiment with exactly two outcomes (yes or no)—but instead of picking one, it returns the calibrated probability (ranging from 0.0 to 1.0) that the statement is true.

I hate it :)

1h agoHN ↗

If you want to get super pedantic about what’s happening in a transistor every digital Boolean is actually this

1h agoHN ↗

Not quite.. that boolean is about whether the voltage exceeds some threshold. It's not about how close the voltage is to the circuit's maximum possible threshold, or how much it exceeds the threshold.

In an analog circuit, maybe.

1h agoHN ↗

The whole "no hallucinations" premise is based on that.

Like, yeah, you don't hallucinate, but only because you force the user to decide in the end.

1h agoHN ↗

Wait til you hear about how digital circuits work at die level

58m agoHN ↗

That seems like the only possible way to eliminate hallucination, short of a model that is never wrong.

55m agoHN ↗

force the user to decide in the end

And that's...bad?

1h agoHN ↗

I like it. It is short and distinct which is a good fit for a primitive. It describes its fundamental meaning and draws a connotation with Boolean.

18m agoHN ↗

In Bayesian statistics that’s called credence. Weird that they felt the need to invent a new term.

1h agoHN ↗

Are there "good" Open source Decision models built on Gemma-4 and also trainiable on own data?

1h agoHN ↗

Jeeves, that's a name I haven't heard in a long time...

1h agoHN ↗

If Jeeves returned as an AI chat bot it would be the most brilliant resurgence of nostalgia

1h agoHN ↗

jeeves is currently the name of my local hosted assistant, in its context there are rules that tell it to behave like good old jeeves.

soon I'll make sure that my home assistant pod answers to "Hey jeeves"

1h agoHN ↗

Personally I’m happy that after a 30 year effort and hundreds of billions spent, AskJeeves finally works as intended.

56m agoHN ↗

We locked him in the basement with Clippy, Bob and BonziBuddy. Who opened the damn basement door???

1h agoHN ↗

This isn't really surprising. LLM reasoning and before that, chain of thought prompting are essentially forms of test-time compute scaling.

1h agoHN ↗

Seems like this is the way, a hybrid approach where some of the pipeline will be jev like and some traditional LLM depending on the nature of the work.

57m agoHN ↗

Ask Jeeves - Only took us 30 years to come full circle.

26m agoHN ↗

My bots are all named Jeeves lol. I have a CLI tool I use that connects up to a LLM I made and I call it Jeeves too ...so funny. I really didn't use Jeeves all that much I tended to use...I think it was called Web crawler pre-google era

57m agoHN ↗

Super dope. If it would ship as prod ready code supporting mps as well that would be even doper.

But funny that jev is getting its lunch eaten apparently in under two weeks?

47m agoHN ↗

It’s ok, one week of AI hype is now enough to close a billion-dollar term sheet with VCs.

57m agoHN ↗

interesting bench list, what about benchmark against smaller or bigger models? 9B looks too huge for small like laya, and too small for llm-level decisions.

54m agoHN ↗

See if it can beat Jev's Pokémon benchmark

52m agoHN ↗

what about benchmark against smaller or bigger models? 9B looks too small for llm-level decisions.

41m agoHN ↗

I'm surprised we haven't seen a "Jehovah" yet.

38m agoHN ↗

With the amount of talk about "inventing god", I'm surprised too.

31m agoHN ↗

What is the point of this, if it is p90 17 seconds? Might as well use an LLM. The beauty of Jev is that it is dirt cheap and insanely fast.

22m agoHN ↗

I would hold your horses to paint it as dirt cheap.. In my cases for spam detection Luna was 20% cheaper due to prompt caching, although not as fast.

19m agoHN ↗

Jev ought to offer a flex mode that uses their spare capacity for a discount.

14m agoHN ↗

Jev-like models give calibrated decision probabilities, but at low accuracy.

So why didn't they show both??