Yet another Hacker News

Jeeves. Reasoning improves Jev-like decision models

73 points nicowaltz 1 hour ago 29 comments

alienbaby 1 hour ago | parent

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 :)

keepitwiel 1 hour ago | parent

Bernoulli

user3939382 1 hour ago | parent

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

kevindamm 1 hour ago | parent

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.

k__ 43 minutes ago | parent

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.

rusk 31 minutes ago | parent

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

doginasuit 30 minutes ago | parent

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

kjs3 26 minutes ago | parent

force the user to decide in the end

And that's...bad?

doginasuit 33 minutes ago | parent

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.

zerop 1 hour ago | parent

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

s2l 9 minutes ago | parent

[delayed]

raverbashing 1 hour ago | parent

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

fishfasell 59 minutes ago | parent

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

grokkedit 51 minutes ago | parent

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"

gizajob 45 minutes ago | parent

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

kjs3 28 minutes ago | parent

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

woadwarrior01 51 minutes ago | parent

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

swader999 47 minutes ago | parent

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.

phplovesong 39 minutes ago | parent

So "askjeeves" has been resurrected?

thm 29 minutes ago | parent

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

tmnstr85 20 minutes ago | parent

this was the comment i came here for

RamblingCTO 29 minutes ago | parent

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?

pavlov 18 minutes ago | parent

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

mxkuzn 29 minutes ago | parent

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.

captainbland 26 minutes ago | parent

See if it can beat Jev's Pokémon benchmark

Naitik88 23 minutes ago | parent

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

AnodicElegy 12 minutes ago | parent

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

jadar 10 minutes ago | parent

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