News for hackers. Stuff that matters.
Re:Jev: New frontier model 40-400x cheaper and 20-200x faster by 8note on Tuesday September 15, 2026 @08:07PM (#49720537)
if it puts a high confidence value on a wrong answer, thats still hallucinating, no?

llm hallucinations are high probability tokens that are incorrect vs the real world

Re:Jev: New frontier model 40-400x cheaper and 20-200x faster by dozerly on Tuesday September 15, 2026 @10:00PM (#49721284)
Yes, there is no magic sauce here that makes stochastic output binary if that’s what people are looking for.
Re:Jev: New frontier model 40-400x cheaper and 20-200x faster by resonious on Tuesday September 15, 2026 @10:07PM (#49721326)
Right and so maybe we should stop saying "can't hallucinate" when it can by definition.
Re:Jev: New frontier model 40-400x cheaper and 20-200x faster by rpunkfu on Tuesday September 15, 2026 @10:15PM (#49721375)
It’s not what people are looking for, but what they wrongly claim.
Re:Jev: New frontier model 40-400x cheaper and 20-200x faster by jubilanti on Tuesday September 15, 2026 @11:51PM (#49721897)
Correct, they have not made a universal all-knowing omniscient oracle, which is what would be required for "can't hallucinate".
Re:Jev: New frontier model 40-400x cheaper and 20-200x faster by spencerflem on Wednesday September 16, 2026 @01:05AM (#49722311)
Not to be tooo pedantic, but a bot that assigned 0 confidence to everything wouldn’t hallucinate.

A calculator either gets the right answer or doesn’t answer.

It wouldn’t have to be all knowing as long as it knew perfectly what it doesn’t know

Re:Jev: New frontier model 40-400x cheaper and 20-200x faster by baq on Wednesday September 16, 2026 @02:44AM (#49722892)
A quantum calculator answers in distributions.
Re:Jev: New frontier model 40-400x cheaper and 20-200x faster by stpedgwdgfhgdd on Wednesday September 16, 2026 @06:59AM (#49724702)
In one universe that is true, in another one not.
Re:Jev: New frontier model 40-400x cheaper and 20-200x faster by baq on Wednesday September 16, 2026 @05:28PM (#49733267)
Copenhagen is a beautiful city.
Re:Jev: New frontier model 40-400x cheaper and 20-200x faster by eru on Wednesday September 16, 2026 @01:07AM (#49722324)
That seems like a weird standard.

I would be happy enough with: only produces what it can verify with sources.

If you eg try to remember a court case (ie produce the reference via LLM token generation only), it's easy enough to check with your data whether it really exists. Similar for following links and other references.

If your data or sources are wrong, obviously your report about them will be wrong. But I wouldn't call that a hallucination.

Re:Jev: New frontier model 40-400x cheaper and 20-200x faster by baq on Wednesday September 16, 2026 @02:43AM (#49722888)
There isn’t a single human in this world and hasn’t ever been that meets your happy-enough standard. Make of it what you will.
Re:Jev: New frontier model 40-400x cheaper and 20-200x faster by hdjrudni on Wednesday September 16, 2026 @03:40AM (#49723211)
Why is that at all relevant?

Humans are known to hallucinate a lot. Ask 10 different witnesses at a crime scene what they saw and they'll all report different things.

A good, non-hallucinating LLM would only report things for which it has evidence. It would consult the facts every single time.

It's a pain in the butt for humans to fact-check everything but LLMs can quickly look up all kinds of stuff. That's what makes them useful.

Re:Jev: New frontier model 40-400x cheaper and 20-200x faster by eru on Wednesday September 16, 2026 @05:45AM (#49724090)
Yes, and for the LLM you can do it in multiple passes.

So you can bolt the fact-check / source-check pass onto whatever other system you have, without having to redesign the underlying system.

Re:Jev: New frontier model 40-400x cheaper and 20-200x faster by eru on Wednesday September 16, 2026 @03:08AM (#49723024)
It's not a binary thing. You can get closer or further away from that standard.

And humans also behave differently in different contexts. A conversation at the pub has more such hallucinations than a formal deposit in court. For the latter, a good lawyer will look at her shoes, when you ask him what colour her laces are.

Re:Jev: New frontier model 40-400x cheaper and 20-200x faster by tedbradley on Saturday September 19, 2026 @07:27AM (#49765619)
That's not true that incorrect sources means incorrect report. Often, LLMs have some sense of what is true, and due to that, they hallucinate plausible sources that appear to back that knowledge up.
Re:Jev: New frontier model 40-400x cheaper and 20-200x faster by adastra22 on Wednesday September 16, 2026 @04:19AM (#49723489)
No, I don't believe so. Hallucinations are not "high probability" in a real sense. They are an artifact of the random walk the inference algorithm takes, which causes it to latch on to and chase attractors in the noise. This random walk behavior is necessary for chat interfaces to be useful, but are less critical to typed output predictors. I'm guessing they found some optimization that is possible if you give up caring about chat.
Re:Jev: New frontier model 40-400x cheaper and 20-200x faster by elil17 on Wednesday September 16, 2026 @06:00AM (#49724192)
What we would want to see if a confidence value that is in line with the actual correctness. If the value is 0.9 for 1000 different answers, then approximately 900 of those answers should be correct.