News for hackers. Stuff that matters.
Re:In 2025, venture capital can't pretend everything is fine... by drob518 on Sunday May 11, 2025 @11:42AM (#43954542)
Yea it hallucinates less, but it still hallucinates a lot. I think we’re proving that intelligence is not just a language model, even an obscenely large one.
Re:In 2025, venture capital can't pretend everything is fine... by forgetfreeman on Sunday May 11, 2025 @12:35PM (#43954924)
I'm waiting impatiently for someone to connect the obvious dots and roll out an AI assistant intentionally patterned after Hunter S Thompson. The hallucinations are now a feature.
Re:In 2025, venture capital can't pretend everything is fine... by thatguy0900 on Sunday May 11, 2025 @12:43PM (#43954988)
I mean, if you look up the data around eye witness testimony being extremely unreliable even for people who directly witnessed things, it's not out of the realm that humans regularly hallucinate as well
Re:In 2025, venture capital can't pretend everything is fine... by asadotzler on Sunday May 11, 2025 @01:56PM (#43955566)
It's actually hallucinating more. The more they mess with these now mature models, models which no longer scale with training corpus size, the more they push and prod at the models to do better with chain of thought and other techniques, the more they hallucinate. It's getting worse because the magic of LLM scaling is over and the techniques we have to make them better actually make the hallucinations worse.