3 ms·
"Hallucinate" is an interesting way to position it: It could just as easily be positioned as "too ignorant to know it's wrong" or "lying maliciously". Indeed,
by xkcd-sucks 2y ago
"Hallucinate" is an interesting way to position it: It could just as easily be positioned as "too ignorant to know it's wrong" or "lying maliciously".
Indeed, the subjects on which it "hallucinates" are often mundane topics which in humans we would attribute to ignorance, i.e. code that doesn't work, facts that are wrong, etc. Not like "laser beams from jesus are controlling the president's thoughts" as a very contrived example of something which in humans we'd attribute to hallucination.
idk, I'd rather speculatively invest in "a troubled genius" than "a stupid liar" so there's that
- shombaboor 2y agothey've got some great marketing to get away without using: bug, defect, error, malfunction
- shombaboor 2y agoor limitation
- euroderf 2y agoIt might be more like reification ? The system finds a satisfying solution and sort of wills it into existence - by verbalizing it, makes it so ?
- zh3 2y ago>It could just as easily be positioned as "too ignorant to know it's wrong" ... GPT-5 is widely predicted to have a Dunning-Kruger level of expertise.
- phainopepla2 2y ago> lying maliciously Malice implies intent, which is even more misleading that hallucination, imo
- taneq 2y agoIt's just 'prediction error' in a feedback loop, imo. I'm sure like any other biological human with mitochondria and stuff, you've occasionally said (or started to say) something and then you (ie. the actively cross-checking self-analyzing enigma that is 'you') thinks 'hang on, no that doesn't make sense' and you self-correct. LLMs are 100% feedforward, there's just one big autoregression going on. No strange loop shenanigans. Honestly I'm really interested to see where LLM-based diffusion models end up. (To be fair, probably mostly because I don't understand them yet so they could still be spooky. :D )