51 Can AI Large Language Models (LLMs) form hypotheses, design experiments, interpret results, and update beliefs like real scientists? What they found is sobering: LLM's are basically libtards! (media.greatawakening.win) posted 173 days ago by catsfive 173 days ago by catsfive +51 / -0 16 comments download share 16 comments share download save hide report block hide replies
LLM predict words.
For example, if you type Roses are red it will then write and, followed by violets, then are, then blue. It is trained that way by many attempts.
There is no logic.
Have you actually used Grok? While some LLMs have not logic - several are quite good at it. Grok's reasoning skills are generally superior to most of my coworkers.
I did not say that all people use logic :-) GROK may be good but it is still based on predictive tokens.