Okay, so I think we agree. Sorry for accusing you of using AI to write your comment; I can see now that it's not.
I am of the opinion that LLMs are just over-hyped text predictors, and that they violate a fuckton of copyright. And they're pretty fascist, too.