And I move on. When I've given it chances, it's been so laughably off base so often that I just can't take anyone trusting its output. Not saying there aren't good use cases, especially when matters of specific correctness are not involved (I rarely have a use for anything like LLMs, unless I'm trying to find something that I cannot through normal means). But, no doubt there is an artificial push towards treating such tech as as a mouth of truth, rather than a language-based probability filter.

Already, I meet people that are replacing their need to think with it. When neutral implants become a big thing, they will go from NPCs to unscary real world versions of Agent Smith.