Peter Krupa on Nostr: this is a good thread. again, the thing to remember about LLMs is they don't "know" ...

this is a good thread. again, the thing to remember about LLMs is they don't "know" anything. they are not going out and retrieving information. they are generating plausible-sounding language. and they are very good at that! so good that they've convinced a bunch of humans that they can think! even though clearly, they cannot. https://old.reddit.com/r/explainlikeimfive/comments/1dsdd3o/eli5_why_cant_llms_like_chatgpt_calculate_a/