The surreal thing about working with LLMs is that sometimes they tell me things I don't know, but sound reasonable! But I don't know how to verify the information 😔. Today Claude explained to me how alignment works (in the memory sense) and IDK if I should trust it