LLMs create content not by agreeing but faking, so when you find some inconsistency it will try to cover it up because training data is written by those who really understands it so it’s probably rare that they realize and change their minds while chatting. learning by llm makes your cyristalized brain a slippery cheeze. I lost my friends to that, I can’t even discuss anything because they start to say idk this or that and they think this is freaking normal! btw I learned the word “cyristal” from the archwiki/Users and Groups page where they went off topic and enlighen us so read the wiki ffs