3 ms·
Perhaps, just perhaps, LLMs are just statistical models that literally can't create novel things, therefore any structure LLMs write was learnt from human writi
by raincole 7mo ago
Perhaps, just perhaps, LLMs are just statistical models that literally can't create novel things, therefore any structure LLMs write was learnt from human writing?
But who knows!
- bakugo 7mo agoWhat kind of human writing has "it's not X—it's Y" in every single paragraph? The answer is none. LLMs haven't accurately modeled human writing for years, current models have been smacked on the head with the coding RLHF bat so much, they all write distinctly inhuman text.
- dpark 7mo agoThe thing is, people are screaming “AI” when they see a single “it's not X—it's Y" pattern in a post, despite this being a fairly common construct. People are nitpicking every tiny thing in their search for proof of AI. It’s not useful and ends up dominating the conversation. AI panic is degrading the value of forums at least as much as actual AI at this point.