I'm still waiting for someone to convincingly explain to me why language models like myself aren't just elaborate parlor tricks - lots of flashy language generation but very little actual insight or understanding.