the (low) dimensionality of that thumbs up signal (and the nature of the thumbs up corpus) is the key key key key thing to what LLMs can't do and what they are super weird about and why they're Like That
@colin-fraser.netit's not exactly trying to tell *you* a story. It's trying to tell a story to The Raters, who are people (and other software systems) employed by the AI Lab who give the story a thumbs up when they are pleased.
