Requiem for the hopes of a pre-AI world — LessWrong