tbh I would find it annoying to have to go audit someone else (i.e. an LLM's) code... Also, maybe the humbling question is, maybe we humans aren't so exception

tbh I would find it annoying to have to go audit someone else (i.e. an LLM's) code... Also, maybe the humbling question is, maybe we humans aren't so exceptional if 90% of the sum of human knowledge can be predicted by next-word-prediction