2 ms·Maybe they shouldn't have trusted their LLM to optimize their kernels so much.by measurablefunc 9mo agoMaybe they shouldn't have trusted their LLM to optimize their kernels so much.