Hacker News new | ask | show | jobs
by rimliu 31 days ago
LLMs are very bad at reasoning. They don't reason, they just backport their "reasoning" to fit what they did.