Comment by sixfiveotwo

> Machines are good at applying rules, so when they fail to apply rules correctly, it means they have incorrect, incomplete, or totally absent models.

That's assuming that, somehow, a LLM is a machine. Why would you think that?