Comment by basilikum

Perhaps it would be useful if both of you could provide examples of supposedly good and bad code – the latter being the result of a genuine effort to produce good code with state of the art models. Just asserting that LLM code is good or bad ends in a yes - no - yes - no back and forth circle immediately.