Comment by thaumasiotes

> It was easy for me to review and understand because I already knew the pattern and it seemed easy for it to get right.

I suggest that you already knowing the pattern actually makes it harder for you to review code that you expect to contain the pattern. You're likely to perceive it as being there whether it is or not. This strikes me as a way of using LLMs that is more dangerous than average.

Relatedly, proofreading your own work is much more error-prone than proofreading someone else's work, precisely because you have a mental model of your own work (created when you produced it) and you're likely to consult the mental model rather than the work.