I don’t intend to be hostile here, but I’m genuinely curious: why does anyone find these kinds of observations about the output of LLMs even remotely interesting or worthy of comment?