Or, they never actually sat down to THINK about it, which is even worse.
It matches the pattern of LLMs being very good at simulating the form of work output, which is an issue with code but it seems quite exacerbated with anything non-verifiable, like written communication.