How hard can it be to define a policy on the use (or non-use) of generative AI in a project like KDE? Judging by this week's uproar, pretty hard.