[pavo-etc]

I would argue that is a net positive, it is valuable to know if a language model was involved enough to be committing itself.

[pinkmuffinere]

+1, it definitely changes the way I interact, and the amount of suspicion I would have for the code.