It was written by an LLM, so... yeah.

Except this isnt using heavily quantised versions of the model thus reducing quality.