Haha, that would be an interesting LLM fine tuning experiment.