Yeah lol he's not going to be 100% perfect. There are definitely improvements that could be made in the request parameters being sent to the OpenAI API but I think the actual training data being uploaded and referenced as part of the response could be improve I'm just not sure how. Open to ideas!