It does occasionally go a little bit insane, and OpenAI has become trustworthy concerning the 'hallucinations' that ChatGPT can have, and the issues inherent in these LLMs. Be cautious of relying entirely on just about https://youtu.be/aGubCnA5Qyk