It does from time to time go a bit nuts, and OpenAI is straightforward concerning the 'hallucinations' that ChatGPT can have, and the problems inherent in these LLMs. Be cautious of relying entirely on nearly anything https://youtu.be/aGubCnA5Qyk