It does often go a little bit nuts, and OpenAI has long been truthful concerning the 'hallucinations' that ChatGPT might have, and the problems inherent in these LLMs. Be cautious of relying absolutely on everything https://youtu.be/aGubCnA5Qyk