It does sometimes go a little bit crazy, and OpenAI has been trustworthy in regards to the 'hallucinations' that ChatGPT might have, and the issues inherent in these LLMs. Be wary of relying wholly on just about https://youtu.be/aGubCnA5Qyk