It does occasionally go a little bit crazy, and OpenAI has become truthful with regards to the 'hallucinations' that ChatGPT can have, and the issues inherent in these LLMs. Be wary of relying wholly on everything https://matteovpod571484.blogchaat.com/29700294/top-guidelines-of-aitips