Doubt this is news to many of my followers, but a quick primer on why not to use #ChatGPT or other chat-only #LLM as fact finding answer machines.

They don't know anything.

Every next word is just a function of the previous words.

It sounds likely and probable because that is what it is designed to do.

It may be factually correct, but only coincidentally in so far as the true answer may also sound likely, so be emitted.

Each word is decoupled from reality, only attached by language use.