Re: Reverend Lionel Preacherbot
As i keep banging on, an LLM is not intelligent, it is a guessing engine. It has no understanding of facts, it has been trained on everyone's opinion of the facts and will spew a weighted approximation of an answer to a question that just asked for the 'fact'. And if the training data for a different 'fact' overlapped, you're going to be in a world of trouble if you trust it. Like a doctor, or a lawyer, or an architect.
If your problem has a load of really good solutions (like JavaScript or C) you're in luck, but in the edge cases with bugger all eamaples (good or bad) where you're relying on the documentation (hello microsoft) you are s**t out of luck.
I'm not comfortable with my doctor using an 'AI' to transcribe our discussion, let alone a diagnosis.
But if it does your school homework for you and you pass (without getting caught), great. Knock yourself out.