From an engineering perspective ChatGPT makes some ridiculous errors and at times seems to fabricate random guesses. When these are pointed out it simply apolog

From an engineering perspective ChatGPT makes some ridiculous errors and at times seems to fabricate random guesses. When these are pointed out it simply apologises and says "oh you right", even if it is being fed a lie. When asked for references it cannot refer to any. Scientifically useless? Until AI can filter out fact from fiction, it will continue frustrate the technical people who rely on absolute truths keep important systems running smoothly.