at this point, how can anyone trust chatGPT when we know it hallucinates things in its responses? It returns things that look like code or answers or whatever,

at this point, how can anyone trust chatGPT when we know it hallucinates things in its responses? It returns things that look like code or answers or whatever, and that's all it's trained to do. The concept of "correctness" only exists insofar as its trained, and you can't train on something that doesn't exist yet, generative AI is not creative in that sense I have no idea why people are letting chatGPT do anything without pouring over everything it says first, and at that point why bother.