It’s way easier to figure that out than check ChatGPT hallucinations. There’s usually someone saying why a response in SO is wrong, either in another response or a comment. You can filter most of the garbage right at that point, without having to put it in your codebase and discover that the hard way. You get none of that information with ChatGPT. The data spat out is not equivalent.