I asked ChatGPT one of my company's interview questions (which involves designing a bloom filter). It got it very subtly wrong; when asked to fix things, it ma

I asked ChatGPT one of my company's interview questions (which involves designing a bloom filter). It got it very subtly wrong; when asked to fix things, it made it worse. I would definitely not use it to interview for jobs yet, but I do think it would be good for training interviewers to detect BS and suggest "hints" that lead things in the right direction. Practice interview without wasting any human's time! Meanwhile, humans have not really struggled with this problem. We tell you what a bloom filter is, you just have to write the code to spec. (This is different than some interviews where you need to know the answer to the trivia question; know what a bloom filter is and how to implement one, or you fail. ChatGPT was not capable of either.) Overall my take on the current state of AI-produced computer programs is that the programmer is forced to take on the role of code reviewer and detail checker, which I think is a more difficult problem than just typing in code in a green-field environment. The thing that's difficult about reviewing code is that things can diverge away from your mental model so quickly, that you don't know which details you should be checking up on. When you're writing it from scratch, you don't lead yourself astray in the same way. In a few years this could be completely different, of course.