Even since ChatGPT 3.5 AI coding has been astoundingly good. Like you, I'm totally baffled by developers who say it's not. And I'm not just unskilled and unable to judge. I have 35yrs experience! I've been shocked and impressed from day one, how good AI is. I've even seen Github Copilot do thing that almost resemble mind-reading insofar as predicting what I'm about to type next. It predicts some things it should NOT be able to predict unless it's seeing into the future or into parallel universes or something! And I'm only half joking when I speculate that!