I’d like to think it would have happened anyway… before it would have happened over a longer timeframe making it less of a “thing”.
I’ve gotten really good results by using a framework that pushes me to make most decisions upfront rather than during implementation. But it’s not perfect, without patterns on what “good code” it will still produce code that is less than stellar and call it a day. I find myself creating skills very often instead of polishing over and over so the next time it’s a breeze.
I’ve been using AI-DLC by AWS but IMO it’s too heavy for anything other than enterprise projects.