AI code is much easier to read than AI text (or book). It's kind of like what people think of the AI generated book cover. That's how I feel about the generic AI writing.
Fun fact, this isn't new. There's an entire discipline that has been doing this for 50 years: machine learning.
Literally the hard part that people deal with in ML is how do you specify the goal of a machine that's just going to blindly make it happen (the other half is how do you optimize that function but that tends to be considerably easier). Like, if I want a good image recognition algorithm what function should I use to compute how good my current approach is. Particularly when I don't have annotations for every output.
We're going to need to extend the methods that we developed in ML for many other fields over to software engineering.