In today's iteration of AI, I think the best ecosystem would be where the entire intents of the program are spit out more in line vs separated into files throughout a codebase, just because that's a little easier to manage for the AI.

But any advantage like that will be short-lived, I'd imagine.

Getting worse?

Any place where you're trying to learn how to programme. Will teachers need to include prompt writing in the syllabus? Is coursework going to be plagued with generated code?

Also cowboys are bad now, what about when they just put in what the AI spits out?

Getting better?

I guess it can shave some time of writing simple stuff you're sure you can validate.

I think it will be hard to see how things shake out.

Ecosystems which emphasize human readability, i.e. Python might become less valuable if the AI can help directly generate more machine-ready code (i.e. better performance). If the AI is also doing de-bugging, how much "human-readable" code do we need?

On the other hand, if the AI is generating code for less technical developers to gut check and work with, you want very human-readable code.

I really think it will be hard to figure out for sure.