Ask HN: Will AI force CS to focus on what to build instead of how to build it?

As AI becomes really good at the implementation part, it seems that what's left for humans to focus on is what to build.

Do you think this will be the case and CS will change accordingly in its focus?

AI is really good at the implementation of small, stand-alone pieces of software. Within this very limited scope your concept holds: Humans may focus on what to build instead of how.

Anything more complicated than that and AI is easily overwhelmed with even the implementation part, trapping itself in endless cycles of contradictory code changes.

So, no, I don't think this is what is going to happen.