This was an interesting take on how AI will impact programmers that resonated with me. It was written by Chelsea Troy who is a staff engineer at Mozilla and CS teacher at the University of Chicago.
The post is definitely worth a read. In a nutshell, programming back in the old days used to be building code out of basic logical and data building blocks. Indeed, programmers are still taught these things in academic environments (and sometimes tested on them in job interviews). However, the job doesn’t involve using the building blocks much. It’s mostly about bolting together higher level libraries and abstractions. Much of this activity is about reading and understanding other people’s code and then fixing it when things don’t work as anticipated.
In terms of delivering understanding, AI is not much better than searching StackOverflow. If you have a good basic knowledge, you can draw between the lines and make the answers work in your context. AI (and StackOverflow) helps, but you still need to understand what is going on so you can add value to the code by making it better (or work).
This is almost exactly my experience with it. AI is pretty reasonable to develop basic scaffolding but it requires a lot of specific experience and understanding to make that code work. I’m definitely going to check out some of Chelsea’s videos on O’Reilly.