Does AI raise the floor or the ceiling in software development? Probably both. But which one does it affect more?
The ceiling argument goes like this. Your best developers already have domain expertise. They already have product expertise. They know the customers, the codebase, the history of why things are the way they are. Give those people AI and they become that much more dangerous. More productive. Faster at the parts that used to slow them down without losing any of the judgment that made them good in the first place.
The floor argument is different. Some companies were outsourcing development to people with no domain experience, no product experience. Those teams were stuck. Moving slowly. Shipping buggy code. Shipping crappy code. And now those same companies have access to an amazing development partner at a fraction of what it would have cost to hire someone with real expertise.
I keep going back and forth on which one matters more, and I don't think there's a clean answer. It depends on where you sit. If you already have strong people, AI makes them stronger. If you were relying on weak development, AI closes the gap faster than anything that came before it.