Harvard finds no productivity improvement with AI. https://fion.ac/jellyfish.pdf We've been here before. It made an unmanageable mess then, too. https://en.wikipedia.org/wiki/Computer-aided_software_engineering Some languages and paradigms require more boilerplate than others, and many business problems aren't novel, and there are AI benefits in those situations. I'm betting that places writing F# or Haskell or Clojure, scientists or academics, or realtime systems might see a much lower benefit.