It's great that thr article really demonstrates what are the presumed productivity gains and how big they are.

Oh wait, it doesn't. How does one know it's a gain without comparing it to a baseline?