This paper created some neat theoretical mathematical models but I can't see anything that indicates that they have anything to do with reality.

I thought they'd come up with a neat mechanism that looked at actual wages in real companies where automation was introduced, but no. It's all spherical cows.