So in other words, the theory is that unemployement due to increased automation is still not a systemic problem - it will just take a number of years (perhaps decades) for the economy to readjust. The somewhat obvious corollary to this theory would be that in the meantime, you want to make sure you're employed in the field which is causing the disruption. This is where the fruits of the increased automation will be seen first.

I can accept this theory. Personally, I believe we'll eventually get to a point where the economy will have to do a dramatic shift in the direction of greater taxation of the richest and more leisure time for most people ("most work is automated"), but we might not be close to that point yet. As of right now, these two outcomes are indistinguishable.