Found it

Found the post I was ranting about below, and I figured out the trick the guy pulled. My memory was faulty, but there was indeed a a large deception.

First, the robot story implies an acceleration of manufacturing productivity (more output with fewer workers equals faster productivity growth), for which there is no evidence.

Can you detect it? Nice and subtle. An “acceleration” of manufacturing productivity. Not an “increase.” It only takes a small annual increase of a few percent to double output with the same number of workers over time. But to “accelerate” the increase of manufacturing productivity is a whole different thing — this means if you measure the acceleration of the increase (which hasn’t happened) you’ll get a nearly-flat graph where it looks like productivity is flat. Which is bullshit.

It’s like saying your car isn’t going any faster because though you’ve gone from 60mph to 80mph, you’re actually not going faster because you went from 10mph to 30mph with much greater acceleration.

Nice trick. I admire the intellectual chutzpah it takes to pull such things because only a small percentage of people will know you are wrong, but those who do know will really know it.

This dude’s graph, which isn’t actually graphing what it claims to, but rather the “acceleration” I mentioned above:

manemp2

Reality:

growth

Note that in the above, the productivity rate is per annum, averaged over a decade. So that means that a 4% per annum increase in a decade leads to ~50% greater output per worker in only 10 years. But it’s not accelerating, and never really has, so you can make a nice flat graph if you plot that. Which is stupid.