Thread by @atroyn

I've long said that the threat to jobs of automation is grossly overstated.

logicmag.io/05-the-automat…

Robots aren't taking people's jobs. The productivity data simply doesn't bear it out - marginal product of capital hasn't noticeably moved for decades.

Automation in general, and robotics in particular, simply can't do the vast majority of even menial work. Robots cannot fold laundry without special setup . They cannot pour drinks without careful preparation.

The recent Berkeley report on autonomous trucking goes into great depth on this;

laborcenter.berkeley.edu/driverless/

In their in depth and frankly surprisingly technically sophisticated analysis, they find the threat is from the erosion of labor protections, and not technical disruption.

We've seen this kind of technical disruption before and the natural labor policy experiments that went along with it, with containerized freight.

Ports adopted different models across the US.

Some ports simply fired en Masse. Others resisted containerization and were out competed by others. (SF vs Oakland)

Others adopted labor-management partnerships and earn outs (New Jersey). Those soon faced labor shortages as the gains from efficiency were astronomical and despite needing only 20 men where 80 had worked before, container freight increased volume 10-20x.

We have policy models and experimental data to support them. But that said, automation doesn't explain wage growth stagnation. The productivity gains aren't there.

So who benefits and how ? That's the question to ask.

It concerns me that automation and robotics, technologies that are like humanity growing an extra species capability, are being used as a stick in the labor-capital battle.

The reaction to capital's attack on labor has the potential to do significant collateral damage to the advancement of technology that can free us all from drudgery forever. That won't do.

Henry Ford, in "my life and work" explains the boon of automation. Where before he hired five men for one machine, now those same five men could run five machines and make their output five times more valuable.

That's simply not what we see happening. If capital could get increases to productivity from automation, it would be hiring skilled work in droves to run those machines. Robots need people, I promise you.

Yeah, the machine in McDonald's can now take your order. Have you ever seen a MDs cashier ONLY taking orders? It frees their time for the value add, actually making the food quickly.

tl;dr robots aren't taking jobs. If they were being widely deployed we'd see it in the productivity figures, but we don't.

Wages are flat for other reasons.

If you want proof go to any self driving car company jobs page and look at how many postings they have for drivers / operators / testers.

These are skilled positions.