Robots vs the middle class: everyone's endangered, white people less so


Originally published at:


Please, let’s not blame the robots, who are just tools. Automation is just productivity growth, and the debate here, as always, is how the benefits of productivity growth are apportioned, not with the pace of technological development. If anything, this is slowing down in its adoption, since growth has been anemic for the last decade.

How productivity increases are apportioned are a political and moral choice, not a technological or “economic” one. For example, if 10 million working-class jobs are automated away by replacing human drivers with machines, a favorite parable of the time, we can respond either by (a) letting capitalists who own the companies who own this technology making off with all of the resulting gains in productivity, or (b) distributing the benefits of this new technology by, for example, making low-cost municipal car fleets available to working people in the city.

It is wholly unsurprising that those with power and influence (here summarised as “white people”) are shielded from the effects of “the robots”, because the actual problem with automation is just a political one.

There may be some people out there who somehow think that there is no longer any high-productivity work available that we might assign workers displaced by automation to, but these are people who simply have their eyes closed against the idea of a future.

If people are left unemployed as a result of productivity gains, the problem is with capital allocation (we are not paying for useful work that could be done), not with technology.


I work with a high school robotics team, teaching them how to be robot designers. Are their jobs also destined for replacement via automation?


Java application developer, internet security specialist, nurse practitioner, dental hygienist, statistical analyst, data mining specialist, physical therapist…

Most of these are jobs a robot can’t do. Is a physical therapist’s job to move the patient’s arms and legs - or to promote healing and wellbeing? Does a dental hygenist just clean the teeth, or also hold the hand of the terrified patient? Even the security specialist is in the business of selling his services. Not many jobs involve only data manipulation skills.


Those are the first to go … That stuff will be done by machine learning.

But by then the robot uprising will be complete.


Engineering jobs have been replaced by automation for a long time. What tends to happen, though, is that increasing automation of routine engineering tasks - like calculations - has freed engineers to do more complicated things, and improved machine tools and production processes have brought down the price of those complicated things. We’re all familiar with the digital computer, which went from racks of tubes holding maybe the equivalent of a thousand gates to SOCs with many billions of gates on a single chip which cost only as much as one of those tubes. But there have been similar revolutions in biochemistry and materials science, they just are not so familiar.

My short answer is that the problem we have in engineering is we can do one of two things: teach people current methods and tools and hope they won’t be too outdated when they enter the workforce; or go back to education and teaching people how to think and learn so that they will adapt to the needs of the workplace and society. In my first year at U I found crystallography a bit boring, and my supervisor said that I wasn’t learning crystallography, but how to think in three dimensions in a structured way, and how to classify things by a variety of metrics. He was right. There’s vocational training and there’s education. There are some vocational jobs, like construction, which can only ever be partly automated (plumbing is a good example - the making of joints is now practically automated, but fitting piping around the system is not likely to be a job for robots for many, many years.) But some of them will disappear. Improved technology in appointment systems means my dentist only needs one receptionist instead of two, for instance. Personally, having made several career changes, I feel that what we really need is more education to learn up to age 18, or even early 20s.


Good points.

It’s still easier to just blame the damn robots. Agitates the masses for…uh…certain ends.


I wonder if robots are going to be considered white before African-Americans get to have unfettered access to the tools of white supremacy ?


Thank god businesses are increasingly amenable to sharing profits with the public.


I think that’s what Cory was trying to say, but he muddled his meaning with a poorly structured sentence.


Why does “robots will do all the work” have to be a bad thing?

We could just, you know, stop bothering to have 'having a job" be a requirement of living comfortably. We’ve been in the age of Bullshit Jobs for a while now, why not just get rid of them and let people figure out what they want to do with their time themselves?

tl;dr we can should must and will go Fully Automated Luxury Communist and probably could have a while ago…


…or just my poor reading skills. I think you’re right!


I know, right? It’s almost like they decided that world problems were what they should be solving, and not the problem of racking up a higher score than whatever phantoms they’re competing with…


I think yours is an excellent and thoughtful post and I agree with what you are saying.

I’m not sure about your conclusions here though and I think you might be underestimating the degree of change that is coming, I think it goes beyond our current understanding of productivity.

It’s one thing for Cory to point out that middle class jobs are going to face significant challenges from automation, but the idea that these “high middle” jobs are somehow safe is wholly mistaken.

All the people displaced by automation are left competing for the “high middle” jobs, so standards required for those jobs will rise dramatically, and much of the work force across the board will simply be displaced, unable to find work, because human labor will no longer be the major component of productivity.

This will be a social shift the likes of which has never been seen before.

Those left displaced by automation won’t have money, so they won’t be buying the stuff that’s produced, participating in the economy. So how are the people in the “upper middle” jobs getting paid? Automation will be encroaching around even those safe jobs.

A significant proportion of healthy adults will be unemployed within decades from now, and they will either be granted basic housing and living conditions by some yet to be devised economic mechanism, but one that is by definition socialist, or they will be a desperate army living in poverty and kept in check by jack booted thugs until they revolt.

I’m not seeing where the useful work you speak of is going to come from.


Lesson? Do things that help people, and don’t be a wage slave.

They will probably participate in an economy, just not that economy.


Required reading:


There is a third option. Lifetime education and a re-skilling. It’s something we should have been doing for fifty years now. It’s not even a drain. A skilled and educated workforce is more productive. There is this persistent myth that the jobs economy is driven by necessity. It’s not. We got along for most of history without most of the jobs that now exist or ever have existed. Jobs evolve to serve markets, markets emerge to fill demand, and demand comes from people. As long as people are willing to innovate, we can find new ways to serve one another doing things machines cannot, and relying on the machines as exponentiators. The only way this fails is if machines can do literally anything a human can, at which point we’re at the level of human-equivalent machine intelligence, at which point I doubt the global economy can survive anyway, but that will be one among many problems if you can make new people with a stroke of the old ctrl+c/ctrl+v.


That’s an interesting thought, and probably along the lines of what @astazangasta was getting at.

I guess work has been shifting toward service industries for a long time now, particularly in the west where a lot of labour has already shifted off shore.

I do notice a tendency in these discussions though to underestimate the impact automation and machine learning is going to have in the near future.

I’m glad to see this kind of discussion becoming more common, it’s a looming crisis and I think we already know that government is ill-equipped and unwilling to act.

Edit: OK so apparently the issue is so topical at the moment it got covered in Tom the Dancing Bug, where everything I said was surmised in one pithy panel. I’m the one who’s out of the loop…

#NeedsMoreLikes (formerly known as "All the Likes")

Disastrously, education and job training are becoming increasingly unaffordable with regular tuition hikes and stagnant wages. Add to that higher education’s pathological resistance to diversify from the existing degree/certification model, and we’re well on our way to winding up with double-digit unemployment that will only snowball as it drags the economy down, widening the gap between prices and spending power eventually leading to runaway inflation. Because whether social services and/or charity keeps them alive or not, vast and growing unemployment will generate huge unsustainable costs one way or another. We’ve already watched this disaster play out in the US in emergency rooms as government failed to seriously reform health care. Now imagine that problem spread to every sector of the economy, globally.

Underestimating the impact of machine intelligence is one thing I try not to do. And it’s why education and training is our only chance of avoiding worldwide economic catastrophe. Unfortunately, the Thatcheresque blowhards currently rotating into the Western halls of power are quite happy to leave the “losers” in the lurch, because they’re too goddamn fucking bloody stupid to understand that a big dispossessed class will drag us all into the shit.


It’s at the very least also a logistical problem. Low-cost municipal car fleets aren’t going to put food on the tables of former truck drivers, but they will take it away from now-former taxi drivers.