Monday, February 6, 2017

A Futurist View of the Welfare State, Part I: Luddites and the 47%

Here is a cool bricklaying robot. Companies can use it to reduce the number of workers on  construction projects, and make their jobs easier. McDonalds is working on full automation for their restaurants (that link is already 2 years old!). Isn’t technology incredible? A full 47% of jobs in the United States are “at risk for computerization”, which is a fancy way of saying that those jobs won’t exist anymore.

Consider the imminently-arriving self-driving car, which I am personally a fan of, as I have to drive in Boston, where I constantly face death at the hands of other Boston drivers. Computers are a lot better at driving than people, and anything that reduces the 30-40 thousand people who die every year (in the US) due to traffic accidents is an unambiguous good and, frankly, a moral imperative.

But while self-driving cars are great for me, they will also “computerize” the trucking industry. And while that’s good for grocery stores, it’s terrible news for truck drivers. In 29 states, “Truck Driver” is the most prevalent job. It’s also an important job, providing above median wages to people of all genders, educations, and socioeconomic class.  

Truck driving jobs are important not only for truckers and their families, but for towns as well. Hundreds of small towns throughout the country are sustained by the consistent revenue that truck drivers bring, eating their food, sleeping in their motels, and pumping gas from their gas stations.

But I want to go back to that number: 47%, and really stop to consider it. What would that be like? An obvious comparison would the Great Depression, right? Let’s take a look:

At the worst of the Great Depression, the American unemployment rate was 25%. Things were famously bad back then, with bread lines, bank runs, Steinbeck, Huey Long, Benito Mussolini, and other sources of worldwide unrest. Things didn’t get much better until the war started, when everyone could work full time killing each other and building war machines.

This isn’t to say that society ground to a halt, but it’s a clear understatement to say that the world didn’t do very well with 25% of the people out of work. And 25 is a lot less than 47.

Now I hear you complaining, imaginary interlocutor. You’re complaining that I’m a Luddite. That the affectation I’m writing with is alienating and maybe a little off-putting. Rest assured, I hear your complaints. But to push back hard against the Luddite complaint, let’s look at their situation: thousands of highly trained and highly skilled craftsmen (weavers and textile workers) suddenly became obsolete with the invention of textile weaving machines. They were worried the years spent learning their specialized craft would be for nothing. So, they tried to break the machines. Which was not a very practical solution, but it certainly got the point across.
Frame Breaking, from Wikipedia

And they didn’t stop trying until enough of them were killed that the rest gave up (which is not generally considered a happy outcome).

Ever since then, textile weaving has been done by machines, or by people in faraway countries who work cheaper than a machine can. The creation of clothes and fabrics became just one more job that the west never thinks about, because for all practical purposes, it doesn’t exist.

Now this is an old argument. The consensus nowadays is that the next effect of this displacement is the creation of more jobs and opportunities. There’s no need to have a traffic cop at every major intersection, now that we have traffic lights, but we still have other things for the police to do. The coming of AI will be the same, people will just do different work.

Respectfully, I have to disagree. I think this is a far more Malthusian situation than most people realize, for one simple reason: people are fundamentally incompetent. Which is to say, they are people, not machines.

Even the smartest, most skilled, most competent person imaginable has a fundamental limit on their competency, some level of skill after which there is no meaningful progression. We’re only thinking meat, and the meat can only do so much (have you tried beating a computer at chess recently?). And as computers and machines improve, more and more jobs and tasks will no longer require people.

This is not a new development; U.S. manufacturing has been continually shedding jobs. Sometimes to globalization, but mostly to automation. Whole swaths of America have spent the last century being automated and globalized out of the middle class and into a pretty upsetting situation. But it’s not as if the same thing isn’t happening other places too, eventually the jobs lost to globalization will be lost again to robots.

Obviously, there will be new jobs. Jobs related to Data science, Machine Learning, and other future-y sounding things you may have heard about. And, yes, some of the workers can learn to do these new jobs. But you’re kidding yourself if you think that all or most of them can. Even assuming the skills can be learned, the fact is that there will probably not be enough jobs (if the ratios are anything like the one for that Chinese factory I linked in the previous paragraph). Also, full disclosure, I've tried to learn data science. It’s hard! And if it’s difficult for someone like me with a college degree in math and science, well, any argument that assumes that the majority of people with at-risk jobs can pick it up in time is either willfully ignorant or deliberately malicious. 

Continued in Part II...

No comments:

Post a Comment