The Robot Apocalypse Has Been Postponed

But just as I don’t actually share Williamson’s specific theological perspective, I have a core disagreement with Yang: His biggest policy proposal, a guaranteed basic income for every American, is a response to a disruption that I’m not persuaded is actually happening.

Yang argues that we need this guarantee because robots and automation are already destroying, and will increasingly destroy, the employment prospects of millions of Americans. His argument assumes that we’re living through an age of accelerating technological transformation — which seems intuitively correct to many people — and that we need a government that’s ready to compensate the losers even as the winners keep on generating breakthroughs.

But intuition might be deceiving. The best reason to doubt Yang’s story is contained in productivity statistics, which measure the output of the gainfully employed and which traditionally rise rapidly during periods of technological change — because even if workers are losing their jobs to the spinning jenny or the automobile, other workers should be increasing their productivity with the new technology’s assistance.

Lately this hasn’t been happening. Instead productivity growth in the developed world has decelerated over the last decade. To quote a recent summary, in mature economies “labor productivity growth rates halved from an average annual rate of 2.3 percent in the period 2000-2007 to 1.2 percent from 2010-2017.” Combine that with the slow, consistent trend back toward full employment in the American economy — again, not what you’d expect if the labor market were being upended by technology — and the story of our times looks more like stasis than automated revolution, more like the stagnation discerned by a number of heterodox thinkers than the acceleration of conventional wisdom.

Yang and I wrangled about just this question when he graciously came on our Op-Ed podcast, The Argument. He suggested that what we’re seeing in the statistics is that automation for now is just holding down wages and shunting people out of industrial occupations and into low-paying service sector work … and that come a few more breakthroughs and the next recession, when companies will inevitably seek roboticized efficiency, you’ll start to get far more significant disruption.

He could be right; he’s certainly right that automation has had some impact on middle-class jobs, influencing the populism roiling Western politics. But it seems equally plausible that the real state of things is captured by my colleague Neal Boudette’s update on the status of the self-driving car, long portrayed as a technological breakthrough poised to throw lots of people — from long-distance truckers to cabdrivers — out of work.

The boosters predicted that thousands of autonomous vehicles would be on the road this year. Not so — and here’s Boudette quoting an expert arguing that the revolution might be “way in the future” instead: Salesky said Argo and many competitors had developed about 80% of the technology needed to put self-driving cars into routine use — the radar, cameras and other sensors that can identify objects far down roads and highways. But the remaining 20%, including developing software that can reliably anticipate what other drivers, pedestrians and cyclists are going to do, will be much more difficult, he said.

This kind of 80/20 problem feels like a characteristic story of our low-productivity-growth era. It’s not that we’ve ceased to make scientific progress. It’s that our research keeps failing to translate into the kind of dramatic real-world change that we learned to expect from earlier expansions of human knowledge. Across a range of fields we’re constantly being teased by the promise of the big change — the alternative-energy economy, the cure of Alzheimer’s or cancer — and getting, instead, a lot of impressive specialized innovation. Which is not a small thing if your disease is cured by that specialized innovation — but on the societal scale, it’s not the acceleration and turbulence that Yang’s prescription assumes we will experience.

On Sunday I wrote about the failure of America, and humanity, to match the greatness of the Apollo project across the 50 years since, or to fulfill the revolutionary, to-the-stars promise of the space program. After writing that column I read a Twitter thread from the technology writer Timothy Lee that imagined “a world where humanity never invented the transistor and therefore never had a digital revolution.”

In that world, the obvious interpretation of economic history would be that the discovery of fossil fuels gave humanity a one-time growth spurt. It would be clear that the invention of the steam engine triggered a 200-year, S-shaped growth spurt. From the perspective of 2019, people would view the moon landing 50 years earlier as marking the point where the S curve leveled off.

Instead, we did invent the transistor, and thus added the internet-fueled boom of the 1990s to the super-expansion that trailed off in the 1970s. So what was just a coincidence of timing — in Lee’s thought experiment, at least — created an expectation that we could just get another innovation boom after that, and still another one, restoring exponential growth, keeping the Singularity in reach.

But if long expansions are actually as hard to generate as they appeared before the 18th century, then we shouldn’t expect a return to pre-1970s conditions, in which all our 80/20 problems suddenly get solved with Edisonian rapidity. (Though Lee himself thinks self-driving cars may be closer than the new pessimistic mood believes.)

Instead we should anticipate an age of slower growth, in which, as Lee wrote, “the living standards of mature postindustrial civilizations … converge on a sustainable level” instead of leapfrogging one another, and many new inventions are like the robot who shows up in an early episode of the near-future HBO show “Years and Years,” in a scene that promises Isaac Asimovian scenarios until it turns out that the robot’s owner just uses him the way so many people use the internet.

Yang was disappointingly muted in the first round of Democratic debates, but he’ll be back for the second round, and it would be nice if his presence and signature proposal inspired a question for all the candidates: Do you think our era is defined by acceleration or stagnation?

A lot turns on the answer. Both the wisdom of particular policy choices and the correct approach to social discontents depend on whether we should fear the coming robot age — or regret, instead, that our robot overlords’ arrival may be permanently delayed.

This article originally appeared in The New York Times.