Welcome To Extremistan! Please Check Your Career At The Door.

Will robots and software eat all the jobs? No. Will robots and software eat your job? Yes, probably. Eventually. Rejoice! …for your grandchildren. You and your kids are likely to have a pretty tough time over the next few decades. Sorry about that.

Everybody who’s anybody is talking about technological unemployment, the notion that technology will soon destroy jobs faster than it creates them; a conversation kickstarted by MIT professors Andrew McAfee and Erik Brynjolfsson, authors of Race against the Machine and The Second Machine Age

Earlier this month, Marc Andreessen, who should need no introduction, weighed in with characteristic optimism:

We virtually never resist technology change that provides us with better products and services even when it costs jobs. Nor should we. This is how we build a better world, improve our quality of life, better provide for our kids, and solve fundamental problems … It is hard to believe that the result will not be a widespread global unleashing of creativity, productivity, and human potential … In arguing this with an economist friend, his response was, “But most people are like horses; they have only their manual labor to offer…” I don’t believe that, and I don’t want to live in a world in which that’s the case. I think people everywhere have far more potential.

Many others are far more pessimistic. One one flank, I give you this misanthropic Hacker News comment on Andreessen’s long-term utopian vision:

Look at the future this guy has concocted in his head: The main fields of human endeavor will be culture, arts, sciences, creativity, philosophy, experimentation, exploration, and adventure. …it’s like he’s never met anyone who didn’t attend a top tier university. Here’s reality: The main fields of human endeavor will be copulating, hustling, consuming low-brow entertainment, eating, and the occasional lunatic running amok.

On the other, Alex Payne responds:

You seem to think everyone’s worried about robots. But what everyone’s worried about is you, Marc. Not just you, but people like you … so much wealth and control in so few hands … Owning a smartphone is not the equivalent of owning a factory … It seemed like a lot of people were going to get rich in the “app economy”. Outside of Apple and Google, it turns out, not so much … Unless we collectively choose to pay for a safety net, technology alone isn’t going to make it happen … Instead, you’re kicking the can down the road and hoping the can will turn into a robot with a market solution.

Andreessen argues:


But it seems to me that there’s a (c): many people will contribute, but not benefit from those contributions, because tomorrow’s jobs will increasingly exist in Extremistan, not Mediocristan.

You won’t be familiar with those terms if you haven’t read Nassim Taleb’s brilliant The Black Swan, which you should. Here’s a primer. Briefly, for our purposes: the remuneration for Mediocristan activities is fixed by boundary constraints — the number of hours worked, the number of clients aided, the number of widgets manufactured, etc. By contrast, the remuneration for Extremistan activities — basketball player, musician, messaging-app co-founder — can scale to an arbitrary amount …

…but only for a tiny fraction of those engaged in the activity. Most would-be pro athletes never make it. Most artists never get to quit their day job. Most startups fail. Few people engaged in Extremistan activities ever become successful enough to start referring to what they’re doing as a job.

I submit that technology is slowly dragging us all, economically, away from Mediocristan and into Extremistan.

Consider college professors: Khan Academy, Udacity, Coursera, edX, etc, allow individual professors to teach hundreds of thousands of students, while legions of adjuncts live in poverty. Consider lawyers: it may be “more cost effective and accurate to utilize software tools to perform many types of legal work” — but the very best attorneys will remain incredibly valuable and expensive.

Consider even doctors: The New York Times warns: “Health care jobs may be safe now,but our sense of what’s safe has been consistently belied by the impact of our technological progress.” Valley legend Vinod Khosla has been arguing for years that expert systems can replace 80% of what doctors do … but at the same time, tech could greatly expand the remit of the best.

It’s not hard to imagine — whisper it — even software engineering moving into Extremistan. Already, everyone wants the so-called “A-players,” but has only lukewarm interest in second-tier software talent, much less the third tier. The best companies hire the best engineers, who, by definition, are a minority; the best engineers work at, or launch, the best companies; technology increasingly allows the best companies to dominate their markets like never before. Extrapolate that twenty years into the future, and what do you get?

This won’t happen to everyone everywhere, obviously — I’m talking about proportions, not the entire population — but I expect the shift will be significant enough to have enormous consequences. While technology will indeed, as Andreessen points out, create new professions and new fields of human endeavour, the fact that technology is an ever-more-powerful force multiplier implies that those fields will increasingly exist in Extremistan, and be partitioned according to power laws; a few will be enormously rewarded, while the majority scrap for crumbs.

I’ve been writing about technological unemployment for some years now, and whenever I do, commenters always chirp, “we just need everyone to become an entrepreneur!” But of course entrepreneurs have always lived in Extremistan … and most of them fail. Everyone who calls for a future of greater entrepreneurship is implicitly calling for us to move ever deeper into Extremistan.

This is not necessarily a bad thing. It seems likely that, considered as a whole, Extremistan is far more creative and productive than Mediocristan. But while its economic pie may be much larger, it is also much more unequally distributed.

I’m not saying this will all happen tomorrow; there will be enormous hurdles en route. (Education, in particular, is as much about context as content, and MOOCs won’t succeed until they figure out a context in which their students thrive.) But it seems apparent to me that the long-term trend is away from Mediocristan and towards Extremistan; in other words, towards a few winners and many losers.

If so, then what will life be like for the losers — by which I mean, the majority? Adjunct professor might not be a bad analogy. Poverty wages, supplemented by odd jobs in the TaskRabbit-esque gig economy, while you and the millions of others like you keep knocking on the doors of success, year after year, knowing all along that only a tiny percentage of you will ever be admitted. Or maybe even no wages at all, like artists or interns today.

Now: do I have any evidence that we’re heading this way? Er. Well. The trouble with speculating about the future in a time of rapid change is that the available evidence is almost by definition insufficient. I hasten to admit that I certainly could be wrong.

But there do seem to be a lot of people singing along in tune with this theory. Lawrence Summers, for instance, who writes:

Technology and globalization give greater scope to those with extraordinary entrepreneurial ability, luck, or managerial skill … Most obviously, the best athletes and entertainers benefit from a worldwide market for their celebrity. But something similar is true for those with extraordinary gifts of any kind. For example, I suspect we will soon see the rise of educator superstars who command audiences of hundreds of thousands for their Internet courses and earn sums way above the traditional dreams of academics.

He doesn’t talk about what happens to the other educators. But Harvard’s Kenneth Rogoff writes

Until now, our societies have proved remarkably adept at adjusting to disruptive technologies; but the pace of change in recent decades has caused tremendous strains, reflected in huge income disparities within countries, with near-record gaps between the wealthiest and the rest. Inequality can corrupt and paralyze a country’s political system – and economic growth along with it.

Will each future generation continue to enjoy a better quality of life than its immediate predecessor? In developing countries that have not yet reached the technological frontier, the answer is almost certainly yes. In advanced economies, though the answer should still be yes, the challenges are becoming formidable.

Jagdish Bhagwati at Columbia concurs: “technological innovation is driving wages down.” Paul Beaudry of UBC observes, “Many higher skilled workers have moved down the occupation ladder and accepted less challenging employment.”

Federal Reserve economists note the rise in “involuntary part-time employment” which “has led to a concern that there is an underbelly of labor market slack not well accounted for by the overall unemployment rate.” The pay gap between college graduates and non-graduates has increased … but mostly because “the average wage for everyone else has fallen 5 percent” over the last decade.

Meanwhile, “For the first time essentially ever, labor force participation & stock market have totally & wildly diverged.”

…All of which sounds like what we’d expect would happen during the beginnings of a general shift towards Extremistan. (Though of course this is hardly decisive proof.)

It’s obviously hard to tease out how much of this is driven by technology, vs. globalization and economic cycles — but last month I attended a dinner hosted by Silicon Valley VC kingpins Draper Fisher Jurvetson, whose partners all seemed awfully worried that tech is destroying jobs faster than people can be retrained. Vinod Khosla agrees: “Technology concentrates wealth in the hands of the creators of technology and the people who fund them … creates more wealth and jobs for a few and takes away jobs at the bottom end of the spectrum.”

My point: if I’m wrong, I’m certainly not alone.

So, if this is all true — what can most people do?

Erik Brynjolfsson cites three categories of jobs immune to robots and software, for the moment: “creative tasks and inventing new things – ideation, some people call it … [those] involving interpersonal relationships: motivating people, comforting people, caring for people … [and] fine motor control – the kinds of things that a barber or a gardener or a cook or a janitor does.” The first category is classic Extremistan, of course. The second and third are not … but they’re notoriously poorly paid, gig-economy type jobs. And, he helpfully notes, “you can imagine machines getting better and better in all three as well.”

The most pessimistic are arguing:

Unless we intervene, the same economic system that has produced this astonishing prosperity will return us to the Dickensian world of winners and losers that characterised the beginning of capitalism … Our birthright as humans – the ability to produce things by our labour that others find valuable – may become economically worthless … on the output side of our new robot economy, we will have material abundance undreamed of by earlier generations. But on the production side we will have an economy increasingly independent of human labour and so unwilling to pay for it. Hence the crisis … The material abundance being wrought by ever increasing automation makes the affordability and sustainability of a universal basic income more credible.

This is probably a good time to mention that I’m actually an optimist.

My own thesis has long been this: technology is eating jobs as we know them, and that’s good, because when you step back and look at them, jobs as we know them aren’t particularly desirable. We want to be on a slow trajectory towards the utopia Andreessen describes; a low-scarcity society where people mostly spend their time doing what they want to do, rather than “jobs,” which largely consist of people being paid just enough to do things that they otherwise wouldn’t1.

To get from here to there, obviously, we’ll have to shed a lot of jobs … so the devouring of jobs by technology is a good thing, an indicator that we’re on the right path.

Unfortunately, our economic and even, to some extent, our ethical systems are built around the assumption that most able-bodied people have jobs. Those systems aren’t going to change as fast as the economy itself. So the transition, for (probably) you and your kids, is going to be more than a little wrenching — by which I mean, a lot of people who didn’t expect it are going to watch their once-prosperous families sink into relative poverty. But, barring catastrophe, their kids ought to have it great. I hope that’s some consolation.

In the interim, what we need is


Indeed. But if more and more people become unemployed — by which I really mean, fighting to get by in Extremistan — then only one safety-net option will work: a universal basic income. Marc Andreessen is exactly right when he says technology can kindle the kind of economic growth we need; from my perspective, we need it to make a basic income a viable option. I just hope that happens before too many lives are ruined because our politics evolve orders of magnitude slower than our economies, much less our technologies.

1This isn’t true for everyone — many people, including me, actively enjoy what they do, and would continue doing it or something like it for free even if economic utopia descended tomorrow — but let’s not lose sight of the fact that it’s true for most people worldwide.