Average Is Over
The future belongs to the brilliant
Description
In 2013, the economist Tyler Cowen published a book with a title that read less like an argument than a verdict: Average Is Over. Cowen teaches at George Mason University and had spent years watching the same puzzle from different angles — why wages for a lot of American workers had stopped climbing, why some households were pulling away while others treaded water, why the recovery from the 2008 crash felt so uneven depending on where you stood. The book was his follow-up to The Great Stagnation, and where that earlier book diagnosed a slowdown, this one made a forecast. The forecast was blunt. The comfortable middle of the labor market, he argued, was being hollowed out, and it was not coming back.
The engine of the change, in his telling, was not offshoring or bad trade deals or a single villainous policy. It was the steady, quiet advance of cheap computing power. Machines were getting good at the routine cognitive work that had employed millions of people in the middle of the income distribution — the clerks, the processors, the mid-skill analysts. What machines could not yet do, and what became newly valuable, was the work that complemented them: managing them, interpreting them, marketing what they produced, doing the things they still did badly.
Cowen's claim was that this split would not stay confined to the economy. It would reach into who marries whom, where people can afford to live, how schooling works, even how governments tax and spend. He was not celebrating any of it. He was describing a machine that, once running, is hard to switch off — and asking what kind of society comes out the other side.
The question we’re asking : If cheap machine intelligence keeps raising the reward for a narrow set of skills, what happens to everyone in the middle?What we’ll see : How a single technological shift reorganizes work, income, geography and the social bargain of an entire economy.
Table of contents
01Chapter 1 — The middle rung is disappearing
Cowen's starting point is an observation that had been troubling labor economists for a while: the American job market was pulling apart at the seams. Employment was growing at the top, among highly paid professionals, and at the bottom, in low-wage service work that resists automation. It was the middle that was thinning. The jobs that had once let someone without a college degree buy a house and raise a family on a single income — factory line, back-office processing, mid-tier administration — were either vanishing or no longer paying what they used to.
The usual explanations pointed abroad. Factories moved to lower-cost countries, the story went, and Chinese exports undercut American workers. Cowen didn't deny any of that mattered. But he argued the deeper driver was sitting on everyone's desk. Computers had become spectacularly cheap and spectacularly capable at exactly the kind of rule-following, information-processing tasks that filled middle-income jobs. A piece of software doesn't need a salary, benefits, or a lunch break, and it doesn't ask for a raise.
02Chapter 2 — Working with the machine, not against it
If machines were taking the routine work, the natural fear was that they'd eventually take everything. Cowen's answer was more interesting and, in a way, more unsettling. The winners of the coming economy, he argued, would not be the people who competed with machines. They'd be the people who worked well alongside them. The decisive skill of the next few decades would be knowing how to get value out of intelligent software.
His favorite illustration came from chess, a domain he knew well. When IBM's Deep Blue beat Garry Kasparov in 1997, the obvious lesson seemed to be that machines had surpassed humans and that was that. But something more subtle emerged in the years after. In so-called freestyle chess tournaments, where humans were allowed to consult computers, the strongest competitors were not the best grandmasters and not the most powerful programs. They were relatively ordinary players who were unusually good at managing several chess engines — knowing which to trust, when to override, how to combine their outputs. The human-plus-machine team beat the machine alone.
03Chapter 3 — A country that quietly splits in two
Cowen was not content to leave the argument in the labor market. His more provocative move was to trace how the same split would seep into the texture of ordinary life. If income increasingly tracks a narrow set of skills, then the divide shows up in where people live, whom they marry, and what kind of future their children inherit.
Take geography. The high earners cluster in a handful of expensive, productive cities, bidding up housing until those places become unaffordable to everyone else. Meanwhile, Cowen predicted, cheaper regions would grow as a kind of relief valve — places with low costs of living where those who'd been priced out could get by, sometimes comfortably, on modest incomes stretched by inexpensive technology and cheap entertainment. He half-joked about Americans building shantytown-adjacent communities in low-cost areas, living decently but visibly apart from the winners.
04Chapter 4 — The meritocracy nobody signed up for
Step back from the forecasts about cities and marriages and chess engines, and Average Is Over is really a book about a bargain coming undone. For much of the twentieth century, a developed market economy carried an implicit promise: that a person of ordinary ability, willing to work, could claim a decent and secure place in the middle. That promise rested on an economy that needed a lot of average work done by a lot of average people. Cowen's wager is that the economy needs that less and less.
What replaces it is something like a hyper-meritocracy — a system that measures skill more precisely than any before it, and rewards it more steeply. This sounds, at first, admirable. If the machines can tell exactly how much value you add, then pay tracks contribution, favoritism has less room to hide, and talent rises regardless of pedigree. Cowen grants some of this. But he refuses to let it read as good news, because a world that measures merit ruthlessly is also a world that offers very little to those who measure poorly, for reasons that are often not their fault.
05Conclusion
The book's title was meant to land as a shock, and it did. But the argument beneath it was patient and, for a market economist, strikingly unsentimental. Cheap machine intelligence, Cowen argued, would keep raising the reward for a narrow band of complementary skills while draining value from the routine work that once filled the middle. The winners would be those who could work with the machines; the divide would spread from paychecks into cities, households and schools. None of it required a villain — only cheap computing and time.
Related books
