Last night, 60 Minutes put a question to prime-time America: will AI take your job? The number underneath the segment came from the McKinsey Global Institute, whose report “Workforce in Motion,” released last week, replaces the abstract fear with specific arithmetic.
The headline figure: roughly 11 million US workers — about 7 percent of the workforce — may need to move into different occupations by 2035. That is 770,000 transitions a year, 3.6 times the historical average, which McKinsey calls the largest sustained workforce transformation in US history. But the striking finding is not the scale of the displacement. It is the quality of the exits. Only one in seven displaced workers has what the report calls a direct pathway — a move requiring minimal retraining and no pay cut. Another 41 percent face “winding” pathways needing moderate retraining. The remaining 45 percent face “unpaved” ones: significant skill gaps, credential barriers, likely wage loss.
Note what the report does not say: it does not say the jobs disappear. McKinsey projects 41 million jobs created against 36 million eliminated — a net gain of 5 million. The problem is not the number of chairs. It is that the people losing theirs mostly cannot reach the new ones. The labor market is not shrinking; it is splitting. An economist mapped this split more than a decade ago, down to the fraction.
The book that called the fraction
Tyler Cowen is an economist at George Mason University. In 2013 he published Average Is Over: Powering America Beyond the Age of the Great Stagnation, the sequel to his earlier diagnosis of stalled American growth. Its central question is the one he said every worker would eventually face: are you good at working with intelligent machines, or not?
Cowen’s answer divided the future in two. Perhaps 10 to 15 percent of the population, he argued, would thrive in a coming “hyper-meritocracy” — those whose skills complement the machines, who can turn an hour of their own judgment into the output of ten. For everyone else, the machines would not be partners but competition, and skills that once earned a middle-class wage would be repriced downward. His title was the thesis: the comfortable middle — average skills, average effort, average pay — was going away.
Thirteen years later, McKinsey’s one in seven is 14 percent. Cowen said 10 to 15. The two numbers measure slightly different things — Cowen was describing who prospers in the whole economy, McKinsey who among the displaced can move without losing ground — but the rhyme is hard to ignore. The fraction of Americans positioned on the right side of the machines keeps coming out the same.
The barbell, now with data
Cowen predicted a barbell economy: wealth concentrating at the machine-complementary top, a scramble for hands-on service work at the bottom, and the middle hollowed out. The report supplies the empirical shape. Of the jobs McKinsey expects to grow, 57 percent sit in the top two wage quintiles. Of the jobs it expects to decline, more than 70 percent sit in the bottom two. Three categories — office and administrative support, retail and sales, transportation — account for three-quarters of all the workers who will need to move. That is the middle’s occupational glue, dissolving on schedule.
The growth side matches Cowen’s map just as closely. Alongside management roles at the top, the biggest gains come in healthcare and construction — work that is physical, local, and hard to digitize. Cowen predicted exactly this migration toward what machines cannot touch, and he was unsentimental about what it looks like: more of America living like Texas, he wrote — cheap housing, plentiful service jobs, thinner public services. The future of the displaced six-sevenths is not unemployment. It is reemployment at a discount.
What Cowen got wrong, and why it makes things worse
Honesty requires noting that the book’s signature example has aged badly — in a way that sharpens its warning. Cowen’s model for human-machine partnership was freestyle chess, where a competent player plus a computer could beat any grandmaster. In the years since, chess engines got strong enough that the human partner now adds nothing. Complementarity, it turns out, is not a destination but a treadmill: the skills that make you valuable alongside this year’s machines may be absorbed by next year’s.
That is the right lens for the reskilling wave now coming. As the report circulates through C-suites and policy shops, expect a flurry of training initiatives. Read them skeptically. McKinsey itself finds that over 70 percent of the workforce will need some reinvention as AI reshuffles tasks within jobs — a one-time certificate program does not meet a moving target. Cowen’s deeper point was never “learn the software.” It was that the economy now pays for a disposition — the habit of continuously reorganizing your work around what the machines just became able to do.
What to watch next
A projection is not a measurement, so watch the data that would confirm or kill this one. The three declining categories — office and administrative support, retail, transportation — are visible every month in the federal jobs report. If their employment holds through 2027, McKinsey overshot. If they roll over while healthcare and construction hiring stays strong, the bifurcation is arriving on schedule, and the 770,000-transitions-a-year treadmill with it.
Watch, too, which end of the pathway problem the policy response targets. Money for training the one-in-seven who already have a route is money spent on the people who least need it. The test of any reskilling program is the unpaved 45 percent.
Cowen ended his book with advice that reads colder every year: the relevant question is no longer whether a machine can do your job, but whether you are the reason the machine is worth more. One worker in seven can currently answer yes. The other six just got their notice — ten years, 770,000 at a time.
This week’s book: Tyler Cowen, Average Is Over: Powering America Beyond the Age of the Great Stagnation (Dutton, 2013; Plume paperback, 2014).
Sources: McKinsey Global Institute, “Workforce in Motion: Skills and Pathways to Future Jobs in the United States” (Sept 29, 2026); Forbes (Sept 30, 2026); 60 Minutes / CBS News, “AI and the future of work” (Oct 4, 2026); International Finance (Oct 5, 2026). Book: Tyler Cowen, Average Is Over (2013).