For most of history, a man who could lift heavy things and keep lifting them all day had a livelihood. Strength got you into the harvest crew, the quarry, the army. It was a qualifier: the trait people picked you for.
Then engines arrived, and over a century or so the market value of raw human strength fell to roughly nothing. American folklore preserves the moment of crossover in the ballad of John Henry, the steel-driving man who races a steam drill through a mountain, beats it, and dies with his hammer in his hand. The song remembers him as a hero; the railroads kept buying drills.
I think we're living through the same transition again, and this time the thing being commoditized is not muscle. It's cognition, the trait we rebuilt the entire economy around after muscle stopped mattering.
Strength didn't disappear. It split.
Here's the detail that makes the analogy useful rather than just alarming. When engines made strength economically obsolete, it didn't become worthless. Its value broke apart, and the pieces went in different directions. As spectacle, strength did better than ever: nobody has been hired to carry stones in living memory, yet millions of people will watch a man deadlift half a ton on television. As a status signal it survives too, which is roughly why gyms exist. What died, specifically, was strength as a qualifier. Nobody gets a job anymore because they can lift the rock, and nobody expects to.
Intelligence looks like it will split along the same seams, and chess already ran the experiment for us. Deep Blue beat Garry Kasparov in 1997, and for years now the engine in a free phone app has been stronger than any human who has ever lived. By the logic of pure capability, nobody should bother watching human chess. Instead the game is more popular than it has ever been. What people wanted to watch was never the best available chess. It was a person under pressure.
So intelligence as spectacle survives. As a status signal it will probably linger for a generation or two, because status is sticky. But intelligence as a qualifier, the thing you get hired, recruited, or admired for, seems to be going the way of the strong back.
Where the value goes
Herbert Simon described the mechanism back in 1971: a wealth of information creates a poverty of attention. Value pools at the bottleneck. When intelligence was scarce, intelligence was the bottleneck, so it commanded the premium. Make it abundant and the premium has to move. The question worth arguing about is where.
My bet for the biggest winner is judgment. When competent output costs nothing, the scarce act becomes choosing: which question to ask, which of a thousand adequate drafts is actually good, which confident answer is quietly wrong. That skill was always valuable, but it used to come bundled with the production work. Now it's being unbundled and sold separately, and the people who have it are going to notice their leverage.
Accountability is the second, and I find it the more interesting one. A model can't be sued, fired, shamed, or imprisoned. It bears no consequence for being wrong, so its assurances, on their own, are worth nothing. Somewhere a human has to put their name on the decision, absorb the downside, and be findable afterward. Working on an assessment platform has made me notice how bad we collectively are at evaluating this: we have a century of refined instruments for measuring what people know, and almost nothing for whether they should be handed something irreversible.
This is also, I suspect, part of why we won't follow machines the way we follow people. Leadership runs on the fact that the leader has something to lose.
Agency is simpler. The machines do what they're told, so deciding what to want, and starting without being asked, stays scarce by construction. Anyone managing a team that uses these tools can already sort it into people who direct the machine and people who wait for the machine to be pointed at something.
And then there's the premium on the provably real. When anything can be synthesized, 'a human actually did this' turns into a label people pay extra for, the way 'handmade' already commands a markup on goods a factory could produce more precisely. I'm honestly not sure how large this market gets. It might stay a niche, like vinyl. But live performance, in-person teaching, the craftsman with a waiting list: the pattern keeps showing up.
The body, unexpectedly
The strangest reversal is physical. Robotics researchers have known since the 1980s, under the name Moravec's paradox, that we had the difficulty ordering backwards: computers matched our reasoning decades before any robot could reliably fold a towel. Abstract thought turns out to be the recent, shallow layer of what brains do; perception and dexterity are the deep old machinery, and they are brutally hard to replicate. Which means the plumber, the surgeon, and the electrician are better protected from this wave than the analyst who outscored them on every test. We spent a century making physical capability economically irrelevant. Some of that value is quietly coming back.
Two predictions people get wrong
Creativity is not the safe harbor it's assumed to be. Generating ideas, fluently and in bulk, is precisely what these systems do best. What they can't do is stake anything on one idea. The defensible part of creative work is collapsing toward taste and nerve: knowing which idea is good, and being willing to be publicly wrong about it.
Empathy is trickier, and I keep moving the line on this one. Detecting what someone feels is partly automatable already, and the models are reasonably good at saying the comforting thing; plenty of people report that talking to one helps, and I believe them. What a machine can't supply is the fact of having chosen you: attention from someone with limited time who spent some of it on you anyway. Whether that distinction keeps mattering, or whether we quietly stop caring about it, I don't know. It's the prediction in this essay I'd least like to be graded on.
Where the analogy limps
I should say where I think this argument is weakest. It assumes intelligence loses its prestige the way muscle did, and prestige is exactly where the parallel is shakiest. Credentialism has survived every technology that made knowledge cheap; the printing press was supposed to be the end of the university too. It's entirely possible we keep venerating smart people long after the veneration stops making economic sense, and that the new hierarchies rebuild themselves with roughly the old winners on top. I think the repricing is real. I can't prove the timeline, and I'd distrust anyone who claims they can.
What we might start selecting for
Still, here's the version of the future I find both plausible and worth wanting. For a few centuries we sorted each other mostly by cognitive horsepower. Schools ranked children by processing speed, careers rewarded analysis, and mate selection leaned on the degree and the job title as proxies for whether a life with this person would be good. Those proxies are losing information value fast. A brilliant piece of work tells you much less about its author when the brilliance might be rented.
What the proxies always stood in for is the stuff underneath: whether someone is steady when things break, whether their word predicts their behavior, whether they can sit across a table without being half-gone into a screen. When the loud signal of cleverness quiets down, those quieter signals get easier to hear. The clearest early casualty is the brilliant jerk. Teams tolerated the difficult genius because the genius part was scarce enough to pay for the damage; once any team can summon the brilliance from a machine, the difficulty is all that's left to pay for.
Whether we actually take the opportunity is open. Character is much harder to measure than knowledge, which is a large part of why we defaulted to measuring knowledge, and it's roughly the problem we've been chewing on while building Agonora. I can report that it stays hard. But the direction of pressure seems right to me: a world that can no longer sort people by intelligence still has to sort them by something, and the honest candidates left are the old ones. Judgment. Trustworthiness. Presence. The way someone treats people they don't need.
None of this arrives on schedule, and the transition will be uglier than essays like this one tend to admit; the industrial version certainly was. But there is one piece of luck in it. The traits moving up the scarcity ranking look trainable to me: judgment sharpens with practice and honest feedback, trust accumulates from commitments kept. A strong back in 1910 had nowhere to retrain into. A person starting now has somewhere to go, and a fair amount of time to get there.