We Are Not the Horses
Tony Seba likes a brutal image. We are the horses.
It is a good image for the end of a function. It is a bad photograph of 2026.
What I see around me is the opposite surface. Software got cheaper. More software became worth building. New products, agents, wrappers, integrations. The demand for people who can make a system actually work went up. I have lived that as a builder. Copilot is convenient. The work did not vanish. The work changed shape and, for a while, multiplied.
That is Jevons. When the price of a useful thing falls, you do not always use less of it. You often use more.
The horse did not vanish on day one of the car. First came more roads, more oil, more garages, more rubber. The horse population held. Then the function collapsed.
Seba’s line is an end-state metaphor for a task. It is not a portrait of this Tuesday.
Two clocks
The first clock is the one we are on now. Cognition-as-output gets cheap. Insight, taste, responsibility, and integration get expensive. Junior boilerplate starts to die. Whole professions do not. New markets appear precisely because the unit cost of making a thing fell. We are the mechanics of the new market, not the animals it replaces.
The second clock is Seba’s. Analyze tasks, not jobs. Tasks per hour per dollar. When a robot or a model can do the task below human cost, the function dies. RethinkX itself says the early robots fill unmet demand: warehouses, factories, inspection, the care that already has no applicants. Mass technological unemployment is a late-2030s argument, not a 2027 headline. Teleoperation may even give humans a strange decade of remote bodies before full autonomy.
Both clocks can be true. People who only see the first call you a hysteric. People who only see the second miss the boom they are standing in.
For a shrinking population the second clock is the exit. Robot output that shows up in real GDP is how a country with too few children keeps paying for the old. The cruelty is timing. The boom in software demand and the hole in the birth cohorts exist in the same decade. Politics will scream “they took our jobs” and “we cannot find nurses” before breakfast.
Who in Europe can take the growth
Robot growth is not “who has the most robots.” It is who has a shortage of hands, something to automate, and permission.
Poland is the realistic first mover. Fertility around 1.1. Emigration to Germany and the Netherlands. Wages up. Robot density still below Germany, but the installation rate has jumped. Auto, metal, food, logistics. German nearshoring. EU capital. Less sacredness around the chauffeur and the national car. A Polish hall that cannot find welders can scale with cobots. That is Jevons in a factory: cheaper output, more orders, more people who install and maintain. Not a civilizational leap. A takeoff.
Ukraine is the highest necessity and the highest variance. Fertility under one. Millions of workers missing. Reconstruction, defence, building, transport. Drones already taught a generation that machines fill holes. If the war stabilizes, the country can leapfrog plants that will never again be staffed like 2013. Until then it is need, not a boom.
Czechia and Slovakia are already on the curve and too small for a headline flight. Italy has the robots in the north and the unions in the story. It optimizes. It does not embrace. Turkey is still too young in the labor market for desperation, even as its fertility slides. Germany is full and protective. Spain has the Super Power and the political memory of unemployment. There the robot looks like a thief until new demand from cheap electrons creates jobs that no one can staff.
So the European robot boom, if it comes, does not start in Amsterdam and does not start in Madrid. It starts in a hall outside Poznań where the applicants stopped coming.
China will feel robot GDP first because it must. The United States will feel it where energy and warehouses already are. Europe will feel it late, unless the East is allowed to be vulgar about machines.
We are not the horses.
Not this year.
The horse is the task, later, especially the physical one, and first where the crib is emptiest.
The next essay leaves the factory and goes into the body. Appetite is already changing. If protein can leave the animal, a small wet country with too many cows will discover that some of its hardest politics were demand in costume.
