You Are The Next Horse
In a Nutshell
Machines are replacing human labor the same way tractors replaced horses, because most jobs are routine tasks machines can do better, faster, and cheaper. The argument that people will simply move to new and more complex work fails as AI takes over intellectual tasks, leaving only limited intrinsic demand for human skills like childcare or live performance. The critical solution is ownership: humans must own shares in the machines and their output, since labor will no longer be the binding constraint on economic productivity.
These notes were generated by AI and may contain inaccuracies.
When was the last time you saw a horse? Most likely, he didn't have a job. Today, there are only about 6.6 million horses in America. The American Horse Council says people raise them mostly for sport, recreation, and therapy. In other words, we keep horses because we love them, not because the economy needs them.
A century ago, this was not true. In 1918, American farms housed approximately 26.7 million horses and mules. In 1920, those animals provided about 89% of the traction power on American farms. Traction power is the force that pulls plows and carts. For centuries, the horse has been the best source of all-purpose kinetic energy we have.
Then, over the course of a few decades, machines took over doing almost all of that work.
I think the same thing will happen to most human workers. Many economists say that people are not horses, and I think they are wrong. You are the next horse.
Horses and mules were used to transport goods and people. They pulled plows on farms and carts in cities. They even operated the machines. On farms, horses walked in circles or on treadmills to operate threshing machines, mills, and hay balers.
In 1850, about 70% of all horsepower in America came from real horses and other working animals. When steam engines were new, their makers measured their power by comparing them to the horses that would eventually replace them. That's why we still measure engine power in horsepower.
An entire economy grew around the horse. In 1910, approximately 440,000 Americans drove horse-drawn wagons to earn a living. More than one hundred thousand others worked as stable workers, saddle makers, and chariot builders.
The horses also had to eat. From 1880 to 1920, about a quarter of America's farmland was used to feed horses and mules. That land was used to grow oats and hay for animals instead of food for humans.
The cities were also full of horses. In 1900, New York City had approximately 130,000 horses. Each horse produced between 15 and 30 pounds of dung every day. So, the city's horses produced about 2 to 4 million pounds of dung per day in total.
Horses were also dying in those streets. In 1880, New York transported nearly 15,000 dead horses, or about 41 horses a day. The horses were also dangerous. In 1900, horses and horse-drawn vehicles killed 200 people in New York. For each individual, this is a higher death rate than what New York experienced due to cars in 2003.
Then the machines arrived, replacing the horse with one job after another. In 1912, traffic statistics in New York showed that cars outnumbered horses for the first time. The last horse-drawn tram in the city ran in 1917. In 1922, the last horse-drawn fire engine in the fire department made its final journey.
On farms, the tractor had the same effect. In 1920, American farms had approximately 246,000 tractors. By 1960, it had more than 4.5 million tractors. Meanwhile, the number of horses and mules on farms has fallen from its peak of 26.7 million to just around 3 million. Thus, by 1960, there were more tractors on American farms than horses and mules.
Simply put, farmers have reduced their breeding of them. By 1940, farms were raising less than half the number of ponies they needed just to replace their existing horses and mules.
People who worked with horses also faced difficult times. Between 1910 and 1930, the number of horse-drawn carriage drivers decreased by 75%. The number of stable workers decreased by approximately 90%.
But the new work moved to the new machines, and many workers moved with them. In those same twenty years, the number of drivers and truck drivers rose from about 46,000 to more than 800,000. The Teamsters Union was founded in 1903 by men who drove horse-drawn carriages, and today it represents truck drivers. The driver learned to drive the truck, but the horse could not.
So, why did the machines win? It was better, faster, cheaper, and safer. This is the underlying principle of all types of automation. The tractor does not get tired, and does not need to be fed during the winter when there is no work.
In those days, one writer estimated that a motorized truck could do two and a half times as much work as a horse. The farmer could not save money by feeding his horses smaller quantities.
In 1983, a Nobel Prize-winning economist raised this point. Reducing a horse's oat ration may delay its replacement with a tractor, but feed rations have clear limits, as do wages. That economist also predicted that machines would do to human workers what tractors had done to horses.
On farms, tractors first reduced the role of horses, then eliminated it completely.
Many economists disagreed with him. In 2015, two researchers at the Massachusetts Institute of Technology wrote an article entitled "Will the fate of humans end up like that of horses?" Their answer was: "Fortunately, humans are not horses." Two other economists expressed this by saying: "Humans have a comparative advantage in new and more complex tasks in a way that horses never did."
In May of this year, an economist published an article entitled "You are not a horse." To be fair, these economists have been right for 200 years. When machines took over the work on farms, people moved to factories and then to offices.
But I say that humans are like horses because the vast majority of what people do to earn money is hard, routine work. Eight of the top 10 professions in America pay less than the average wage. Those eight include the cashier who bills you, the fast-food worker who takes your order, the store clerk, the worker who stocks the shelves, and the office clerk.
For 59% of American jobs, the typical requirement to start is a high school diploma or no formal qualification at all.
But don't people always move on to new and more complex tasks? That's what economists say. But over the past century, more and more of those new tasks have been intellectual work, and intellectual work is exactly what artificial intelligence is learning to do.
In November 2025, McKinsey looked at technology that had already proven effective. It was estimated that, theoretically, this technology could automate about 57% of the working hours Americans spend. In a five-day work week, this means all of Monday, all of Tuesday, and most of Wednesday.
The rule applies to humans just as it applied to horses. If a machine can do a job better, faster, cheaper, and safer than a human, you will choose the machine.
You may have heard this phrase: "Artificial intelligence will not take your job, but the person who uses artificial intelligence will take your job." An economist said that at the World Economic Forum in 2023 and people have been repeating it ever since.
Two years ago, I made a meme about that phrase. In it, a farmer speaks to his horse and says: "You will not lose your job to a tractor, but to a horse that has learned to drive a tractor." This ironic image is ridiculous, and that's the point. The tractor never needed a horse, and eventually, the farmer lost his job too, and even that old tractor was replaced by bigger machines.
The Teamsters are the real-life version of that ironic image. They actually learned to drive trucks, and it worked for about a century. But now, vehicles are learning to drive themselves. In March of this year, Waymo said its self-driving cars were delivering 500,000 paid rides every week. So, about 50 times a minute, someone gets into a car with no one behind the wheel.
That's the problem with saying someone is using artificial intelligence. It assumes that a machine will always need someone to use it.
So, what's left for the horse? Its old function is gone. Nobody wanted the horse itself. Farmers and wagon drivers wanted traction, the amount of power applied to the plow or wagon. The horse was simply the best way to get that power until it stopped being that way. Economists call this kind of demand derived demand.
What's left is the leisure horse, which we keep because we love it. The rider wants a real, live horse, and no machine, no matter how good, can replace it. That's what I call intrinsic demand.
There will be intrinsic demand for humans, too. Some people will pay more for a real human to look after their children or to play live music on stage. That demand is real, but it's limited. These specialized jobs won't justify a full labor market.
But there's one difference between humans and horses that really matters. It's been identified by MIT researchers. Humans can own capital, while horses cannot. A horse has never owned a share in a tractor. That difference is the other half of post-labor economics.
If machines do most of the work, then people should own a share in those machines and what they produce. That's why I define post-labor economics as a model where human labor input is no longer a binding constraint on economic activity or productivity.
Horse labor ceased to be a binding constraint by the middle of the last century. Now it's human labor's turn. So yes, you're the next horse. Learning to drive a tractor won't save you, but owning a share in that tractor might.
Keep David Shapiro in your library
Save the videos and channels worth coming back to, and find them again in one place.
