The Communism of Machines

Elon Musk says that within twenty years work will be optional and money irrelevant. If he is right, capitalism will deliver the world Marx promised, and nobody has prepared for what people do with it.

Orlin Chotev Orlin ChotevEditor of record · signed column. The facts here are sourced; the conclusions are his.
Illustration: AI-generated for Fact Refinery

At the US-Saudi Investment Forum this year, the richest man in the world made a prediction that would have got him expelled from any economics department a decade ago. "My prediction is that work will be optional," Elon Musk said. "It'll be like playing sports or a video game." Give it long enough, he added, and "money will stop being relevant". Saving for retirement, he has said, "won't matter". Not a universal basic income, he insists, but a universal high income, with "no shortage of goods or services".

It is easy to laugh. Musk has promised fully self-driving cars "next year" since 2016. But the question he raises does not depend on his calendar. Thinking is getting cheaper faster than anything humanity has ever made. The same quality of machine reasoning that cost $20 per million words in late 2022 costs about 40 cents today; Epoch AI measures the fall at between five and ten times a year, faster than computer chips in the microprocessor revolution or bandwidth in the dot-com boom. Call where this leads superintelligence, SI: machines that out-think us at almost everything, at almost no cost. What happens to the people whose lives were organised around work?

WHERE WE ARE

The early numbers are not apocalyptic, and they are not reassuring either. Entry-level professional postings in the United States have fallen 29 percent since January 2024. Unemployment among recent college graduates reached 5.6 percent in early 2026, a level last seen after 2008. Stanford researchers, using payroll data, found a 13 percent relative fall in employment for 22 to 25 year-olds in the jobs most exposed to machine intelligence, while older workers in the same roles held steady. The aggregate still looks healthy: job openings are near a two-year high. The machines are not yet taking jobs. They are taking the first rung of the ladder.

The money, meanwhile, is real. Five American companies will spend close to $800 billion on data centres this year, more than the GDP of Switzerland. That matters for everything that follows: if SI produces the wealth, the wealth belongs to whoever owns those buildings, not to the people it replaces. The question of the century is not whether the machines will work. It is who will own them.

THE IRONY

Marx described the "higher phase" of communism in 1875: "from each according to his ability, to each according to his needs". He was clear that it required abundance. In The German Ideology he imagined a man who could "hunt in the morning, fish in the afternoon, rear cattle in the evening, criticise after dinner", never becoming a hunter or a critic. Communism failed because it never produced the abundance. Planned economies made too little of almost everything.

Musk's optional work is that sentence with a robot added. If it arrives, it will arrive by the road Marx despised: private capital, competition and profit, building machines that make labour unnecessary. Capitalism would achieve the goal communism announced and could not reach.

But there is a catch in the word "universal". A machine that produces everything does not pay anyone. A universal high income is not a market outcome; it is a transfer. Someone has to tax, or own, the machines and hand the proceeds out. Capitalism produces, the state distributes. That is not the end of redistribution. It is redistribution as the only remaining political question.

Income can be engineered. Meaning cannot.

WHO LOSES

It is tempting to see this as a blow to the parties of the left, which have lived for a century on taxing work and redistributing its fruits. The opposite is more likely. When income comes from machine rents instead of wages, the oldest programme of the left becomes the entire agenda of government. What the left loses is its base: no labour, no labour movement.

The right loses something too. Its moral language, from the Protestant work ethic to the self-made man, assumes that work is how a citizen earns his place. A society in which nobody needs to work has no use for that sermon.

And there is a third loser, the one worth worrying about: the citizen. The Gulf has run this experiment for half a century. In Kuwait about 80 percent of citizens with jobs work for the government, and Kuwaitis are under 4 percent of private-sector workers; in Qatar 82 percent of nationals are in the public sector. Oil pays, foreigners work, citizens receive. Political scientists call it the rentier bargain, and its rule is the reverse of the American founding: no taxation, no representation. A population that neither works nor pays for the state has little leverage over it. Musk's world would make every citizen a rentier, and every government an emirate.

WE HAVE BEEN HERE BEFORE

Life without work is not new. It has simply been reserved for a few.

  • Athens. The Greek word for leisure, scholē, is the root of "school". Free time made philosophy, theatre and democracy possible for citizens, because slaves and women did the work.
  • Rome. Under Augustus around 200,000 citizens received free grain. Juvenal gave the arrangement its name: bread and circuses. A people fed and entertained, he complained, had stopped caring who ruled.
  • The tribes. Marshall Sahlins famously claimed hunter-gatherers worked two to four hours a day. Later studies, counting every task, put it nearer 40 to 45 hours a week. What they lack is not work but the separation of work from life. Nobody "goes to work"; living is the work.
  • The gentlemen. Charles Darwin never held a job. Neither, in the sense we mean it, did Lavoisier, Boyle or Byron. The leisured class of Europe produced much of modern science, and a great many idle sons.

VICE OR ELEVATION

The evidence says both, and that the gap between them widens.

When Swedish lottery winners were followed for up to 22 years, between 85 and 89 percent kept working, mostly taking longer holidays. Their life satisfaction rose and stayed up. When Sam Altman's research group gave 1,000 Americans $1,000 a month for three years, they worked on average 1.3 hours a week less, searched harder for better jobs, and spent the money on food, rent and transport. Nobody stopped. On the other side, the "deaths of despair" that Anne Case and Angus Deaton documented, overdoses, suicides, drink, track joblessness with a precision that should frighten anyone selling the end of work as liberation.

The difference is not the money. Lottery winners keep their place in the world; the laid-off lose it. Work gives most people three things besides a wage: a reason to get up, people to see, and a measure of their worth. Remove the wage and you can replace it with a cheque. Remove the other three and you need something no data centre makes.

What follows is the editorial view.

Musk may well be right about the machines and wrong about the people. Abundance is an engineering problem, and engineering problems tend to get solved. Meaning is not one. History's leisured classes gave us the Origin of Species and the Lyceum, and they gave us gambling, opium and decadent heirs, often from the same families. A world without work will be Athens for some and Rome for many, and the line between them will run through every household, not between classes.

So the honest answer to the question "is capitalism achieving what communism failed?" is: it may achieve the abundance, and inherit the problem communism never had to face, because it never got that far. What do free people do when nobody needs them? The societies that answer it will be the ones that treat education, family, faith, craft and citizenship as the real work, not as hobbies for after hours. The ones that answer with a cheque and a screen will rediscover what Juvenal knew. A people that is fed and amused stops asking who owns the bakery.

What to watch

  • Entry-level hiring: whether the 29 percent fall in junior professional postings spreads to mid-career roles [Indeed Hiring Lab, Federal Reserve Bank of New York].
  • Policy: who proposes the first tax on SI compute or a national dividend from machine income, and from which side of the aisle [Congress, UK Treasury, EU Commission].
  • The Gulf: whether Saudi Arabia and Kuwait succeed in moving citizens into private-sector work before oil revenue falls [IMF Article IV reports].
Orlin Chotev is the editor of record of Fact Refinery and has hosted the Bulgarian technology programme Микрофон и мишка since 1997. He writes this column weekly. Corrections →
Was this useful?

💬 Comments

📜 Comment Policy