Why I Doubt UBI Is Coming

Arvid Kahl posted on X that he was surprised we’re having all these AI debates without constantly mentioning universal basic income. I quote-posted: “UBI unfortunately is unlikely, the likeliest answer will be dystopian levels of poverty and wealth disparity. Oligarchy on steroids is far more likely and backed up pretty much all of human history. Hope I’m wrong on that.” It got a lot of replies, so this is the longer version of why I think that.

Robots become the measure of wealth

My bet is that robotics becomes the only moat, and the way to gain more wealth. Suppose you own a workforce of a billion robots, all pointed at producing food and doing every other level of production in a society. Logically, you’re now the wealthy person. In that world we’d have shifted from counting dollars to counting how many robots you have.

AI and automation lead to more robotics, and eventually to humanoid robots that do the tasks humans do, 24/7, better and with fewer errors. Once that happens, capitalism as it works today will favor them over employees. So I think the people with the most money right now will be the ones who can amass armies of robots. They’ll use those to keep their positions and widen the gap between themselves and everyone else.

Think about catching up from the other side of that gap. How long would it take one person to amass even a small force next to a billion robots? You won’t be able to enter the game, because a few families at the top will control everything. Prices could fall a lot, since you can produce so much more of everything, including services and entertainment. But the one who controls the means of production is the one in power.

None of this is happening yet. I wrote about how McKinsey added agents on top of its people, and that’s the picture at one of the biggest knowledge-work firms right now. What I’m describing is the step after that, where the labor is physical and the owner of the robots doesn’t need anyone to supervise.

That’s a prediction. Where we’d be starting from is already lopsided. The Federal Reserve’s distributional financial accounts put the top 0.1% at 15.0% of US net worth and the rest of the top 1% at 17.5% in the second quarter of 2026. That’s 32.5% for the top 1%, and the bottom 50% holds 2.3%. The 49% in between hold the rest.

The math on UBI

If robots do most of the work and most people have no job, UBI has to be everyone’s whole income, and paying for that would require taxation across the board. Take a full year of US gross domestic product, at today’s numbers, and tax all of it at 100%. FRED, using BEA data, has Q2 2026 GDP at $32.563 trillion at an annual rate. The Census Bureau’s July 2025 estimate is 341.8 million residents, of whom 269.8 million are adults. That works out to about $95,000 a year per resident, or about $120,700 per adult.

That’s a ceiling. GDP is total output, not a tax base, and the figure assumes the government spends nothing on anything else. At that point the government has become the employer of everybody.

Taxation at that level is communism on steroids, and I think we know it has never worked. Communism, and the way it came apart, bred some of the largest and most egregious oligarchies in the world. The Russian oligarchs came out of the 1990s sell-off of Soviet industry, and the loans-for-shares scheme is where that class took shape. Human nature is to hoard resources, especially when we think they’re scarce and useful. I don’t believe the billionaires at the top will willingly give up their wealth to live on $95,000 a year. I also don’t believe we have the money to fund UBI without something of that nature.

It would be a cataclysmic shift in how the government operates and how billionaires operate. And from where I sit, almost everything they’ve done for the last 50 years has been to pay less tax and hoard more wealth. ProPublica’s reporting on leaked IRS files found that the 25 richest Americans’ wealth rose by a collective $401 billion from 2014 to 2018, and the federal income tax they paid came to what ProPublica calls a true tax rate of 3.4% of that growth.

I think trickle-down economics has never worked. As far as I can see, it hasn’t raised wages, and all it has done is consolidate wealth at the top. A study of 18 OECD countries over five decades by David Hope and Julian Limberg found that major tax cuts for the rich raised the top 1% income share and had no significant effect on growth or unemployment. So I don’t think we pull off UBI when so many people would be displaced from their places of power and wealth if we tried.

What happens when you don’t need to work

Say it does happen. If you no longer have to work to feed your family, you no longer have to work to do pretty much anything. Entertainment is virtually free, or constant stimulation from whatever source. Medications and things like that are taken care of. Then what are the motivating factors that drive us?

There’s creativity and art. There’s having kids. But most people are working because they think they can overcome their circumstances and provide a better life for their family. If there’s no longer a purpose to what you’re doing other than to exist, what is motivation? It’s a philosophical question that’s really hard to answer, and I don’t have the answer.

The pilots we’ve seen don’t tell us much about it, because they only reduced the need to work. They didn’t eliminate it. OpenResearch gave 1,000 low-income people $1,000 a month for three years. The NBER paper found they cut their work by 1 to 2 hours a week, and their partners cut back by a comparable amount. Alaska’s oil dividend and Finland’s two-year trial didn’t reduce work at all (Jones and Marinescu; Finland).

Those are small payments that people still had to supplement with work. Eliminating the need to work raises a whole different set of questions. Do you just sit at home playing video games all day? Do you stream? How do you get social interaction, and what kind of interaction matters at that point? Do you even go to a store, and if you do, are all the stores run by robots? Or do some stores stay human-run because people want to do it?

Could it end up a utopia? Absolutely. Is it likely? Most of human history suggests it’s not, because it only takes a few bad actors at the very top to really screw it up for everybody.

What history says

When I said oligarchy on steroids is backed up by pretty much all of human history, I meant the French Revolution, the Russian Revolution, feudalism. Basically any time a ruling body amassed more wealth than the masses and kept the working class in terrible conditions. Walter Scheidel’s The Great Leveler argues that violence has been the main force behind declines in inequality.

What I’d do

I honestly don’t know what people should do. A lot of software will change over the next several years, and it’s going to change more rapidly than we could imagine. Any prediction I make about the next few years of software would probably be wrong.

Right now I’m trying to create and provide as much value as I possibly can, and embrace the paradigm we’re in. The only thing you can do is participate and try to help shape where things go. If we don’t use the tools at hand, one, it runs out of control, and two, other people are in charge of it as it runs amok. So my suggestion is to embrace it and build as much prosperity and wealth as you can now, while it’s still possible. Then hope the oligarchy doesn’t end up on steroids, and hope something like UBI does happen.

Maybe superintelligence yields a solution we didn’t think of. Maybe we collectively decide some jobs should be done by humans and never by AI. We could end up with something like the Butlerian Jihad from Dune, where humans built thinking machines, fought a war over it, and agreed never to do it again.

My point about the oligarchy is that we may have to go through a period where things get really dark before they get really bright and we potentially get that utopia. The US went through the Great Depression, FDR’s New Deal and a world war before it came out the other side into the postwar boom and the baby boom.

If you think I’m wrong, tell me in the thread on X. I’d like to be.