Elon Musk got into it with the sitting Vice President on X on June 20. JD Vance had floated something about the government taking ownership stakes in major AI companies, like a sovereign wealth fund does with strategic industries. Musk shot back. "Better just to send money directly to the people from the Treasury," he wrote. His math: AI and robotics will produce more goods and services than the money supply grows, so prices fall instead of rising. He's been building toward this for years now, and he's got a name for it. Universal high income. Not a floor. A very high floor, funded by taxing AI profits and mailed out as checks. He's called it the best way to deal with unemployment caused by AI.
The whole idea leans on one number, and that number is where the debate should actually be happening. Production has to grow at least as fast as the checks do. If it doesn't, the plan doesn't fail loudly. It just quietly turns into something else.
Here's what that looks like. Picture an island with a hundred people on it. Two doctors, three plumbers, five farmers, one electrician. Give every person on that island a hundred thousand dollars, deposited the same morning. Nothing about the island changed except the numbers in people's accounts. Still two doctors. The farmers didn't multiply overnight. The electrician is still one guy with two hands.
People start spending fast. They eat out more, finally renovate the kitchen, call the plumber for things that used to just wait. Demand spikes. Supply doesn't, because supply was never what changed. So prices climb. The plumber raises his rate, he's got more calls than hours in the day. Construction costs jump because the four people on the island who can actually frame a house are suddenly the four people every renovation is fighting over. Wages climb too, and at first that looks like the system working. The plumber who made eighty grand a year is pulling two hundred and fifty now. But past a certain point, services just stop being available at any price. The salon has a four month wait. A restaurant closes because nobody can staff a kitchen at the wage it would take to fill one. A windfall turns into a rationing problem, just rationed by price this time instead of by empty pockets.
Nobody here was being greedy or foolish. This is what happens when the claims on a thing grow and the thing itself doesn't. Money is not goods and services. Money is a claim on goods and services, and a check is just a bigger claim.
This is where the AI argument tries to close the gap, and it's a real argument, not a dodge. If robots can clean houses and drive trucks and handle a meaningful share of diagnosing patients, the island's actual output can grow to meet the new claims. A hundred people plus enough automation might produce what used to take ten thousand. In that world, the checks aren't just claims chasing scarce output anymore. They're claims chasing output that's finally big enough to support them.
So the whole thing comes down to whether that catch-up actually happens, and on the island's own terms it's not looking even. Widen the lens past the hundred people and out into a real economy, and the pattern holds. Automation is already coming for the accountant and the driver, the jobs built mostly out of patterns a model can learn. It is not coming, at least not yet, for the nurse or the electrician, the jobs that need a hand in the room or judgment nobody's written down. Those are the jobs immune to the productivity multiplier. They're also the jobs a high income check can't conjure more of. You can mail someone two hundred thousand dollars. You can't mail them a home health aide who doesn't exist yet, in a country that already doesn't have enough of them.
Which means even in the generous version of this future, there's a stretch of years where checks grow faster than the caregiving and skilled trade workforce does. That stretch, funded at national scale, is the actual policy problem. It's the part of this conversation that keeps getting skipped in favor of arguing over whether the top line number sounds inflationary.
Say it resolves anyway. Say production eventually catches up and the checks stop outrunning what the island can build and staff. There's still a second question sitting underneath the first one, and closing the gap on production doesn't close this one too.
Who decides what the island needs enough of.
Once the immediate scarcity clears, people don't automatically spend new money on the things that keep a place functional. Some will. Plenty will spend it on streaming and gambling and whatever feels good for an evening and leaves nothing behind, and that's not a moral failing. It's what happens when a species raised for two decades on engagement optimized feeds suddenly has more discretionary income than it's ever had. The market responds correctly, in its narrow sense. Wages for nursing and teaching and sanitation work rise, because those jobs go short of people. But a market correctly pricing a shortage is not the same as a market solving it. Nothing guarantees the corrected wage pulls enough people into caregiving fast enough to matter to the people who need care right now.
Every society has nudged this outcome somehow. Scholarships that steer people toward medicine. Loan forgiveness for teachers. Tax credits for families. None of that is new, and none of it is sinister. It's an admission that markets optimize for what people want in a given moment, and what people want in a given moment isn't reliably what a place needs sustained over years. Wants and needs pull apart, especially for anyone who's spent the last twenty years inside a feed built to blur the line between them on purpose.
That's the real shape of the gap in universal high income, once you follow it all the way down. The production math might work, under the right automation curve. But the proposal only ever answers the money question. It says nothing about whether the people getting the money can tell the difference between what they're drawn to in a given hour and what would actually leave them better off a year from now. The systems currently best at answering that question, the recommendation engines, the casino floors, the infinite feeds, are optimizing for attention, not for flourishing. A check doesn't touch that layer. It just hands that layer a bigger budget.
I don't think the fix is a government committee deciding how everyone spends their money, and I don't think a big enough check solves it alone either. Both treat this as one distribution question when it's actually two, stacked. Can the island produce enough. And do the people on the island have any real way of knowing what they need before they spend. Universal high income, at its most generous, might eventually answer the first one. It was never built to answer the second, and right now almost nobody arguing for it is even admitting the second one exists.