A blank wooden reserve sign at the edge of a meadow beside an unused gravel lot, late-afternoon light, no people.

Gates wants some jobs reserved for humans, and a tax on AI tokens. That’s the essay.

Bill Gates published a long essay today. I am not walking you through all of it. The piece is here. Two proposals are the ones that survive a second read.

He wants some work set aside for people. He calls that domain Human Reserved. And he wants a tax on AI tokens and robots, because the current tax code already pays you to replace a worker with a machine.

That is not a conversion story. The Verge framed him as newly bleak. He did not call himself deeply pessimistic. In the essay he still says the same thing he has been saying: if the world takes the right steps, “AI will be a force for good and leave everyone better off.” He also says, in the same breath, that “right now, we are not preparing adequately for that transition.”

Human Reserved

Gates’s father died of Alzheimer’s in 2020. Paid caregivers understood him when he could not say he was hungry. “Something in the care they gave my dad was irreplaceably human,” he writes. “No robot could or should have done it.”

From that he draws a rule, not a eulogy. “I believe that as AI and robots improve, we’ll set aside certain things for only people to do. I’ve started calling this domain Human Reserved.” He likes the phrase because it sounds like a nature reserve: land you could pave, and choose not to, “because the loss would be too great.”

Sometimes the reason is economic. A lot of people cannot just switch careers. “You can’t tell a 55-year-old who has worked in construction their whole career that they need to go work at an elder care facility and expect them to find it fulfilling.” Sometimes the reason is that the machine should not be the one in the room. He asks you to imagine a robot delivering an incurable diagnosis. “There’s no technical reason why it couldn’t. Yet it shouldn’t.”

He does not hand you a statute. He lists the questions he does not have answers for: who decides, by what criteria, how you stop firms from cheating, what happens to trade when one country lets robots do the job and another refuses. Education and mental health, he says, will be a mix, with a human in charge. Japan may want a caregiving robot. Another country may insist on a person. I am not going to invent the missing mechanics. He didn’t.

Tax the tokens and the robots

The second idea is older, and he knows it sounds strange. “I believe we should tax AI tokens and robots.”

His argument is the imbalance already in the code. Hire a person, pay payroll tax. Buy a robot, write it off. “The tax system nudges you toward replacing people with machines.” A tax, he says, would “slow the rush away from human labor a little and raise money for retraining and a stronger safety net.” It would have to be targeted so it does not tax the uses he still loves, “like making medicine and education cheaper.”

A robot gripper on one side of a workbench and a paper payroll stub with coins on the other, no people.
Hire a person, pay payroll tax. Buy a machine, write it off. That is the nudge he wants flipped.

He proposed a robot tax years ago. Most people treated it as a gag. “I’m still a big proponent of it. Although it is not the whole solution to the threat of AI, it is part of a wise response.” He does not name a rate. He does not name a collector. Anyone who writes those in is padding.

He still thinks a plan can make the good outweigh the bad

The rest of the essay is the weather around those two bets. Jobs will vanish and not come back with the next cycle. “Many jobs will disappear forever.” Entry-level work gets hit first. Robots, he thinks, start competing on some physical tasks in construction and hospitality by the end of the decade. If someone had “a credible plan for slowing down AI advances globally,” he “would likely support it.” He does not think that plan exists. Incentives run the other way.

He also writes the other half, the part a doomer recap drops: hospitals that lack a specialist, farmers who finally get a forecast, a student who is allowed to struggle before the answer shows up. “Maximizing the benefits is just as important as minimizing the harms.” The line he keeps returning to is not despair. It is the missing plan. “There is no plan to ease the entry into the AI era.”

So here is the short version I would send a minister in Mexico City, or anyone else who will not read 6,000 words. Gates is not asking you to halt the labs. He is asking you to fence off some work as human on purpose, and to stop giving a tax break to the replacement. Those are the two choices in the title that are actually choices. The rest is him telling you the transition will be ugly if you wait for the unemployment print.