Nick Hanauer says Bernie Sanders's 50 percent AI ownership plan is common sense, not socialism

Nick Hanauer has found the one AI panic take designed to make LinkedIn libertarians spill cold brew on their quarter-zips: Bernie Sanders’s idea that the public should capture 50 percent of AI value is not socialism. It’s “common sense.”

On The Diary of a CEO, Nick Hanauer argued that the money being minted by artificial intelligence is not just clever founders and GPUs sweating in harmony. It is, in his telling, a private cash machine built on the unlicensed intellectual exhaust of everybody who ever wrote, posted, coded, photographed, explained, argued, translated, or accidentally made a useful Reddit comment.

the whole valuation that AI is predicated on is job disruption, right? You can’t you can’t get to those numbers unless you’re displacing lots of jobs.

That is the cleanest version of Hanauer’s case. If the AI boom is being sold to investors as a labor-replacement event, then don’t be shocked when people ask where the labor-replacement tax goes. You can’t tell Wall Street the robots will eat payroll and then tell everyone else, relax, this is just the calculator all over again.

The capitalist case for taking a slice

Hanauer is not some dorm-room revolutionary discovering Marx between vape charges. He’s a venture capitalist, which is what makes the claim land. His politics are redistributionist, sure, but his language here is ownership, licensing, and markets. Norway turned oil into a sovereign wealth fund. Hanauer wants AI’s common raw material, human-generated data, treated with the same kind of seriousness.

I don’t call that socialism.

I don’t know. Just common sense.

This is where the episode gets more useful than the usual AI content treadmill, which tends to oscillate between “learn prompts or die” and “my calendar app is now sentient.” Hanauer’s claim is structural. The question isn’t whether one junior marketer can do five people’s work with Claude. The question is who pockets the other four salaries.

data is a common good and it is a common asset that has been um sequested illegitimately by these companies.

Verdict: the diagnosis is stronger than the prescription. Hanauer is right that AI companies have benefited from a giant gray-zone harvest of public and private human output. He is also right that “innovation” sounds less magical when the business model depends on not hiring the entry-level person who used to do the work. But the 50 percent figure is more slogan than machinery. Fifty percent of what, exactly? Equity? Profits? Model revenue? A licensing fee on training data? The episode gestures at all of it and nails down none of it.

Daniel Priestley would like to see the manager

Daniel Priestley plays the necessary crank in the room, and not in the fake “just asking questions” way. His counterargument is brutally practical: even if you invent the perfect public mechanism for AI wealth, someone has to run it. That someone may be the government. Priestley does not hear “government” and picture Singaporean competence in a tailored suit. He pictures an HR portal from 2009 asking you to reset your password with a fax machine.

In the UK government, you are 10 times more likely to die than to be fired for poor performance.

It’s a spectacular stat, the kind that should come with a tiny Union Jack and a sigh. Priestley’s answer to AI disruption is not a bigger state balance sheet. It’s smaller companies, smaller teams, and more people using AI to start businesses that would have been impossible five years ago. He gives the example of a husband-and-wife video agency in northern England that used AI to build software, sign up 1,500 customers, and start hiring a team of 10.

That is the optimistic version of the AI economy: the two-person shop becomes the 10-person company. Hanauer’s darker version is that the 10-person company becomes one founder, a stack of agents, and a Stripe account. Both can be true, which is exactly why Hanauer’s ownership argument has teeth. AI can create small businesses and crush entry-level work at the same time. The machine does not have to pick a lane.

Steven Bartlett, who has built a whole content empire out of asking CEOs what they’re scared to say out loud, keeps returning to the entry-level problem. Call centers can shrink through attrition. Junior coders can feel useless watching agents write the code. Sales teams can become AI-fed closers rather than prospectors. The first rung of the ladder is starting to look less like a rung and more like a software subscription.

Hanauer’s best move is refusing the comforting fiction that this is just another productivity tool. If AI really is only a better calculator, then nobody needs a sovereign wealth fund. If it is a new ownership class built on everybody’s data and aimed at everybody’s job, then the political question arrives fast and wearing steel-toed boots.

The next politician, founder, or podcast prophet who wants to sell the AI future has to answer a simple question: if the models trained on everybody, why do only the shareholders get the dividend?