Why Would This Time Be Different?

In 1930, John Maynard Keynes made a promise. Within a hundred years, he wrote, technology would make us so productive that the problem of work would be nearly solved. Living standards would be four to eight times higher. We would need "three-hour shifts or a fifteen-hour week" just to stay busy.

The hundred years are up in 2030.

The productivity came. The leisure never did. As David Graeber put it, "technology has been marshaled, if anything, to figure out ways to make us all work more."

Keynes saw the cost coming. In that same essay he coined a term for it: "technological unemployment." He treated it as a passing phase on the road to abundance.

Now read Sam Altman, writing in June 2025 about the 2030s:

"There will be very hard parts like whole classes of jobs going away, but on the other hand the world will be getting so much richer so quickly that we'll be able to seriously entertain new policy ideas we never could before."

Same promise. Same decade. Same order of operations: the jobs go now, the sharing comes later. And when the old jobs go, Altman says, new ones will appear. He even predicts that the jobs of a thousand years from now will look "very fake" to us, and says that's fine. Graeber had a name for work like that: bullshit jobs.

We have heard this before. Why would this time be different?

Two promises

We are being offered two promises.

The first is abundance. Dario Amodei, CEO of Anthropic, writes that AI could give us "the next 50-100 years of biological progress in 5-10 years." Altman says "intelligence too cheap to meter is well within grasp." That phrase has a history. In 1954, the chairman of the Atomic Energy Commission predicted that the atomic age would bring electricity "too cheap to meter." We still get the bill.

The second is generosity. Nan Ransohoff predicts a "third wave" of American philanthropy, with AI fortunes adding $37 to $100 billion a year in new giving. Her models are Carnegie and Rockefeller.

Both promises skip the hard part: who gets the gains. Amodei concedes that eventually "our current economic setup will no longer make sense, and there will be a need for a broader societal conversation about how the economy should be organized." He doesn't say who gets to have that conversation.

History does.

What happened last time

Workers' share of what the U.S. economy produces is now 52.8 percent. That is the lowest ever recorded. The Bureau of Labor Statistics published the number this month. For most of the postwar era, it held between 60 and 66 percent.

For thirty years after World War II, pay rose with productivity. From 1948 to 1979, productivity grew 112 percent and typical worker pay grew 90 percent. Then they split. From 1979 to 2019, productivity grew 60 percent. Typical pay grew 16.

The difference didn't disappear. It went up. RAND estimates that in 2018 alone, the bottom 90 percent of Americans would have earned $2.5 trillion more if incomes had kept growing the way they did after the war.

The gains didn't come back as time, either. The best-paid now work the longest hours. At the bottom it's the opposite: a third of retail and food service workers can't get the hours they need, and two-thirds get their schedules less than two weeks ahead. Nobody got the fifteen-hour week. Some got fifty. Others can't get enough, and don't know from one week to the next which hours they'll get.

That is how the last wave of technology was shared.

Who owns it

OpenAI was founded as a nonprofit to benefit humanity. After its 2025 restructuring, the nonprofit owns 26 percent of the company. Microsoft owns 27 percent.

Read that again. Microsoft owns more of OpenAI than the nonprofit that is supposed to control it. Employees and investors hold the rest. Altman says his industry is "building a brain for the world." Three-quarters of his company is privately owned.

The owners are about to lock that in. Anthropic and OpenAI both filed confidentially in June to go public. Anthropic is reportedly aiming to list as soon as October.

Look at what the technology is being used for first. For five straight months this year, AI has been the leading reason U.S. employers gave for layoffs. More than 112,000 job cuts through July were blamed on it.

I spent more than a decade at Facebook. I saw what incentives for power and revenue do to a company's stated goals. Altman himself uses social media feeds as his example of misaligned AI: systems that are "incredible at getting you to keep scrolling" by exploiting "something in your brain that overrides your long-term preference." The same forces are now at work in AI. In February, OpenAI began testing ads inside ChatGPT conversations. By August, the ads had spread to eight more countries.

And look at what it's made of. These models were trained on the writing, speech, art, and code of billions of people. None of us were asked. Almost none of us were paid. The largest payment so far went to the authors of about 400,000 pirated books, $3,000 each, in a $1.5 billion settlement with Anthropic. Courts have said training itself can be legal. The law is behind. The moral question isn't close. It was taken from all of us.

The philanthropy promise

When John D. Rockefeller asked Congress to charter his foundation, Congress wouldn't do it. Theodore Roosevelt said "no amount of charity in spending such fortunes can compensate in any way for the misconduct in acquiring them." Samuel Gompers suggested the one endowment the world would gladly accept from Rockefeller: one "to help other people see in time how they can keep from being like him."

That is the model we're being offered.

The Giving Pledge turned fifteen last year. The original U.S. signers who are still billionaires have grown their fortunes 283 percent since 2010. Mark Zuckerberg and Priscilla Chan grew theirs more than 4,000 percent. One living couple has fulfilled the pledge. Eighty percent of what pledgers did give went into private foundations.

And foundations grow faster than they give. In 2024, U.S. foundations gave away a record $100 billion. Their assets grew nearly 12 percent, to $1.64 trillion. The median foundation pays out 5 percent a year, the legal minimum, and has for five years running.

Gates says his foundation will spend $200 billion and close by 2045. When he announced that in May 2025, its endowment was about $77 billion. By this June it was $92 billion. The promise is growing faster than the giving. Chuck Feeney actually gave away everything, which proves it can be done. For almost everyone else, the pledge stays a pledge.

Then ask who governs the money. Take the Ford Foundation, among the most progressive in the country. Its board is chaired by the founder of a private equity firm who sits on the boards of Uber and TSMC. Another trustee chairs the private equity firm Warburg Pincus.

That isn't public governance. It's the powerful governing the money that was supposed to repair what power broke.

From passive recipients to active directors

Our relationship with technology is broken. The owners build, and we wait to see what they'll share. Keynes's grandchildren waited. Workers waited. Now we're being asked to wait again.

Amodei is right that we need a broader conversation about how the economy should be organized. We shouldn't wait to be invited to it. The window is now, before the IPOs lock in ownership for a generation.

Here is what we should demand.

Safety as the price of going public. No IPO without independent testing before deployment, independent audits after, and real financial and executive liability for harm. Investors, underwriters, and exchanges can make this a condition. The largest asset owners can refuse to buy without it.

Benefits for people, not just margins. Companies that replace workers with AI should disclose it in their filings, and part of the savings should fund those workers' transitions. Public contracts and subsidies should go to companies that create good jobs and raise wages, not busywork, and not cuts.

Shared gains, globally. This technology was built from all of us, so all of us should own a piece of it. Start with a global tax on extreme wealth. The economist Gabriel Zucman has already designed one: a 2 percent minimum tax on the world's roughly 3,000 billionaires, raising $200 to $250 billion a year. Then build a global AI wealth fund. Alaska pays every resident a dividend from its oil. Norway holds its oil wealth for its citizens. Our words are the oil.

Foundations that answer to the public. If AI fortunes become foundations, those foundations should be governed by the people they claim to serve, not by private equity chairmen.

In 1930, Keynes wrote that greed would have to be our god "a little longer still." The hundred years are nearly up.

The technology was never the constraint. Ownership and governance are. This time won't be different because they promise it will be. It will be different only if we demand it, and write those demands into the rules that govern the money: the offering documents, the tax codes, the foundation charters.