Top White House Advisor: The US Empire Is Declining And Socialism Is Coming Next! | David Friedberg

It's inevitable that socialism will happen in America, but I don't want to lose my agency, my liberty, my freedom.

"Do you think the US as an empire is in decline?" The answer is yes. I was appointed to President Trump's Council of Advisers on Science and Technology, and one of the great lies is that the government will save you. But the more government has tried to create services for people and help bring them up the ladder, the more expensive those things get.

63% of Americans live paycheck to paycheck. People can't afford to pay their bills, and it needs to be fixed.

What will ultimately happen? All my politician friends are about to chop my head off about this, but—

David Freedberg, what is the most important subject of our time?

AI. We're entering this golden age, and people don't realize we're in it. I don't buy into the narrative that Dario, Elon, and Sam end up having all the value in AI. In the next 10 years, there will be a billionaire who emerges with zero net worth today, who downloaded open-sourced AI and built a company that made them a billionaire. They'll be someone who comes from nowhere, and there will be a lot of stories like that.

But this narrative that AI will take jobs away—you don't believe it?

From data, from evidence across all periods of technological revolution like this, we've never seen jobs decline. The average person doesn't see it, but oh my god, the stuff we can do is unbelievable. Right now we have a billion people in the world starving and hundreds of millions dying from curable disease. These are the sorts of things that are going to get resolved, because AI unleashes this capacity for human advancement and it's transforming every aspect of biology and our understanding of the life sciences.

So I wanted to ask you—do you think we're going to be able to live forever soon? The truth is—

This is super interesting to me: my team gave me a report showing how many of you who watch this show subscribe. Some of you have told us you randomly unsubscribed from the channel, so I have a favor to ask. If you're a regular viewer and you like what we do here, please check right now that you've hit the subscribe button. We're approaching a significant subscriber landmark on this show, so if there's one simple, free thing you can do to help us keep this show free and improving year over year and week over week, it's to hit that subscribe button and double-check you've hit it. It's the only thing I'll ever ask of you. Do we have a deal? If you do it, I'll make sure that every single week and every single month we fight harder and harder to bring you the guests and conversations you want to hear. I've stayed true to that promise since the very beginning of the show, and I will not let you down. Please help us—really appreciate it. Let's get on with the show.

Asked how he would label himself, David Friedberg says he is deeply curious about the world and the universe—how it works and why it works the way it does—and a big believer in human agency to affect the universe around us. That is his core motivation, and much of the rest of his work is about how to enable it.

From Astrophysics to Silicon Valley

Describing his professional journey, he explains that he started college majoring in math and physics, then switched to astrophysics. While he was in college, the dot-com boom hit the heart of Silicon Valley. A kid in his dorm started a website selling DVDs online and sold it for a million dollars in their junior year—an unbelievable, unfathomable amount of money at the time. Friedberg got enamored with the idea of using businesses to change the world, not just scientific discovery, and decided to go work in Silicon Valley.

He didn't get many job offers, but poker helped. He had made money playing poker during college—after a year working in a pool hall for $4.25 an hour, where the manager and a bookie took all his money in a home game, he decided to learn the game properly, reading all the strategy books. He put poker on his resume, and investment banks wanted to interview a kid who knew how to play—this was before internet poker and ESPN poker, so it was an interesting novelty. He landed a job in investment banking working on tech companies, where he learned business, finance, accounting, and mergers and acquisitions.

He then joined Google when it was just under a thousand employees, left a couple of years later, and started a business called the Climate Corporation. He has spent his whole career in Silicon Valley.

The Climate Corporation

The company was sold in 2013 for a little over a billion dollars—a great outcome for investors and shareholders. The software, built for farmers, is now owned by Bayer, the large pharma and agriculture company, and is used across 200 million acres of farmland every year. As he puts it, it has done fairly well.

The interviewer notes that Friedberg was 33 when he sold the company for $1.1 billion—the same age the interviewer is now.

Since selling the company, the speaker has taken on other roles — including, as of March this year, an appointment as an advisor to Donald Trump. He was appointed to President Trump's Council of Advisers on Science and Technology.

Asked what that means in reality, he explains that the composition of PCAST — the President's Council of Advisers on Science and Technology — has reflected different themes across administrations. Bill Clinton, for example, set up a PCAST focused specifically on the internet. The current council under this administration is heavy on AI: its members include people like Mark Zuckerberg, Marc Andreessen, and Sergey Brin. It has always also included people involved in biotechnology and the sciences.

The council's role is to help guide science policy for the administration. That means conversations with people in the administration about what is happening in industry and on the frontiers of science and technology, in order to inform the policy decisions the administration is trying to create or enforce.

Asked what the most important subject of our time is, the answer was AI. Right now, the big question in the United States from a policy perspective is the regulation of AI: what role the government should play in determining where AI goes, how it gets built and deployed, who gets to use it, who owns it, and where the value lies.

Factions have already emerged inside the government—even within the same political party—over questions like whether to allow Chinese open-source models to be used by American businesses. These Chinese models are as good as, and in many cases better than, private American models. Because they are open weights, you can download them and run them on any computer you want, and many companies want to use them because they are cheap: effectively 50 cents per million tokens of output—a token being roughly like a word—versus $50 from Anthropic, which is a private model.

The question is whether to allow these Chinese models to become ubiquitous in industry, and if so, what that does to American AI and the country's capacity to compete, and whether there is a security risk. A lot of this debate is underway, and while it may not be the hot topic tomorrow, today it is the big thing.

From a science and technology policy perspective, this administration will probably be viewed through two lenses: on one hand, the enablement, acceleration, and regulation of AI; on the other, the still-open question of what science is in an era of AI, since science itself is changing in a big shift underway.

The interviewer noted they would go into those subjects specifically—AI and everything it interfaces with—but argued it is important first to get a backdrop of the economic and social state of the nation, because all of these things feed into the big questions on AI and inform everyone's biases: if you are at a different socioeconomic position or feeling something different, your perspective on AI will be radically different. So, focusing first on the United States—which is in some respects a proxy for the rest of the Western world—where is...

Where is the United States right now—is inequality rising, and what factors help us understand the backdrop of the nation?

The biggest drivers shaping what public policy and private markets will do are, first, the unaffordability crisis. 63% of Americans live paycheck to paycheck—they don't have enough savings to cover a fiscal emergency in their household. Costs are rising faster than wages, so things feel less affordable to most Americans.

Then there's wealth inequality, which is partly a perception issue and partly the result of policy failures. America is in a crisis of classism, an "us versus them" divide that manifests in many ways. You had Ray Dalio on, and if you think about his cycle of empires—the idea that there have been six empires in the last 500 years, with the United States in decline and China rising—it can be measured across several dimensions. One is this internal struggle, which often manifests as external struggle, like the Iran war. Why can't we stop having external wars? A nation that's happy with itself generally doesn't go to war; a nation that's unhappy generally does. A lot of this ties into the economic distress the United States is feeling.

My contention—which may not be popular—is that the more government has tried to create services and help people up the ladder, the more expensive those things get.

This is one of the great lies of U.S. policy over the last half century: that the government will save you by providing more of the essential services you want—education, healthcare, housing. But there are people offering those services and someone paying for them, so as government enables more money to flow into those markets, costs have simply gone up.

Education shows how this plays out. Thirty years ago, administrative staff were about 10% of colleges; today they're about 60%. There's no rational reason for six times the administrative staffing. The real reason is that the federal student loan program doesn't discriminate based on how well a university does with its graduating students, the degree, or the individual borrower—it just lets more money flow into universities. University administrators face no constraint on tuition, so they raise it 5%, 8%, 10% a year and hire more people, who then earn higher salaries as they climb from manager to director to senior director. That's natural for any organization—nonprofit, university, or private company: people want to grow and earn more year after year. Because universities can collect more money, they charge more tuition, and there's no market force saying tuition is too high, since the government provides unlimited funding through federal student loans.

As a result, education has become so expensive that everyone says something must be done—and both parties respond with more government intervention, which is like putting fuel on the fire. The challenge now is that people feel left behind, wealth inequality is massive, and things are extremely unaffordable—yet the embraced solution is for government to do more, which in effect leads to socialism.

What's wrong with socialism? I would argue that the key measure of a country's success is how many people transition from labor to capital each year.

It means this: every year, I'm working paycheck to paycheck, collecting a check. When I graduated from college, I had $19,000 in debt. I went to work, got paid, saved up, and paid down my debt until I got out of it. Then I worked at Google, we had an IPO, and suddenly I had assets—money in the bank account. It was an incredible moment for me. I realized I could live off that money. I didn't need to live off a paycheck every month. I had savings, I had assets. That is capital.

Everyone talks about capital versus labor, but the real advantage of America is enabling people to go from labor to capital. Every individual who works, saves, and invests—whatever they need to do to accumulate enough capital—ends up in the capital bucket. Then you're no longer stuck going paycheck to paycheck to pay your bills. That's the dream. That's the real American dream. It's not just about owning a house. If you own a house but carry a bunch of debt, have insurance payments to make, and still have to work every month to cover them, you're not necessarily living the American dream yet. The American dream is being freed of that burden.

>> Because you can live off the interest of your assets.

>> That's right. Say you have $100,000 saved and you can make 10% a year on it—that's $10,000. If you get to a million dollars saved, you can make $100,000 a year and live on that. That's economic security, financial security. What we should give every American is the ability to own assets that give them the freedom to choose what they want to do in life, rather than be stuck in work.

The problem with socialism is that it takes away that transition moment. It eliminates it, because it says everyone has to be labor all the time for the rest of their lives—and then no one has freedom. That's what America was founded on: the ability, through freedom of choice and agency, to transition from labor to capital. I think our goal as a country should be to get 2% of Americans across that line every year; then we'll be successful going forward. But if everything is about giving people more stuff and having the government do it, we'll just keep raising the debt, making things more expensive, and inflating the cost away—and it's inevitable that socialism will happen in America. It's absolutely inevitable.

Asked whether the US as an empire is in decline — in terms of the power of the dollar and the prosperity of Americans — the answer for the average American, or the majority of Americans, is yes. But the potential hasn't been deleted; decline is not inevitable. We are not yet at the point where it's a runaway train.

In the history of democracies, there has never been a place like America, where someone like him — moving from South Africa, attending public college, graduating with debt — could work his way out of that debt and reach his current position with no nepotism and no handouts. You could argue about privilege and other advantages he had over his life, but it's an amazing place where that story can be told a million times over — not just one or two people, but millions. The problem is the hundreds of millions who never got to live that opportunity. If that can be fixed, America is not in decline; it's a prosperous nation with centuries ahead of it.

When you look at the national debt, though, it does look very much like a runaway train.

That's where inflation comes from: printing more dollars to pay for the debt. The Federal Reserve, which is separate from the government, buys the debt and issues dollars to the banking system to buy it from the government. It can send dollars to banks, the banks loan money out, and it makes its way into the economy — now there are more dollars in it. That money was owed back to the government at some point, but the Federal Reserve may never get paid back; it may just keep adding and adding, printing more and more dollars.

Is that sustainable? No.

What force will intervene to stop this compounding debt? The outcome is socialism — that's the moment we're in, and that's why we're seeing this wave of it. He had been saying for five years that America would face a wave of socialism, and predicted after Donald Trump was elected that the next wave in American politics would be socialism. People thought he was crazy, but it's absolutely inevitable. It is de facto socialism when the government employs roughly half the people of the United States: counting federal, state, and local agency employees, contractors of those agencies, and people living off a government pension or Social Security, the total is close to 50% of Americans. People are reaching the point where the only place they can go is more money from the government, and eventually the government becomes the whole economy.

If socialism was the outcome, what was the cause — capitalism?

Capitalism plus policies. Say you have a million dollars making 10% a year: next year you have $1.1 million, then $1.21 million, and so on. That's compounding — after a couple of years, your million turns into $10 million. And you never have to sell shares along the way; you just own them as the value keeps going up, never paying taxes along the way. Meanwhile, if you're working, taxes are taken out of every paycheck, every time you get more money. So if you don't get enough capital or savings to compound, you'll never get there. From a tax policy perspective, labor — working for your money — should never have been taxed higher than earning returns on your capital.

So you're saying tax the rich? No — tax capital at a rate equivalent to labor. Wealth taxes.

Wealth taxes are asset seizures — definitely no. One of the founding principles of the United States is private property rights. Go back to Europe 500 years ago: there were very few places where an individual citizen had the right to own private property, where you could say "I have this thing and it doesn't belong to the king or the local lord." Everything was the property of the upper class, vested in them by birthright for centuries. The same was true in Asia, the Middle East, and Africa — every nation started with a system where people had no rights to private property, and those above them decided who got what.

The United States was founded on the principle of private property for the individual. When the Declaration of Independence was drafted, it was reportedly about life, liberty, and the pursuit of property, later changed to the pursuit of happiness. The pursuit of property was really the measure of what you've accumulated as an individual: the right to build value, build wealth, build a house and have your family in it. That was a founding principle.

I'm all for taxes when people transact. When you turn stock you've held for a couple of years into cash and buy something with it — boom, you get taxed on that. That's the right model: tax the moment you turn an asset you own into another asset. That's what we do today, as an income tax or capital gains tax when you realize the gain.

But if the government goes into your shares and takes 5% of them, what are you losing? You can't use those shares to buy anything, because as soon as you try, you owe tax on them. Taking away your shares before you've transacted with them — or taking your home, or giving the government the right to come into your garage and audit everything you own and impose a 2% tax this year — is antithetical to what the United States was founded on.

If your property goes up in value and you eventually turn it into other property, that's a transaction, and that's when you get taxed. So if we have an issue with people having too much wealth, that's the moment to raise the rate: capital gains tax and income tax on high earners. Those two things can get you 97% of the way there, and a wealth tax won't really gain much more. Tax the rich — all for it — but tax them when they transact, when they sell, when they have a capital gain.

Borrowing against assets is another flaw in the current system. If I'm very wealthy, own a bunch of stock, and have never sold any of it, I can borrow against that stock to buy a yacht and pay no taxes. That should be a taxable event, and it's an easy fix: I'm not taking away your private property, but as soon as you borrow against it, it's effectively like transacting on it — a moment to pay tax. If we think the rich aren't paying enough, raise income or capital gains taxes.

Most people who own stock in big companies just borrow against it and never pay a tax.

I didn't know this until I was about 26. My company had gone public, and a European bank approached me and said they'd lend me 20% of the value of my stock in the company — a tax-free loan. I said, "Explain this to me. You'll just send me millions of pounds and I don't have to pay tax?" They said I wouldn't have to sell my shares. I told them to look at the rest of my portfolio, which had some Facebook stock at the time. They offered a 50–60% loan-to-value: if I had a million dollars of Facebook stock, they'd transfer me 700 grand today, tax-free, to spend however I wanted. I thought, "This is what the rich do."

That's exactly right, and the truth is that no one talks about this behind closed doors. Everyone knows it needs to change, but no one wants to say it.

What's interesting is that if I have 10 million of Facebook stock and I take a loan, they give me 5 million in cash. What a rich person then does is take that 5 million and go buy other assets, and then they start acquiring more and more assets. That's what's unfair about the system: once you have capital, you're accumulating more capital, while labor—if you haven't made that crossing from labor to capital—you're still grinding every month, every week, trying to pay your bills, while these guys have all these shenanigans going on.

It is wrong and it needs to be fixed. But the answer is not taking private property. As soon as you start taking private property, those gates are open. By the way, it's not even going to be constitutional in the United States—the Fifth Amendment has this takings clause, so it's legally not going to win in federal court. It could work in some states: seven states actually have a constitutional ban on a private property tax like this, while in the other 43 states it's game on, and we'll see how it plays out in the courts as they try to pass these wealth taxes. But these are asset seizure taxes.

You can have a wealth tax by raising capital gains, or making high-income or high-asset people pay more when they sell or buy things. There are a lot of ways to tax wealth. But taking assets from people, whether after they've paid their taxes or before, is not the right way to do it.

The ultimate outcome will be that 51% of people say, "Give me all the stuff that the 49% have." What would stop them? You can take as much as you want from a minority and distribute it to the majority—and that minority starts with a few billionaires, then a couple more 50 millionaires, then 10 millionaires, then millionaires, then people with 100,000. There's no limit to where this goes. And that is socialism.

Asked how this affects where talent moves, the speaker points to real examples: "I know billionaires. Many of them have moved or are planning to move out of California—just on the threat of this." Some respond, "Good riddance," but the top 1% of earners pay 40 to 50% of California's total personal income tax, depending on stock market capital gains performance. Those are the people leaving the state because they see where this is headed, and California probably doesn't want that. They are productive people who create companies, hire workers at good wages, and support the economy—not just through taxes but as a big part of the capital flowing through it.

France offers a well-studied case: it had a wealth tax, and total tax revenue declined by more than the wealth tax brought in because everyone left. When France stopped the wealth tax, people came back and the income came back.

So neither a wealth tax nor an asset seizure tax is the solution. Instead, we have to fix the other problems discussed earlier and fix the government spending problem. If we do those two things, we can bring inflation down, reduce wealth inequality, and give more people the ability to make the transition from labor to capital.

Every politician should be talking about what percentage of Americans each year move from labor to capital—how many people can say, "I've got enough cash to retire if I wanted to." That's the goal we should push everyone toward. One of the great lies is convincing everyone that the American dream is owning a house. For many Americans, a house means putting all your capital into one asset, and the whole system depends on housing prices rising every year to keep people moving up the ladder. Fast forward 30, 40, 50 years, and young people can't afford to buy a house anymore because asset values have been pushed up so much to keep existing homeowners ahead.

We're told that the central idea once you leave university is to figure out how quickly you can get a mortgage or buy a house.

It's one of the great lies. In many countries, people have been able to accumulate capital and jump the ravine from labor to capital much more quickly when they don't buy a house, because they can invest instead. You can buy the S&P 500 on an E*Trade or Robinhood account, and on average it will make you 10–11% a year. Put your money in that and you're making 10–11% a year, it can go up every year, and you don't have to pay taxes, property tax, or insurance, or deal with repairs and maintenance. So a lot of people would have been better off over the last 30 years actually owning the S&P 500 than owning their home and paying rent on a home — depending on the market, since some markets ballooned.

One thing I see in the comment section whenever someone makes that point: someone who bought a house 20 years ago says, "Well, I bought a house 20 years ago for this amount, and now it's worth this." Every single time this argument is made by financial advisors or investors, that is always the top comment — "I bought one and I did good."

That actually indicates a problem: if your house went up in price so much, that house is much less affordable to the young person buying it next. That's what we've fundamentally done with residential real estate — created a system where prices go up so much, with a whole bunch of policies designed to make that happen.

There's a graph here showing all the things that have gone up, and housing is one of them. The stuff on the bottom is what the government doesn't touch; the stuff on the top is what the government touches. The more the government is involved, the more expensive these things get.

But all the policy around residential real estate has allowed the American middle class — which holds the majority of the wealth, and I'll give you the statistics in a minute — to grow their asset base and become wealthier. All the boomers are so wealthy because they owned these houses. And now young people graduating college can't afford a house. It's a bad fact that a lot of people got very wealthy by owning a home, because it means the next generation is going to say, "I want socialism because I can't afford a home." That's the moment we're in.

45 million Americans graduated from college in the last 10 years. Do the math on that: it's a large number of people who are now, on average, encumbered with some amount of student debt and facing a challenging job market. And then the only way to buy a home is to come up with a million bucks. It leads to this problem where everyone's like, "Man, I need mom and daddy."

So what would you say to the 63% of Americans — the statistic changes year to year, but that's roughly the number — who have less than $500 in savings, are working paycheck to paycheck, and can't pay for a fiscal emergency?

For the 63% of Americans living paycheck to paycheck, where buying a house is unaffordable and potentially not a good investment anyway, what's the prescription for building assets?

People will challenge this, but I would have never thought podcasting would become such a massive business, or being an Instagram or TikTok influencer, or that individual artists could publish online and make money without a record deal, sell their work on eBay and Etsy, or that craftspeople could find jobs through the internet. The internet has been an incredible enabler of millions of new ways to work and earn an income.

A lot of people roll their eyes and say, "But that's not what the system was supposed to set me up for. The system failed us." Maybe the system did fail you, and maybe the government's solution wasn't the right one, but human agency is what got all these people there. Did the government give you your job as a podcaster, or teach you how to build this media empire? I don't know if the government played much of a role in it — I think you did.

This is important for people to observe: the government isn't going to save individuals. In fact, it often makes things harder. If I want to start a barbershop, I need cosmetology licenses, thousands of dollars for school, health inspections — a lot of government layers that stop me from doing that work. The government inhibits people's agency and makes it harder for them to progress.

The second thing is I never discount human agency. There was a great podcast years ago by the guys at the Hoover Institute about China's success. China grew GDP per capita from around $3,000 to $30,000, and many people said, "Chinese central planning is brilliant." These guys argued — which I discounted at the time but often think about now — that it was actually the agency, entrepreneurship, drive, and motivation of individuals, not central planning. The government enabled free markets: it allowed people to work, trade, and create value for each other. People looked around, said "I think you want this," and started making it. When the government loosened the shackles on individuals, that population grew the economy — even though central planning still plays a big role in China.

Think about all the success that's come from the United States: it's not the government announcing a central plan. It's individuals making their own choices — seeing what people want to learn or watch, creating that value, delivering it, and earning something people are willing to pay for.

And I don't think we teach this in schools. We've lost the value of agency, of the individual taking their own action. Every conversation is about the government — government this, do this, do this. For me, what individuals can do is limitless.

I tell my kids this: every day, ask yourself, what did you do today that someone else valued? If you focus on figuring out something you can do that someone else values, then do it over and over again, rinse and repeat. I think there are paths.

Let me offer a couple of counter-arguments so we can stress-test this thinking. You pointed at me and said I started podcasts, started businesses, and those businesses grew—and you were generous in giving me more credit than I deserve. But if I were to rebut that, I'd say: I was born in Africa. Would that have been possible in Nigeria, where my mother's from, or in Botswana, where I was born? Is it the fact that I moved to the UK, where I had this foundation of stability because of the government—school was free for us, healthcare was taken care of—so I never had to think about those things, which allowed me at 18 years old to not go to university, take a risk, drop out, and be in that room?

When I was sitting in that room in Manchester, I actually at one point printed off the Jobseeker's Allowance forms—basically, the government gives you free money if you reach the point where shoplifting pizzas is no longer a reliable way to feed yourself. I had that form on my desk. There's maybe something psychological about knowing I had this underlying safety net: I could make a 300-mile round trip back to my parents in Plymouth, live in the house, or fill out that form and they would pay me. That allowed me, maybe at some deeper level, to take the risk. I never even realized the privilege of it—there's healthcare, there's laws, there's roads, there's this Wi-Fi thing that let me build the website I built with a laptop in my room. So that's the government's involvement in my success: they gave me this great foundation of stability and infrastructure.

That's a great role for government to play. But the problem is that's a slippery slope, right? There's balance here—nuance between personal agency and the government giving you a good foundation. And the government shouldn't employ you. That's the slippery slope we end up on, and that's where I worry about our ability to come back from it. If you end up with 40 or 50% of Americans getting paid some check from the government, it becomes very hard to get off that check, and then people vote for that check to get bigger and bigger.

The other counterpoint is that there's a privilege we both have in terms of just our nature—nature, nurture, whatever it might be. I said nature because upbringing implies our parents were great to us and that's what made us whatever we are, but the opposite can be true: think of Elon Musk—his dad wasn't that nice to him, so not necessarily a lovely upbringing, but the circumstances were there. He was never without a meal, and neither was I. Those are important facts of privilege. The circumstances of our upbringing, and maybe our nature, maybe our biology, may have set us up to be a certain type of person for whom agency—however that's defined—comes naturally.

I think you're exactly right, and that's what education should be teaching—that value. That's where I think the question of what's the role of a public school, what's the role of this infrastructure government builds around us, comes in. If it's teaching you that your next job is to come work for the city or the state, it's not teaching you that you have agency. If it's telling you that the only path out of this rural school is to go work in the military, it's not teaching you about your agency. I'm not saying it's wrong to do civil service or military service—those are important parts of what all of us should strive to do, supporting the government that makes our society possible. But the idea that we should all be proposed the government role, the government path, the government work as what we should all be stuck doing for the rest of our lives is the wrong setup.

The opposite is also probably the wrong setup — the idea that we should all start a business or become a creator. I think that's a mistake too. What matters is that you can find a path where you provide value to someone, whether as part of an organization, as an individual, or as a leader — you can choose your path. But the objective should be accessing your capacity to create value for someone else.

"Do you think there's always going to be inequality?"

"Yes. It's either inequality or lack of progress. Those are the dance." China is very good at this. If you watch their policy, they just recently shifted to a free market policy — they literally just have a throttle. I always imagine Xi has a lever in a big room where he goes "free market, not free market," rolling it back and forth.

When you have a free market, people push the frontier. They discover new things, make things never done before, create new value, and the economy grows. The problem is it doesn't grow evenly initially. Say you came up with a goose that lays golden eggs — an AI — and this technology is making golden eggs, and you're getting very wealthy. Then the goose makes two geese, three, four. It starts breeding, and you start selling the goose. It takes years before everyone has a golden goose. During that time, you've raced ahead because you've got dozens of geese making lots of golden eggs for you. A lot of people say, "I want a golden goose," but by the time they get one, you've got a million golden geese. It feels like inequality, but the truth is now everyone has a golden goose — and if you'd told them 30 years ago they'd have one, they'd say, "No way, that'd be amazing." Every day they're making golden eggs, printing money, buying stuff with it.

That's the problem with progress, whether it's medicine. Many medicines today cost a million dollars a dose — CAR-T therapies, for example. These can save you from cancer, but only 50,000 people in the US can get them every year, and hundreds of thousands die because they don't have access. Availability is limited. There's technology like AI: not everyone has access to it, not everyone owns a piece of a model, not everyone's using it — and the people using it are getting ahead faster than those behind.

"In your golden goose analogy, by the time the golden geese get down to the lower end of the socioeconomic ladder—"

"I'd use the word diffuse."

"—those golden eggs buy less."

"That's right. Along the way, the government says, 'Let's give everyone that — let's give everyone access to more stuff to make up for that difference' — and then things get more expensive."

Part of the problem is that when the nation is prosperous and GDP is high, four-year election cycles push politicians to get in by offering voters lots of free stuff.

"The kid that got elected president in middle school was the kid who said he's going to make the vending machines free. That's always going to be true in any election." That's how the system grows, and we don't talk enough about the four-year cycle as a causal factor behind short-termism among leaders who just keep spending more money.

It drives me nuts that when I go to Washington every couple of weeks and meet with senators and representatives, literally all of them talk about is getting stuff for their constituents. That's their job and how they get elected: "I'm going to get you something. Vote for me."

The founding founders had a different idea — that people would rotate into public service, do their tour of duty, and rotate back out into private life. There's a lot of writing about this in the early days of the republic. Nobody believed America would evolve into a world of career politicians whose job is simply to be a politician and constantly get re-elected. All my politician friends are about to chop my head off for this, but I think it's nuts. I don't think "politician" should be a job title at all. The people representing us should have the long-term interests of the people in mind, constrained to serving a limited period and then returning to private life.

That means term limits: never more than two terms as a senator or four as a representative, and if you hold one public office, you shouldn't be able to transition into other public offices under those same limits. Otherwise you've built a system where your proven skill is getting stuff for people, and that causes the government to inflate, be mismanaged, and generate waste, fraud, and abuse along the way.

Everything is a trade-off, though. Even with an eight-year cap, someone could serve their friends while in office: run the transportation division, hand nice benefits to a friend at Boeing, then step back into private life as that company's chair.

That's fixable with a law — you go to jail if you do it. You can't buy stocks, can't own things, can't trade. If you serve in government, you shouldn't be able to do what a private citizen can do, and you shouldn't be able to enrich yourself as a private citizen afterward. But we should also set the incentive structure up so good people come work in government and there's healthy rotation between public and private life.

It's a hard problem. This stuff inevitably happens because we hand out lots of free things without thinking about the long-term consequences.

Has anyone fixed this globally? Is there any nation you look at as a model for a better system?

A lot of people point to Denmark and places like that and say they do a great job—there are a lot of public services and people pay high taxes. But there's not a lot of great progress there; no entrepreneurial success comes out of those countries. What happens in America—the freedoms afforded, the individual agency, liberty, and rights to private property—creates an incentive structure that says: go try and do crazy stuff that could work, and if it works, there's tremendous return, so it's worth taking the risk. That level of risk is what pushes the boundaries. It's what discovered CAR-T therapy, gene editing that can lead to curing human diseases, new advances in agriculture, new physics, new engineering, and AI. All of these came from the risk-taking system we've set up in the United States, built around individual incentives. And it works.

You can have great comfort for a great many people in Denmark, and people will be very happy living there and have a good life—but you will not have progress like you have in the United States. We push the frontier because we create this incentive structure, and that's where progress comes from.

Would you say the vast majority would choose to live in a place like Denmark, though? Looking at some of the stats: they have higher societal trust in government, police, institutions, and strangers; a standard 37-hour work week, five to six weeks of paid vacation, and a cultural emphasis on balance; a lower ceiling of extreme wealth but a higher floor, so less poverty and financial ruin. And according to global happiness metrics, Denmark consistently ranks significantly higher than the United States. Most people would choose that, right?

Absolutely.

Central planning might stop a few very high-agency, ambitious people, but the vast majority would theoretically be happier with a Denmark model.

Asked whether most Americans, given two buttons, would actually press the Denmark one, the guest argued that what makes American culture different is the idea of individual progress. You could offer people the Danish arrangement, but with a condition: you don't get to progress. You don't earn more next year, you don't have a shot at becoming a millionaire, building something great, or buying a second home. You're limited. On the other side, some things are guaranteed to you, but not much more—yet you could have the world. You could make a million dollars, own a second home, buy a second car, pay off your kid's college tuition debt if you earn well. There are more options here, and fewer guarantees.

Would the 63% who prefer the Danish model still choose it under that condition? He thought some within that group would, and that this tension explains where the United States is headed: people want more of a guarantee. The question is whether you want to maximize the number of people who are happy or the number of people who are truly free. Being truly free increases the rate of progress across the whole population, and that progress may be asymmetric and diffuse—it starts with some and makes its way to others over time. That, he said, is what America has chosen to be for the last 250 years, though he's not sure it's choosing that for the next 250 given what's happening across the country.

The interviewer pushed back: maybe people aren't happy, and maybe happiness was always the more important thing than progress.

Important in an aggregate way, the guest replied, but he returned to the individual. We choose to live in a society because it gives us the foundation described earlier—the safety and security to take risks, make our own decisions, and potentially build a media empire. Maybe the interviewer would have been happy in a system without that, but would he have reached his potential as a human?

The interviewer admitted he wasn't sure he would have been happy—it goes back to his own nature. Put him in a room and he'll try to create and build something. But he's learned over time that not everybody has that bias or those desires—not everybody has the childhood trauma of being the only Black kid in an all-white area, thinking money was freedom and selfishly important. So when he designs solutions, they probably shouldn't be designed for him. It took him a long time to reach that realization.

The challenge for America in this moment, the guest said, is whether you can give enough Americans the Denmark answer without taking away the liberty that Americans who want it have today—and he's not sure the answer is yes. As soon as you start taking private property from the successful ones, you take away their incentive to be in that bucket in America. He sees people wanting to leave. You won't solve for everyone, but if you solve for the Denmark solution, you lose a lot of the other side.

The UK is an interesting example for exactly this reason. One big subject of conversation there is that millionaires are leaving in high volume. The millionaire migration maps show them going to Dubai or coming to America, and many of the interviewer's friends building technical companies in particular have moved to the US for this reason.

The UK seems to have a spending problem, but also a productivity problem. Asked whether the two are related, the answer was: yes, of course. If you're collecting checks or receiving provided services, the individual incentive to be more productive goes away. And as a system, it becomes much harder to create economic incentives for productivity gains when tax rates are very high and you don't actually get to reap the benefits of the risks you take.

I've been catfished by furniture my whole life—something looks fantastic in the picture, whether it's a sofa or whatever, I order it, get excited, and then it arrives as something entirely different: the quality is significantly worse than it looked online, the texture is different, or it doesn't hold up the same way. That's why I love our sponsor Wayfair, who helped us fit out our green room behind me. Their Wayfair Verified system takes away all that second-guessing: products are hand-picked by Wayfair specialists for quality, so you can feel more confident once you've found the thing you love. If you're looking for furniture for any room of your house, go to wayfair.com to start your home refresh today, and make sure you use Wayfair Verified.

"Hello, I wasn't able to make that finance meeting. Can you tell me what I missed?" What I have in front of me is Whisper Flow's new notetaker, and for me it's been an absolute game changer. I spend most of my life in very long meetings across lots of teams and departments—40, 50 meetings a week, sometimes more—and I also sit in podcasts where interesting things get said while I need to focus on the conversation. Whisper Flow's notetaker captures the entire context of everything that happened throughout my week, labels who said what, and lets me ask about things I might have forgotten, including actions I said I'd do but haven't done yet. It builds this broader context of everything happening in my life, and it's an incredible superpower. Many of you come up to me in the street knowing I'm an investor in Whisper Flow and tell me how much it's changed your life—it's the worst-kept secret in productivity. If you want to give it a go, head to whisperflow.ai/stephven to download it and try it. Once you do, you'll realize why we let them sponsor this podcast.

That's why I wanted to start with the backdrop. Think about it: I'm already struggling, I'm in the 63%, and now you're telling me I'm going to lose the job that's paying me while I'm living hand-to-mouth. And what's the offer? We might get UBI — we might be sent checks in the post at some point. But actually, I like my job. It's a community, a sense of purpose, it gives me something to do. That's the narrative.

Yes, that is the narrative.

You don't believe the narrative.

I believe what everyone's feeling. But this narrative that AI will take jobs away — I don't believe it. I recently pulled up — you can go to archive.org — a Newsweek article from 1963 about how computers will take all the jobs. The mainframes were what got put in the basement in the '60s, and everyone had to walk past the scary mainframe with its blinking lights, buzzing, and sounds. The story was that the mainframe was going to destroy all the accountant jobs, all the typing jobs — all these things we get paid to do. And people thought: I like working in my office, I like my colleagues, I like coming in every day and having lunch. That's a good life, and this mainframe is going to ruin it. There was even an effort to legislate against these machines at the time, which didn't get through Congress.

Then in the 80s there was another wave, when desktop computers arrived and it was said they would replace workers — one desktop computer plus one worker, and fifty people disappear from the workforce.

But empirically — from data, from evidence across all periods of technological revolution like this — we've never seen jobs decline. We've always seen jobs go up, demand go up, and therefore incomes rise when new technology enters the workforce. That's because the new technology makes an individual more productive, so they can do more per hour with their time and, as a result, get paid more for that work. Imagine instead of having a painter painting a wall, he's in charge of the five robots painting the wall. Now he gets paid more because so many more homes get painted in a day.

People will say, "Well, there's only so many homes." That's always the argument — the fixed-pie argument, that the pie doesn't grow. But what always happens is that the pie grows; that's how to think about the economy. As automation or technology gives individuals more leverage to do more things with their time, more things get done. People come up with new product ideas: hey, we've got excess capital — let's build new homes, let's build a building.

But who benefits from the expansion of that pie?

In all cases, deep economic value has been created for the entire workforce, all the way down to the lowest level. It doesn't happen overnight, though. During that in-between period — when the mainframe is in the basement or the desktop computer is on your colleague's desk and you're thinking, "Oh my god, I'm going to lose my job" — there's a gap before the time your company is growing its revenue considerably, giving everyone raises and bonuses, adding more product lines, and doing more stuff. Then you think, "Okay, I feel better." That would be the argument from the '60s and '80s, and even back in the Industrial Revolution.

The question today is how work changes. Remember, in the Industrial Revolution, people moved from farms — 60% of labor was in farming, today it's about 2% — into factories, where they got paid a lot more and lived a better life.

They could actually afford healthcare and so on, and the transition in any technological revolution is net beneficial for all parties involved. But there is a transition. So what's it going to look like? Are we all going to lose our jobs? AI is definitely different, but I don't think AI is this world where humans are going to look around and say none of us have anything to do anymore. I don't believe in that narrative.

The reason is that I already see extraordinary creative things being done with AI that weren't possible years ago. People are creating TV shows single-handedly that would have needed a staff of 20 people, using AI-generated content and putting these shows out on YouTube, TikTok, or Instagram. You may not like how it looks today—there may be a lot of AI slop out there—but as this stuff gets better, the creative output of people using AI on the channels where we all consume media is only going to go up, and people are going to make money from that output that they weren't able to make before.

My own work is an example: the four hours a day I spend writing documents today, I can get done in 10 or 15 minutes, and then spend the other three and a half hours doing other things that create new value for my company.

Asked to return to the robot painter analogy, he clarified: instead of a guy painting on his own, he can now have five robots painting, and produce five times as much with his time. Think about the desktop computer revolution. When Adobe Photoshop came around, everyone said photographers were out of work, because a big part of the job was post-editing, cleaning up, and airbrushing—that whole industry was supposedly destroyed. What actually happened was that far more media was created, including print media, because Photoshop created leverage to produce many more beautiful images. Magazine circulation grew considerably as Photoshop came out, because it allowed much more production of high-value photography.

The reason this moment feels different is that the industrial revolution moved work from physical to cognitive, whereas Photoshop created new cognitive work out of cognitive work. The Adobe analogy raises the question: if AI takes the cognitive work, and robotics—as you said—takes much of the physical work, what is this third type of work?

I think it's still cognitive and physical. It's no different from what we've seen in the past. That doesn't mean there isn't an individual who programs the robot—which rooms to paint, how to paint them, how to do the detailing. The homeowner doesn't have the level of skill the painter does.

I couldn't do that.

Maybe at some point—you could make that argument on a continuum where eventually nothing exists. But along the way, new things emerge. Go back 50 years: there was no yoga instructor, no dog walker, no Instagram influencer, no podcaster. A million versions of new work emerged as the internet took hold and people could spend their time marketing themselves or their capabilities in other ways.

So I think it's a pessimistic view to assume robots replace all humans and humans are left with no logic, no agency, no capacity for creative work. My view is optimistic: AI and robotics enable individuals to do far more than they ever could have contemplated, giving everyone much more ability to realize their potential through their own agency.

Even this podcast, I think, will be significantly disrupted in some part. A couple of years ago we ran a test to see whether a synthesized voice of me could match my average view duration. I did a Steve Jobs episode reading the stories of founders through history—my voice, but AI had written the script and synthesized my voice, so I had no involvement. Forty to fifty percent of listeners stayed to the end of the hour, and they knew it was AI because it said at the top that this was an AI podcast. So I thought: there's a lot of the informational part of what you and I do—sharing pure information—that will be...

The informational part of what we do—sharing pure information—will be taken away. But listening to your answer about creators and art, you're saying there will be this irreplaceably human component. Yoga teachers aren't just telling you yoga; it's also community, gathering in real life, being there among other people.

I think the in-real-life job is going to explode and become extremely valuable. I always had this view that coffee shops would be automated, but I actually think the opposite is probably true. There will be a premium on coffee shops—I'd pay $40 for a latte because a human is making it and I get to be with other humans. I get to go to a workout class with humans, spend time at a concert, at dinners. This event, live, experiential part of our world that we think of as luxury today might become more core—maybe we spend 30 or 40% of our time in these spaces, while automation generates income for us and supports whatever work we're interested in. We'd spend far less time working and far more paying a premium in the live-space economy.

It's hard, isn't it? We don't know what these jobs are.

We just don't. I mean, imagine telling someone who worked in a factory in 1923 that a hundred years later there'd be people making X dollars as personal trainers, or someone whose job is dog walking in their neighborhood. I don't know if that would have fully grokked.

But the pace of disruption here seems different from the industrial revolution. It's targeting cognitive work, and it's built on this platform called the internet, which has this—

I'm telling you, I just don't see it. At my work, everyone's using AI, everyone's just doing more, and then they ask for more employees. You can think of a business as having two sides: revenue and costs. Revenue is making stuff that people pay you for; cost is how you operate the business. The argument has always been that you'll lose customer service reps—you'll lose the cost side. No one ever says you'll lose the revenue side. The truth is, on the revenue side, you create new products you couldn't create before because of AI.

Say you're a material science company whose job is to discover new synthetics for microchips, computing, travel, or airplanes. That field has been challenged by its experimental cycle: you create an experiment, make a product, test it—all manually, all creative ideas, all human potential. AI can now screen through millions of material science ideas in a computer, print them out, test them, and you can automate that throughput. So that materials company can have more revenue and make more products.

The interviewer asked whether people like the two of them—founders with capital, thinking about how to push intelligence to create new things—occupy a particular band. He could relate to the guest's point: AI has probably made him more ambitious about hiring, especially in the near term. But he caveated that there's a certain type of role he now hires much less for—especially entry-level roles, where he has to pause a lot more, because agents and tooling can do some of the things he would otherwise have hired for. Generally, though, he feels he's hiring as fast as he possibly can. He pressed the point: for the 21-year-old fresh out of university, the DoorDash driver, or someone assembling parts on a factory floor, the leap to this new world is so great that there's going to be a messy middle.

The guest disagreed. As new products are created and new revenue comes in, companies don't cut costs—they grow costs, hire more people, uplevel and train them. That's what he sees on the ground: the US just published one of its lowest unemployment rates. The idea that your current job gets disrupted and you lose it is not how enterprises transition. Companies transition by doing more things and bringing people up the ladder. He argued there's a media narrative around layoffs because it underlines what the media has been promoting: that AI causes job loss.

The interviewer pushed back: the narrative didn't come from the media—it came from the AI guys. He noted he listens to the All-In podcast and has barely missed an episode since the beginning, so he understands the guest's perspective well. Oddly, he tends to side a bit more with Jason, whose arguments seem to carry the most nuance—and when you consider people's incentives, the show often frames it as a media narrative when it's actually all the guest's friends.

The guest replied that none of them are his friends, and restated a point he's made on the show: there is a deep arrogance in making these proclamations—like "we're not going to work anymore." It effectively discounts the human, as if what you do today is your limit. He believes the "jobs will go away" idea is sourced by people seeing the crazy things they're building and concluding, "This is amazing, I'm responsible for this, we're going to change the world, we're going to take all jobs and they'll all sit in my computer." He doesn't think that's what happens, because humans have a profound capacity for doing things these technologists don't necessarily see. The great technologists see technology; they don't see people. You have to see people—the potential of individuals—and believe in it.

He refuses to be pessimistic about people. Humans are amazing in their adaptability: we can live in any climate, eat any kind of food, work all kinds of jobs. Arguably, many humans have been limited in their potential, and AI enables more of those individuals to realize more of it.

Much of our conversation about the US economy has centered on incentive structures leading to inevitable outcomes. The older I've gotten, the more I just look at incentives to figure out behavior and also to figure out who to trust.

When I look at the incentives of the people building AI companies: Andy Jassy has said there will be fewer people doing some of the jobs being done today, and then tens of thousands of jobs were cut. Mark Benioff, who has been on your show, cut 4,000 jobs and said he needs fewer heads because of AI. Maybe they're saying that because it's a good narrative for their share price—cutting costs plays well. But then I look at the AI CEOs—Dario, Elon, Sam Altman—and their incentives, with the exception of Elon, have just changed with whatever's good for them at the moment. Actually, I think Dario might be somewhat of an exception, because he seems more nuanced.

If you go back through their quotes: Sam Altman in February 2015 said, "Development of supermachine intelligence is probably the greatest threat to the continued existence of humanity." Elon in 2014 said, "With artificial intelligence, we are summoning the demon. It's far more dangerous than nukes." And the statement that "mitigating the risk of extinction from AI should be a global priority alongside societal-scale risks such as pandemics and nuclear wars" was signed by Sam Altman, Demis, Dario, Bill Gates, and hundreds of others. Altman in the Senate in 2023: "I think if this technology goes wrong, it can go very wrong." The OpenAI superintelligence letter warned that the disempowerment of humanity or even human extinction could be caused by superintelligence. On your show, Altman said jobs are definitely going to go away, full stop. At Barraud Army Day in May 2025, he predicted a white-collar bloodbath—half of entry-level white-collar jobs gone, unemployment at 10 to 20% within one to five years. That's pretty high. And Altman at the Federal Reserve in July 2025 said whole job categories will disappear, customer support first—that one is "totally, totally gone."

Then suddenly the message changed. Before his pre-IPO discussions, he said things could go pretty wrong, and now he says he's changed his mind, that he's delighted to feel he was wrong, and he's talking about how OpenAI and ChatGPT will benefit all of humanity—and that's their goal. Should we trust the same people who told us to be scared when they now say don't be scared?

First of all, there were a bunch of companies that said early on they were cutting all their customer service reps—the big company that starts with a K. Klarna. But I called the founder live on this show.

He said that he's been misinterpreted in the press.

He said he had 7,000 employees by the end of last summer, and he hired on 3,000—but the media framed it as: I had 7,000 employees, tried to cut some, realized I couldn't, and went back. That's not what happened. I see that story repeated on all these podcasts, including yours.

But he did say they are cutting because of AI, and he's confirmed that—he's DM'd me on X, I have the DM, and live on this show he confirmed it to one of our listeners.

Is his business still growing?

The business is growing. What he did say—and this is a point of nuance—is that with the cost savings, he's able to offer better white-glove services to his highest-ticket clients. Which makes a lot of sense.

You could apply that analogy to another business, where they would take that capital and invest in offering new products and services — but then you need employees to deliver those new offerings. So you do actually hire up, and I think that's what I'm seeing on the ground. Even while this business owner is doing an acute restructuring, I just see people hiring more because they can do more with AI.

It's counterintuitive, but if you're a business manager, you can invest $10 and make one product a year, or invest $20 and make five products a year. You're going to invest the $20 to make five products. You're going to spend more, and I think that's what AI does: as soon as you can leverage how much more productive people are, you'll put more capital toward hiring and scale up. That's just what I'm seeing.

Look, I could be wrong — all jobs could go away in five years and we could have 20% unemployment — but I don't think that's the case. I think we can go back to the question of where people are going to be spending their time. I do think there's going to be big growth in media, encompassing everything: digital content creation and so on. I think there's going to be big growth in entrepreneurship, and big growth in IRL stuff.

The interviewer brings up an earlier interview, noting that the media reported the CEO had doubled down on AI and then reversed course because it didn't work out. Since their last conversation was nearly a year ago, he asks for an update on Klarna's business, AI agents, and related efforts. The clip played is from five months earlier.

The CEO explains that Klarna was early in releasing AI to support customer service, which initially meant more calls were handled by AI — something customers liked because those chat messages were much faster and of higher quality. Since then, that has expanded slightly. But the company also tried to communicate that in a world where AI is cheap and available, the value of human interaction will be regarded as higher. So the future of customer service VIPs is a human, and Klarna has doubled down on providing more of that. At the same time, efficiency gains within the company have continued: Klarna used to be about 6,000 people and is now less than 3,000, two to three years after it stopped recruiting, while revenue has doubled.

The host pushes back: this is a startup, and a startup that grew very quickly and then cut heads is a pretty common phenomenon. He isn't going to concede or argue that jobs aren't going to change, but he questions how much restructuring was really about the "AIification" of the business versus normal startup dynamics. The real questions, he says, are how many of the customer service staff were actually fully occupied on the phone — four hours a day, two hours, or eight? Were they truly fully employed? — and what they are doing now, and whether that is defensible against the spread of intelligence.

They could be lawyers, he concedes, but they could also be doing DoorDash. And Dara has said that the 9,000 people who are riders won't have jobs — the ones driving food around. That connects to DoorDash's big investments: the host was with Dara recently, and he was talking about their autonomous vehicle investments. Asked what Dara thinks those drivers will do, the answer: he loosely hypothesized it might be data labeling.

The host then reframes the debate about AI claims and backtracking. The important question, he argues, is whether we should stop AI development — and whether we really can affect what China does, what Europe does, or what some lab in Indonesia does if they get a group of people together. There are moments in human history when technology advances and you can't just put the catapult away and say no one can develop it anymore; once the catapult's out there, it's out there. So the question is what enables the most productive path forward for US policy and US citizens — and China is asking itself the same question for its own citizens. China's answer, he thinks, is to make models free and ubiquitous, which benefits a large part of the global economy but doesn't do well for Sam Altman, Dario, Google, or others who have invested billions or hundreds of billions in model development — investments now reduced to rubble because China has a model that's as good or cheaper. The more important question, then, is what the US does to navigate: if we stop building data centers, those data centers are going to be built somewhere else.

They'll be built in Iceland, in China, in Indonesia—anywhere you can get access to energy and plug into the internet. So the real question is whether we want those data centers to exist in the US, creating jobs and economic value, or built offshore. Do we want AI models built in the US or by China? Those are the important policy decisions. As for the idea that you can slow or stop AI development—unfortunately, I think it's too far too late. The AI is out there.

Whether it's bad or good requires looking at what's actually happening on the ground. There are plenty of stories: look at the number of companies doing things that were never possible before because of AI. Take delivery—companies are deploying robots on the ground, robotic cars, drones. Those companies aren't just employee lists. There are 15,000 new jobs being created at the factory making the robots, or 15,000 jobs to service them because they break down constantly, or 15,000 jobs to make the food that goes into the robots—though maybe that gets automated too. As you think about the impact, the economy doesn't just shrink in terms of jobs. We're just not seeing it in the data—it's not happening. And when it does happen, that would be a good time to have a conversation about AI taxes and all of that.

The question is what metric or piece of evidence would actually trigger intervention. The answer: job loss — and wage growth. If wages are declining and unemployment is climbing, that's a red flag, an alert moment when we have to address it from a policy standpoint. But it's very speculative right now. And if you get in the way — if you stop advancement, or the ability to create higher-paying and new jobs — it's going to happen in China, in Europe, in South America, and the US gets left behind. Maybe that's a decision we choose to make, but tracking the consequences we're worried about is the most important thing to be doing at the moment.

There was a study — I think Stanford did it — showing a 13% relative decline in employment for workers aged 22 to 25 in roles most exposed to generative AI. Could it be that by the time we see significant job loss, it's a bit too late to have the conversation, because the decline could be quite rapid given the nature of this technology and its adoption?

No — what you're describing is exactly the right time: when we see the first sign.

Digging into that study, in software engineering what's happening right now is that young people aren't getting hired. You can use AI to write a lot of code, but to be really good at using AI to write code, you need experience writing code yourself. So there's a fight for senior, experienced software engineers and disinterest in junior ones — you're no longer putting a junior engineer in a seat to write their first line of code, which is what you used to do. With AI you can turn one engineer into a hundred, but the junior person can't do that; the senior person can. So companies want to hire more senior people, and that's what's happening across software right now.

Beyond software, operational functions that used to be done by people are also being done by software. You might have had a personal assistant whose job was to respond to your emails, check and organize them, and put them into your Monday board. Now I have an agent that goes across all of my inboxes every single day and pulls everything into one central system. I used to have someone doing that kind of work; now it's a software job — it's Claude doing it.

"So now the PA can use Claude to do it?" I don't need the PA. I can just do it myself.

"What else could you use that PA for that you couldn't even contemplate in the past?" At the moment, it's real-world stuff. I probably would have continued to hire more people on that team, but now I don't need to, because I have this agent that helps me with my scheduling.

"What are you going to do with the extra money you're not spending?" That's a good point. You're not going to just leave it as cash in the bank — you're going to invest it in something.

I could theoretically go buy the S&P 500, and that money goes to individuals who are then able to employ more people. That capital finds a way to get spent; it goes into the economy. That's my point: this is why economies grow when technology shows up. The leverage you gained gave you more capital. You could either build a new product and hire more people to do it, or you could choose to buy the S&P — but then that money ends up in other people's coffers, and they're using it to hire people and do new things. At some point, the whole point of the economy is that money flows, and it will ultimately flow into individuals that are productive.

"Do you not think a lot of it is going to flow disproportionately into the billionaires and the wealthiest people who are building the robots?"

I'm a huge advocate for open source for this reason.

What open source means

Open source just means the software is free for anyone to use. Go back to the early days of the internet: Netscape made server software and a browser. To produce a website, you paid for Netscape's server software and installed it on a server; to browse the internet, you paid for Netscape Navigator. Charge, charge — gatekeeper, gatekeeper.

Then the Mozilla Foundation came along and made an open-source browser called Firefox, and the Apache Foundation made an open-source web server. Now anyone could download Apache, make their own website, plug it into the internet, and it would be available to everyone in the world for free. Anyone could download Firefox and browse the internet for free. Open source enabled all the entrepreneurs who built money-making websites to proliferate — they weren't gatekept.

Open weights for AI

In the world of AI, open-weight, open-source models are free. You don't have to pay Anthropic, Claude, Gemini, or ChatGPT for AI — you can use the model for free, and no one is making money off your use of it. There's no one getting paid; the only cost is finding a computer to run the model. You could use your own computer or rent one, and that cost is coming down like crazy too.

This is why technology without government intervention is so amazing: it always gets cheaper, and then everyone benefits.

The risk argument

There could be a risk, though. Isn't one of the key arguments against open source that if anybody can download these AIs onto their computer, they can jailbreak them and use them as...

Two things are being fought over here. First, government people are saying, "That's coming from China—we don't know if there are back doors." That's not how open-weight models work, by the way; people who are technically deficient don't fully grok what an open-weight model really is. It's not some backdoor program you're running. It's literally just a list of numbers in a document—trillions of numbers, that's all it is.

The second argument is that now that AI is everywhere and everyone has access to it, you could use it to design a cyber weapon, a bioweapon, a physical weapon—all these things you can now do. The same is true with DNA printing and DNA sequencing: everyone can buy a DNA sequencer and a DNA printer and could theoretically make a virus. That exists today. We haven't stopped DNA sequencing or DNA printing, and we don't have constraints around it. Similarly, we haven't stopped people from buying guns in the US—we can roll our eyes at that one.

But look, there's a very good argument to be made that any technology that could theoretically cause harm should be banned, or that the government should control it and decide who gets access. The flip side is that if it's allowed to proliferate, there's far more benefit than risk, and the risk side needs to be mitigated—that's where you build good defense systems. We have technologies for monitoring, cyber defense, bio defense, and so on, which are critical especially in a post-AI era: what are the big attack vectors? I think that's probably the way to think about this.

And that might help with the inevitable inequality—the golden goose collecting into insight. I will bet you anything that in the next 10 years there will be a billionaire who has absolutely zero net worth today, downloaded open-source AI, and built a company that made them a billionaire five or seven years from now. They'll be someone who comes from nowhere, and everyone will say, "Wow, look at that." There are going to be a lot of stories like that—that's what we saw with the internet.

This is something I've made for you. I realized that the Diary audience are strivers, whether in business or health—we all have big goals we want to accomplish. One thing I've learned is that aiming at a huge goal can feel incredibly psychologically uncomfortable, like standing at the foot of Mount Everest looking upwards. The way to accomplish your goals is by breaking them down into tiny steps, which we call the 1% on our team—and this philosophy is largely responsible for much of our success here.

So that you at home can accomplish any big goal, we made these 1% diaries. We released them last year and they all sold out. I asked my team over and over to bring them back, with some new colors and minor tweaks, so we now have a better range for you. If you have a big goal in mind and need a framework, a process, and some motivation, I highly recommend getting one before they sell out again. You can get yours at thediary.com, and the link is in the description below.

If the button below says "subscribed," you're already subscribed. If it says "subscribe," you're not yet—and if so, please hit that button. It helps the show more than you know, and according to the algorithm, you're someone who watches our show but hasn't hit it yet. Thank you so much.

When Elon says that working will be a choice, you think he's wrong.

He's selling robots, but... I do think he's right in a sense.

Think about how many hours a week you work. The interviewer works all the time whenever his fiancée isn't around, and the speaker says he's the same — every moment he can, he works. But they're both more entrepreneurial. Look at Denmark, for example: 37 hours a week. What was the average number of hours worked 50, 60, 100 years ago in the US, when you did farm labor managing your own farm as your whole business, or worked in a factory? There were 12-hour shifts, and far more work per week. Going from 100 hours a week to 80 to 60 to 40 is part of the progress of humanity: we're no longer encumbered by this labor category — being told what to do, how to do it, and when, in order to get money to pay our bills. We're moving toward a world with more capital, and that capital is also time. He has now accumulated the ability to choose what to do with 40 hours of his week.

Asked which part of his AI hypothesis could most easily be a miscalculation, he names mass unemployment — specifically, the assumption that people will choose to move to higher-paying jobs when they see an opportunity. That choice is the key to the whole thing.

He illustrates it with New York taxi medallions — the right to drive a cab in New York. The city limited the number of cabs, so medallion prices rose to $500,000, and Citigroup was a big lender, loaning drivers the money to buy them. Then Uber came along — one of the big battles around Uber at the time. You could get paid more and there were more rides; Uber grew the number of people taking car transportation around New York considerably. When Uber demand shot up, many drivers said, "I'm not going to be a cab driver anymore. I can make more money as an Uber driver," and the number of Uber drivers grew. There's been a lot of New York regulation since then, but everyone made the choice and moved; the market found its way. Individuals realized they could earn more, pick their own hours, and not rent the medallion.

The interviewer pushes back that this is a very similar sidestep — going from taxi driver to taxi driver for a different firm using an app. The speaker concedes the point: he could pontificate all day and they'd both be wrong, because he doesn't know what the leap will be for an individual. We have to start pulling these stories out of the economy.

As an example, he cites a Palantir promotional video interviewing a nurse (he initially says a hospital administrator, then corrects himself). She said she normally spent most of her time filling out paperwork on the computer, but the Palantir software takes care of all that, so now she can spend more time doing rounds face to face with patients. It's a better job, and better for her — a poignant story about what AI actually does on the ground. There are many examples like this; people could counterargue back and forth all day and probably both be wrong, but this is what he observes is actually happening.

The interviewer opened by noting that they agreed jobs would become more human in every sense of the word — the yoga teacher example, the nurse who can spend more time with her patients and become even more valuable as that work grows scarcer. His remaining concern was the transition itself, especially retraining at speed and the question of human agency: should we be preparing for such a shift? He believed that, absent existential problems, we would net out in a better place, but the journey there remained a big question mark.

The speaker agreed, saying he wasn't defending the transition as a clean, smooth line — only that we will get to a better place. We could argue all day about displacement or dysfunction along the way, or about people smiling as they jump to new jobs and find their lives better. He was optimistic in that sense, but agreed we have to track it: if people are out of work, can't find work, and struggle to pay their bills, we have a serious problem.

Asked what the solution to that problem would be — whether it was UBI — he answered, "Well, it's going to be socialism." Asked whether we would adopt socialism in that messy middle, he said yes.

Pressed on what that would look like in reality — just giving out money and free food, like the Mundan grocery stores he had seen — he said that approach wouldn't work, though it would take time to fail and would look great in the meantime. Instead, he expected more checks and more government employment: the government creating jobs to pay people. These would not be productive jobs — if you're paid $100, you don't create $120 of value; you might create $60 or $80. But either way, there would be more government payments, whether in the form of UBI, government job growth, or some other financial support.

The direction of travel at the moment is toward a more socialistic America. Asked how he feels about that, the speaker says he doesn't want to lose his agency, capacity, liberty, or freedom, and feels sad for people who embrace it because they don't realize they're losing their freedom in the process too. It offers them a solution where they feel no other solution is being offered.

It's also logical to point out that billionaires want to stay billionaires, aren't as philanthropic as they claim, and are protecting their own interests — they want to remain free.

The wealth tax conversation with Ro Khanna

The speaker describes a call with Ro Khanna, a congressman from California's 17th district in Silicon Valley, a former moderate Democrat who has pivoted to hard-left true socialism. Khanna asked why people were so outraged at his support for a wealth tax — not just billionaires, but anyone who has built something and created value: entrepreneurs and successful people. Khanna couldn't see the problem: it's just 5%, 2%, or 1% a year, so why do people care?

The answer: it's the first time in American history that you'd be taking something from someone after they've paid their taxes, when it's now their private property, with the government deciding how much of their private property they get to keep. It doesn't matter that it's 1% or 5% and that "the billionaires will be fine" — it's fundamental to private property rights. Anyone can imagine what it feels like to list out everything you own, send it to the government every year, and have them decide what you keep. That's how it feels to everyone.

The math doesn't work

Financially, a wealth tax doesn't solve the problem either. Total net worth in the US is $183 trillion. Billionaires' total net worth is $8 trillion; even extending down to everyone with $50 million or more, it's $23 trillion. The remaining $160 trillion belongs to everyone else. The bottom 50% of Americans hold only $6 trillion — the net worth sits with the middle, the population people call boomers or Gen X. So taxing 100% of billionaires' net worth yields $8 trillion — one year of government spending. A wealth tax doesn't solve a fiscal problem; it just makes everyone feel better about taking money from billionaires.

"Are the billionaires paying their fair share of taxes?" is the politician's trick: comparing tax paid as a percent of assets versus as a percent of income. Income is when you sell stocks or earn money for services. The speaker pays 53% tax every year as a wealthy Californian — roughly 14% California plus 39% federal — and just pays it rather than avoiding it. In California there's no difference between capital gains and income anyway, which, he argues, is how it should be at the federal level.

Could the US lower income tax and raise capital gains tax? Absolutely. Could income tax be lower for people working for their money than for people earning gains on capital assets? Absolutely — that would solve the problem. But taking assets doesn't: even 100% of billionaires' assets is one year of spending. The tax rate is the fiscal problem.

For the 63% who may be championing socialism because they need a solution and it sounds like one—make my bills cheaper—what they don't hear from other politicians is a better solution, and this one sounds great.

This is what's crazy to me. I don't get why politicians are fighting over a wealth tax when there are literally 15 other things you could do to make this better. Make capital gains and income tax the same, cut stupid spending at the federal level, and suddenly things become more affordable. People have more take-home pay and can afford more stuff.

But let's be honest, you're not going to be able to cut spending at the federal level. They've tried, and because of the four-year election cycles we just talked about—Elon went in, he tried to cut it, and he bounced straight back out.

Yeah. That to me was the most depressing part, because I thought, "Oh my god, Elon's going to go in, he's going to blow up spending, we're finally going to cut all the nonsense out. And if Elon can't, no human can."

Then I went and visited DC, went to all these Congress meetings, and sat with senators, and I realized none of them want to cut it. They're all on board with it. This isn't even a party thing—everyone's doing it. This is the system. This is how it's set up.

Asked whether we'll be able to live forever soon, or at least extend life significantly within the next couple of decades, the answer is yes — because we've discovered the mechanism of aging, the most fundamental of human diseases.

How genes make proteins

Our bodies have 10 trillion cells, and every one of them carries a copy of our DNA in its nucleus. Segments of that DNA are called genes: A, C, and G are the nucleotides, and a bunch of them together form a gene. A copy of a gene travels to a protein printer in the cell called the ribosome. The DNA acts as the programming for which of 20 amino acids to use — three letters code for one amino acid. The ribosome takes a segment of DNA copied as RNA and prints out a sequence of amino acids defined by that gene, producing a protein. A protein is simply a chain of amino acids that, as soon as printing is done, collapses into a three-dimensional machine.

These machines do remarkable things. A protein in your mitochondria spins 6,000 times a second as hydrogen protons pass through it — that's what generates energy in your cells, making ATP, like a turbine. Other proteins carry oxygen from your lungs to your cells. Genes are simply the programming for making proteins in the cell.

Epigenetic switches

Your DNA isn't a loose strand; it's a compact cluster, with small molecules sitting on top of it that turn genes on or off — histones or acetyl markers. This whole system is what's called epigenetics. For any given cell, the genes that are on are producing proteins, and the genes that are off are not. That's what makes our cells different: every cell has the same DNA, but eye cells, brain cells, heart cells, and skin cells just have different sets of switches.

Every time you drink alcohol, get sun exposure, smoke, or eat burnt food, these things can cut DNA. Your cell is very good at repairing the break — this happens millions of times a second across your body. But each repair carries a chance that a switch moves slightly. If it moves to the wrong place, the cell starts making the wrong protein or stops making the right one.

Aging as the core disease

Statistically, over time, this drift accumulates. As a baby, everything is tuned correctly and all the right proteins are being made — your skin isn't wrinkled, you can run, breathe, and see perfectly. Around 30, vision starts fading, running gets harder, wrinkles appear. By 40 and 50, the switches have moved to the wrong places, cells become dysfunctional, and if enough cells are affected, tissues start to misbehave: wrinkles form, the heart doesn't beat quite right, the retina transmits light less clearly so vision blurs, hearing declines, and thinking and reaction slow down.

That's what aging is — the core human disease. As epigenetic switches get displaced and tissues misbehave, we develop all these diseases, and this could be one of the core drivers of cancer.

Why does aging seem so even across the body? Because all these cells are suffering the same statistical effect of time: DNA breaks happen in every cell in your body, and as those breaks occur, the epigenetic switches move around at the same rate in all parts of the body. That's why everything starts to suffer at once—"my hip hurts, my eye, my brain." Young viewers may not appreciate this, but being 46, I'm starting to, and it's becoming more of an issue now.

In 2006, Shinya Yamanaka discovered four proteins that, when applied to a cell, move all of that cell's switches to the position a stem cell has — essentially the state of an embryo. The cell then operates like an embryonic cell, and other proteins can move those switches further to turn that stem cell into an eye cell, a skin cell, a heart cell, a lung cell, or a brain cell. So in the mid-2000s we suddenly had the technology to take any cell and turn it into a stem cell — changing one cell type into another, a process called differentiation.

Years later, someone asked: what if instead of applying a lot of these proteins, we applied just a little? They took the Yamanaka factors — just those four proteins — and applied a microdose to a few cells. The switches started to move, but the cells didn't revert to stem cells. Instead, they became young versions of their own type: an eye cell became a young eye cell, a skin cell a young skin cell, a heart cell a young heart cell. A small dose of these proteins moves the switches just back to their correct position.

This is now called epigenetic reprogramming, a new area of science that has already shown striking results: in mice, where the animals lived to the equivalent of 100-plus years old; on monkey tissue, where it removed wrinkles; and one of David Sinclair's companies has dosed the human retina to reverse blindness in people with retinopathy.

Epigenetic reprogramming opens an entirely new lane of medicine. The question is not just whether we can address the downstream effects of aging, but whether we can go to the core and reset aging in all the body's cells — and so far, the answer is yes. The risk is that resetting too far turns the cell back into an embryo-like state and it becomes cancer. That's what researchers have been working against for the past few years: finding the right dose of the right proteins for the right cells so they reverse their age and become young again.

At this point it's really an engineering exercise rather than a biological proof of concept — the science and the foundation are there, and what remains is the engineering. I'm extremely excited and optimistic: in the next couple of years we're likely to see more human therapy trials that not only address the core of aging but also target specific outcomes — reversing wrinkles, improving cardiac muscle and heart tissue, as was done with the retina. I expect a lot more of this coming to market over the next few years.

AI is expected to accelerate this work, since it can theoretically make new discoveries and run tests faster.

To give a sense of what happens inside a cell: it contains about 10 billion proteins interacting with each other—building things, breaking things, creating new molecules and proteins. If each protein were the size of a human, the cell would be the size of Manhattan. Imagine a Manhattan-sized city with 500-story skyscrapers and 10 billion people living in them, going back and forth, building things together, talking, making babies—non-stop, without sleeping, for 80 years. That entire set of interactions is one second in one cell. That's how complicated biology is. And we have 10 trillion cells in our body, all constantly shooting out proteins and messages to each other.

AI lets us build models to understand what's going on and to make predictions about proteins we could put into cells that would change the epigenome on the DNA, without needing a deterministic understanding of the system. With AI, I can generate a million ideas for a new protein, test it in silico—on computers—to see how it would interact with other molecules, whether it has the same effect as the original protein, and whether it would be more efficacious. That reduces my candidate list, which I then take to a lab. Labs are now being automated, so I can test candidates on cells to see if they actually reverse aging, and test them in models. The result is a probability of success in 10 years that goes from 1% to something like 75%.

That's how AI plays a role in life sciences discovery. It's transforming not just epigenetic reprogramming, which is now a whole major avenue of medicine, but every aspect of biology. We can use AI to simulate these immensely complicated interaction sets and understand what they drive toward, instead of a scientist saying "I've got a good idea, let's try that," then five weeks later, "another good idea, let's try that." All of that gets reduced to silicon.

Asked whether he trains his own AI or uses publicly available frontier models, he answered: in his world, both. They build their own predictive models from their own data, and separately use off-the-shelf life sciences models plus some general models—a great combination. The things they can now do are unbelievable, and they're hiring more people because of it. AI didn't eliminate jobs; the frontier has gotten much wider. That's why he's optimistic that AI won't destroy jobs—all he sees is opportunity: new companies, new ideas, new pathways, new breakthroughs in science, materials, and engineering—things we could never have tackled before that we can now tackle with AI.

The interviewer conceded it was a persuasive argument.

Asked what the great lies in America are, the speaker named three:

1. If you go to any college with any degree, you're going to have a great job waiting for you at the end.
2. If you buy a home, you've achieved the American dream, and all of your assets should be in that home.
3. Social security will take care of you when you get old, and it's a fair system.

The truth behind the social security lie, he argued, is that everyone should have ownership in stocks, because all the money that went into social security went out to the government to pay bills and was never actually kept in the trust. As a result, social security is going to go bankrupt — one of the great lies about being safe and protected when you retire.

Public pensions and 401(k)s

Public pensions suffer from the same problem. In a defined benefit plan, money is taken out of every paycheck with the idea that accountants and investors will invest it well and you'll receive your benefit — say $1,000 a week — at retirement. If they do a poor job, the pension plan goes bankrupt and you don't get your retirement check. If they do too good a job, or too much money is put in, the plan is overfunded and you should be getting more money out than you actually receive.

Companies realized pension plans are unfair — they either go bankrupt or end up overfunded — which is why everyone moved to 401(k)s. The middle class discussed earlier holds $160 trillion of assets largely because many of them had 401(k)s. Meanwhile, the bottom 50% who rely only on social security were making less than 3.5% a year on their money. The great lie, he said, is that these public retirement funds would take care of their people.

On healthcare

Asked for a fifth great lie, he said he doesn't have a great answer on healthcare, because he believes there's a moral obligation there. He could make the libertarian argument but won't, since he thinks it's probably right that in a civil society no one should be without healthcare.

The hard question is how to make that work when the government is paying the bills and costs go through the roof. That, he said, is a six-hour, very technical conversation about making healthcare accessible — and it definitely has to be a multi-tiered system.

Looking at the track record of US foreign policy decisions over recent decades, very few actions have led to positive outcomes. The US often finds itself trapped in quandaries, contrasting sharply with China's approach. In line with Sun Tzu's philosophy that having to fight means you have already lost, China demonstrates a capacity for influence without kinetic action—relying on economic gamesmanship and trade incentives.

By contrast, the US military and advisory apparatus is structurally primed for action, often pushing leaders to intervene when they perhaps shouldn't. At the same time, China's posture is partly enabled by the US acting as a global police force, leaving China's own military power largely untested at scale.

Oil Reserves and the Standoff with Iran

In Iran, the situation presents serious economic risks, particularly regarding energy supplies. With the US Strategic Petroleum Reserve (SPR) significantly drawn down—estimated at around 24 million barrels, or roughly a few weeks of supply—disruptions pose severe domestic threats to gas prices.

This dynamic creates a strategic stalemate:

• Even after leadership strikes, Iran has little incentive to concede and can simply wait out US pressure.
• Iran has threatened retaliation against energy infrastructure across neighboring Arab allies, including Saudi Arabia, Qatar, and the UAE.
• The US cannot easily escalate due to the devastating economic and security fallout that retaliation would bring to regional allies.

Political Pressures and US Strategy

Domestic politics heavily influences the administration's posture. Entering the midterm election cycle, military conflict is deeply unpopular across the political spectrum, creating strong pressure to de-escalate.

From an outside perspective, US policy appears erratic—shifting repeatedly between threats of strikes, cancellations, and proposed deals. While some view this vacillation as strategic unpredictability, it largely reflects the difficult constraints of avoiding an unpopular escalation while facing limited viable moves.

That unpredictability can in some cases be an advantage. People have made the case that the president's unpredictability makes it hard for others to find a clear path to attack him. My poker player buddies—the best in the world, who I'm friends with—base their whole strategy on creating randomness in their gamesmanship so it's very hard for someone to exploit their gameplay. I just don't know what you're going to get with President Trump. You can meet with him, hear him say one thing, and then he'll do something different later.

I understand the merit of that argument, but in this situation what I'm observing is that he's become predictable—in terms of not taking action. He threatened the biggest strike since World War II, then called it off.

What would you do if you were him? Honestly, there are only bad outcomes, so it comes down to which bad outcome I'm most willing to accept.

For someone like Trump—I don't know him, but I'd guess he's probably thinking about which one would save the most face.

That's probably the right objective function for him. I think the worst thing is that all outcomes, in my view, lead to the Strait of Hormuz becoming basically a tool of Iranian negotiation, which it wasn't before.

A new point of leverage.

Yeah, a new and powerful point of leverage in that region. But it may create alternatives—there's conversation about building other pipelines out to Israel, and so on. Maybe at the end of the day it creates an incentive for other channels to open. TBD. I do think it's not just about the Iranian program, but about where all the enriched uranium is. If you asked a military adviser, that's probably what they'd see as the core issue, because it takes a long time to enrich uranium for nuclear fuel like this. So securing the enriched uranium is probably a key objective.

Asked whether a Republican will win the 2028 presidential election, the speaker said no — he thinks it will be AOC. His reasoning comes down to everything they had discussed about the state of the nation: people feel desperate and want a candidate who can promise to make their lives better. They are less concerned about issues like freedom of movement and much more concerned about paying their bills, getting healthcare, and getting child care.

He believes she will run, and that she would beat Buttigieg, Kamala Harris, and Newsom. Newsom, he argued, looks and acts the part but will struggle to sell himself to the rest of America on a national stage. As for Kamala Harris, no one wants to take the risk on her because she has lost before. AOC, by contrast, is like Mamdani in that she brings a very different voice, and he thinks the DSA movement has real legs — he sees few young people speaking out against it, mostly old Democrats.

He hedged, though: "But look, what do I know?" The interviewer noted that on the All-In podcast the speakers had predicted a Democrat or a socialist would win in 2028, to which the response was that there is a lot of runway between here and there.

The Golden Age Hypothesis

The speaker then turned to the profound frontiers ahead. Many of his friends hold a "golden age hypothesis" — that we are entering a golden age without realizing it, and that AI unleashes a capacity for human advancement. A billion people in the world are starving and hundreds of millions are dying from curable disease, and he believes these problems will be resolved more quickly than we realize.

The productivity gains from AI won't just be a battle over whether the painter has a robot or not, but about entirely new frontiers: a profound revolution in material science, and travel to the moon — which might seem like a billionaire's fantasy but is becoming a real industrial frontier everyone will participate in. Extraordinary abundance arises from this, not merely making what we already have cheaper, but the new frontiers that entrepreneurs want to pursue, that dreamers see as the hope for tomorrow, and that pragmatists are acknowledging as legitimate for the first time in our history.

That is the light rising off the horizon, and he feels very good about it. Maybe AOC does get elected — that is just his view of the sociopolitical timing today. But he believes everyone's life will profoundly change in the next couple of years, and he is profoundly optimistic. Despite all the discussion of socialism and government spending, his experience on the ground is that the amazing breakthroughs are here, and there is no stopping them.

When you hire lots of people, you learn that behind every announcement you make in your company—raising money, whatever it might be—there's a second question every team member is actually asking: what does it mean for me?

On the ground, it's the 63%. They look up and see a trillionaire—the world's first trillionaire. They see AI, they see job displacement coming, and then they say: I can't pay the bills. So what's the mechanism, right now, for helping people on the ground benefit from advances in this technology?

From a public policy perspective, what we need to change is 401(k)s, so people have direct ownership in these things and can look on their phone and see how they're benefiting.

Everyone thinks AI value is going to accumulate to two or three companies. I don't think that's how it will play out. Everyone is using AI, and everyone is using open source AI now—meaning every company will benefit from AI, which means every company will have more money to pay employees, hire more people, and train them up. I don't buy the narrative that Dario, Elon, and Sam end up with all the value in AI. I think the Chinese blew that up, and I think the open source stuff is the best thing to happen.

Is this an AI bubble?

Doesn't this mean there's going to be an economic collapse? Those data centers still need to exist to run the models—are we in an AI bubble in that regard?

If the cost of running a Chinese model comes down by some factor compared to a proprietary model, I'm going to run more models. As I do more, I'll need more compute and more data centers. That's where it goes.

On the data center argument, I totally buy people's concerns about data centers driving up the cost of energy. The water thing I think is BS, but on energy: every data center should be on its own grid and should generate enough power not just for itself, but to pump energy back into the grid. If you set that as a standard for every US data center project going forward, you've solved the data center problem—electricity gets cheaper for everyone in America, the data centers come off the grid, and everyone can own a piece of them. They can be in REITs, many are publicly traded, so they should be in your 401(k), and everyone can own a piece.

David, we have a closing tradition on this podcast where the last guest leaves a question for the next guest, not knowing who they're leaving it for. The question left for you is: what is one idea you have that is uncomfortable for people to hear, that you don't typically say out loud?

I think we should end the federal student loan program in the United States. Education costs would come way down, and the private market would still provide all the student loans that people need. Because the federal program doesn't discern between colleges, degrees, or individual performance, ending it would improve the quality of education in the US dramatically, drop the price dramatically, and give more people access to higher-quality education. I know it's counterintuitive, but that's what will happen. It's like when Milei removed rent control in Argentina and rents declined. We've propped up the price of education in the US with the federal student loan program.

If the federal program were replaced by one that is accountable—meaning the person underwriting the loan knows that loan has to perform, so the school has to be good or the degree has to be worthwhile—the market would correct itself. People would get a loan depending on the degree they choose: pick your basket-weaving major, and there's no loan available; pick a degree that's useful and likely to get you a job, and there is.

But for those people that can't get a loan because they want to do basket weaving, they're going to have to go find something else to do.

So they might have to go work somewhere else. The alternative was we put them in $200,000 of debt and then they didn't have a job afterwards.

Isn't the counterargument that the most privileged will get the best education—whether that privilege is intellectual or financial—creating a divide where some go to Harvard and others don't?

You still have a scholarship system. Scholarships don't change; they solve for that.

And education itself is changing with AI. But beyond that, a lot of these accredited universities provide a terrible education—and then they give you debt. You graduate with a degree from whatever crappy university, having assumed that because the government gave you the loan, you'd have a job at the end.

Yeah, it's a scam.

It's a scam. It's one of the great lies. Imagine if it was all privately funded. There's plenty of lenders—the money going into student loans is actually a small percentage of what goes into home loans—so there's plenty of bank capital to fund them. Student loan debt would suddenly become: if you go to this crappy university, we're not giving you $200,000 of debt.

I see what you're saying. It would also force people to pick fields linked to the employment market rather than, say, basket weaving. Maybe I'd be pushed toward things with a higher guarantee of return because of the private markets.

It's a retraining mechanism in that regard. And I think this is a great reset that's needed, because I don't think you can reform federal student loans—it's just too complicated. But if you shut it down, or transition it over time and let the private market underwrite it, you'd have a better path.

David, thank you—to you and all your co-hosts on the All-In podcast. It's one of my favorite podcasts, and I listen to it all the time; it really helps me develop my own perspective on the world. I appreciate that you all take the time to do it among your very busy schedules.

Well, thank you for having me. Hopefully we can continue this conversation once we get more of these answers and the fog clears a little bit.

Thank you so much. I'll link all of your links below for people to read your stuff. Is there anywhere else in particular people can get more of your content?

All-In, really—that's where I'm at. And follow me on Twitter, at Freedberg. I don't post very much, but that's it.

Thank you, David. YouTube has this new crazy algorithm that knows exactly what video you'd like to watch next, based on AI and all of your viewing behavior—and the algorithm says this video is the perfect one for you. It's different for everybody looking right now.