Vinod Khosla on the End of Jobs and the Future of Capitalism

[00:00:00] I've never understood why rational people believe in God. It just makes no sense. I didn't sour on philanthropy. I soured on the efficiency with which philanthropy works. Are you going to go all in for AOC in the president? I feel like you're much more of a... Don't your fellow venture capitalists want to murder you over this? Is Vinod Khosla in 25 years a socialist?

[00:00:23] He called technology as religion before AI existed. Now he thinks it can deliver something no religion ever could, material abundance for everyone on earth. Today I'm talking to Vinod Khosla, founder of Khosla Ventures, one of the earliest VCs to bet on open AI. I couldn't get my head around why he seems to support tons of liberal programs but hates AOC and Mamdani. We're talking post-capitalism, the Silicon Valley New Deal, and why this billionaire thinks a little differently from his friends in venture capital.

[00:00:53] I'm Eric Newcomer, author of the Newcomer Substack, where I write about startups and venture capital. Now let's get into it. I was looking at an old 2014 interview you did where you said, technology is my religion. And I think that dates back, you know, I think that was a take you started cooking up maybe in the early 2000s.

[00:01:20] And now we live in this world of AI. Obviously, you're the first venture capital investor in AI. It feels like AI could create a religion of its own. Yeah. Is technology still your religion? It is. And, you know, since I was a little kid, I've had an opposition to religion because especially in India, it's about manipulating people to do your bidding and really give you revenue.

[00:01:49] If I look at the Catholic Church, all the money it collects from all over the developing world is just criminal. It goes into gold accompaniments in the Vatican. It's a poor use of money.

[00:02:11] When you look at technology, it has the ability to benefit humanity in a way no other technology does. If I look at where we are today, probably 10% of the global population has a rich lifestyle. And by rich, I mean rich in housing, rich in health care, rich in entertainment, rich in transportation, rich in education.

[00:02:42] If we get to 100% of the population, being rich in everything will destroy the planet. If we do it the way we've done it for the top 10%, 10 times the number of cars, 10 times the amount of cement or steel or energy or everything else.

[00:03:01] The only solution that's necessary, it's necessary but not sufficient, is to have technology multiply resources. And in fact, I've long believed that and said technology can help.

[00:03:19] And with the advent of AI, which was not really a thing when I first said this, it's become so real to me that the world can be such a different place in the next 25 years. By 2050, we can have a very different world. There's so much we're going to entangle, I think, with this one idea, right? Because there's the piece of it that is improving quality of life really the main thing we can sort of deliver to people.

[00:03:49] Because I think we're also in this crisis of meaning. There's the climate element. There's the how technology, what it's like to be a billionaire 100 years ago and whether that's better. You'd rather be alive today or the super wealthy of 100 years ago. So there's so many pieces of it I want to get into and we will get into. Just the first, when you say, like, technology is a religion. You don't believe in God, right? I don't believe in God. Yeah, we share that.

[00:04:15] I grew up, my mom's partner was a Unitarian Universalist minister in Macon, Georgia. So I was very much, you know, the liberal moms in the South arguing with people about, oh, there's no God. I literally, in public school, have teachers on the board trying to, which I love, trying to make arguments for, you know, religion. Because it was, you know, so I share that sort of, you know, not believing in God. But I think at the same time, there's this crisis of meaning in the country right now.

[00:04:45] Do you think technology and the technology industry can really, like, give them, give the public sort of the meaning they crave? And it's just like improving their quality of life. The same thing is giving them an answer to this question. So those are very different questions. Right. I've never understood why rational people believe in God. It just makes no sense. I believe and admit there's some infinitesimally small probability God exists, but there's no proof.

[00:05:16] In fact, all the proof is to the contrary. But you still have to allow for a probability that's very, very low. But if you're rational, you believe in the highest probability things, which is no God exists because there's never been proof of God. Why do people believe in it? Because their brains have been programmed from age three or four to do this.

[00:05:41] And this, so, being religious is being programmed by an ecosystem. And so, my worldview is meaning shouldn't come from being programmed. Then you're a slave to God and open to exploitation by religion as a surrogate for God.

[00:06:05] I'd rather say, what are you programmed to do? And I won't go into the meaning of life, but… Well, that is sort of what I'm getting at, though, because I get… Certainly, I agree with you. We shouldn't even… Let me explain that a little differently. What should the brain be programmed to do? In the old world, there was religion.

[00:06:32] It was the only way to keep people under control. In fact, 500 years ago, you couldn't even get to a remote location. Maybe by horseback, it took five days. You couldn't enforce any laws. And so, you had to have the fear of God in people as a law enforcement mechanism to be good. Otherwise, God would punish you.

[00:06:57] So, maybe it even served a really useful purpose a long time ago of law enforcement was the purpose of the fear of God. Today, if I look at last 50 years, what do we program kids with? We program kids with, if you're five, go to school, do your homework, do well. You go to college, you'll get a good job.

[00:07:25] So, all of it is about income and supporting yourself. It's not about meaning. I think what changes dramatically is for today's five-year-old… Child born today. I had a grandchild born two weeks ago. You had a grandchild born. I had a grandchild. Congratulations. My first. I have a seventh-month-old daughter. So, I'm new to this game. But it's life-altering. It's an exciting time.

[00:07:54] It's amazing. It's very exciting. To me, I mean, meaning at the moment is reading a children's book. Like, if I had to ask directly, like, what is meaning in your life? Even though she doesn't really understand it right now, just having her pay attention and reading a book with beautiful art and rhyme and human language. So, that is a different kind of meaning. Whether you're a human being or an animal. Looking after our young is just instinctive.

[00:08:24] Otherwise, we wouldn't multiply as a species. So, what will my grandchild, when they're five, be programmed with? Hopefully, to be human. To care about the things that today people don't have time for. Looking after their kids. Looking after their grandparents. Parents. The relationships that matter. I think that should be meaning.

[00:08:52] Striving to be whatever you want to be good at. And this could be pretty random. I want to be awesome at surfing. I want to be awesome at singing. I want to be awesome at painting. I want to be awesome. Are you post-capitalist now? Because you're saying that the story recently was, do all the right things, climb the ladder, get a great job. Now you're saying, go surfing, spend time with your family. Like, is this post-capitalism? I'm saying for the first time, kids born now will be able to exercise freedom.

[00:09:23] Look at most of the jobs in this country today. They're not jobs. They're not meaningful jobs, even though people call them meaningful. If you're a farm worker in Salinas, in 100-degree heat, picking lettuce eight hours a day for 30 or 40 years, and then don't have retirement income, so you depend on your kids, that's not jobs.

[00:09:47] If you're an assembly line worker at General Motors mounting a tire on a car for eight hours a day for 30 or 40 years or till your back gives out, that's not jobs. That's servitude to survival. I think that's what it is. So do you think the American social contract, like if you're, I'm going to spend time with my kids and surf, should America provide you health care and the basic standard of living?

[00:10:12] Or do you need to still find, even if it's just 10 hours a week, something that the economy says is meaningful? Well, first, in 25 years, we can have great abundance. So people don't have to strive for a living. They don't have to do everything so they can pay their mortgage or support their kids or feed their kids.

[00:10:38] And it's sad when you have to do that, but it's necessity today, an economic necessity. That will go away, I think. All goods will become abundant. Almost every level, you'll have so much abundance. Distribution won't be an issue. People will still be people. People love competing. So today it's, I have a bigger house because I have a better job and I can do more.

[00:11:06] I think it'll be competing in the Olympics. It'll be competing in X Games or... You're saying that's where we'll get meaning or that's where we'll get money? That's where we will get meaning. Okay. Because competing is a human trait. Right. And we'll still need money to allocate like zero sum resources like houses. Yes, there are few and housing is the one area that is the hardest to see how it's dramatically impacted.

[00:11:34] I could see the cost of housing going down by 20 or 30 percent. And we can go down that alley later. It's hard because it's so much physical goods. But will we have shortage of copper or lithium? No. So is Vinod Khosla 25 years a socialist? I believe we will have so much abundance and depends on how you define socialism or capitalism.

[00:12:03] Socialism is not everybody shares equally. My personal view is everybody gets a base level of guaranteed lifestyle provided by the government. And that's the socialist part. But then they can do more whether they choose to pursue more income or their passion and happens to generate income. I don't think Beyonce got into singing for making more money.

[00:12:33] She got in because she has a passion for singing. But it also happens to generate money. So I do think people will follow for the first time. Kids born today will be free to follow their passion and be more human than they've been allowed to be in the face of economic pressures. I think that's the big liberation. I think that's the big liberation. AI will free humans to be humans and do what they come naturally. Spend time with their kids.

[00:13:02] Spend time to their parents. Spend time on their passion. I think there are a lot of people watching this that are going to say, why not cash in now? Like we have a technological abundance today. You know, we could produce the food that we need. Why not sort of create the utopia today? Why is it always sort of, you know, a couple decades from now? Today, there's a distribution problem. Okay. Because goods aren't cheap.

[00:13:27] You know, there's no question in a few years, even now, you could provide free primary care to every citizen in this country. In fact, every citizen off the planet, all billion and a half people in India could get free primary care, free… This is like a bridge open evidence. You're saying use AI tools? Yeah, I'm saying AI doctors, not AI tools. Tools are for humans to use.

[00:13:56] I think we are past that. Free AI tutors for every child. And I'm doing such a project in India to make free AI tutors for every child. It's a nonprofit effort. Free AI doctors for every person. And it's not only just the primary care doctor because the same AI is also a great endocrinologist or a great gastroenterologist, AI oncologist. Everybody gets that and it's almost free.

[00:14:24] It's cheap enough where one person can do it for a country like India. And I hope we can do that. I'm also doing an AI agronomist for the small farmers. So those things should all be free. Entertainment should be near free. The cost of production is going down dramatically. People talk about AI slop, but it's on the learning curve of improving at such a fast pace.

[00:14:53] Soon we will have AI that's tuned to each person that they love based on their watching history and maybe their rating history. So entertainment, legal services so people can access their legal rights, education. Education, healthcare. Care, those can all be free. Anything that relies on expertise will cost a dollar an hour.

[00:15:16] I mean, a key insight here is that things that don't require asking another person to do work for you, where something can be autonomously provided is something that can be a benefit versus, you know, sometimes the socialist case goes wrong when they're sort of insisting upon other people acting on their behalf and taking their time for themselves. In relying on the efficiency. In relying on the efficiency of people doing things because people have other motivations. AI will be just as efficient in government as outside.

[00:15:46] You know, it's something people haven't cracked. Efficiency comes from humans or inefficiency in government comes from humans and bureaucracy and all that. All that can be automated away. So my view, we enter this world of abundance. And back to your original question, where will people follow meaning?

[00:16:12] Because they will be programmed from age five to express their desires and be themselves. Whether it's their relationship with their parents or it's a relationship with their passion for sports or art or something else. I think that's, it's a very liberating kind of idea. It's very structured. It's very, you're sort of trusting the average person to create their like self story.

[00:16:37] I feel like a challenge of the Trump world is that people want these leaders that, you know, you know, fascism is appealing because it's like the strong man leader who you can project all these views onto. Like, who do you think like should be our society's like heroes today? Who are our heroes today? Right. Right.

[00:16:59] I would say used to be, you know, I like to say used to be only the basketball stars, right? The sports stars. And then when entrepreneurship really became a thing in the late 90s and it was existed long before then. But people got to recognize a Larry Page or a Mark Zuckerberg and they became heroes. So lots of people started thinking they could be, they could be an entrepreneur.

[00:17:27] I think the next big turn. First, most of corporate America should be four times more efficient, which means to get the same amount of goods and services, maybe one fourth or one tenth the number of people per enterprise. And what happens to the rest of the people? They can be micro entrepreneurs. How many people want to work at IBM four levels down at a desk?

[00:17:56] Or would they rather be their own boss? I think micro entrepreneurship becomes trivially easy because AI does your accounting for you. You're like, I'm your hero. I'm a solo owner. Now I have a small team, but yeah, small business owners. Yeah, all the whole creative phenomena is because it became easy to be a creator if you had some talent. Right.

[00:18:20] And so we've seen millions and millions of creators created almost overnight in the last couple of years. That projected forward many, many times. If you're a great baker and make the greatest bread, it's very hard to be in business. You have to get an accountant. You have to get a lawyer. You have to incorporate. You have to know marketing. You have to know sales. You got to know all these things.

[00:18:47] If AI does all those functions and there's startups that say, well, will the creator in a box where you're the creator, everything else is taking care of you. That will happen for every micro entrepreneur, whether you have a passion or skill at carving wood or baking or entertaining. It doesn't matter.

[00:19:08] I would guess by 2035, 10 years from now, that's a very short time, we will have one-fourth the number of corporate employees, maybe less. And we will have 50 million micro entrepreneurs doing their thing, being their own boss. And some people won't want to do that. It's actually a lot of work. Being a micro entrepreneur, it's easier to be an employee. You make more money.

[00:19:36] It's obviously amazing to be able to run a business and you can sign a big deal and make more money, but it is a lot of work. Back to the meaning of life. I don't think people are afraid of work. They're afraid of working for a boss who doesn't appreciate them. And most people in the country would say that's true. Most people in the world would say that's true. I think what matters is when people have agency.

[00:20:02] And that agency is being a micro entrepreneur. Doing your thing, you're the boss, you're responsible. People love taking responsibility. I think that's where meaning comes from, from more people. And their life will be more meaningful than working in a cubicle or an assembly line or on a farm. You've, you know, looking back at the 2014 article, I need to, it's like the business, it's a publication in India.

[00:20:31] I can't remember exactly the name. But you were early to be sort of soured on philanthropy. I think the tech industry sort of has grown more skeptical. Though now you're talking about some philanthropic efforts. I didn't sour on philanthropy. I soured on the efficiency with which philanthropy works. There's a big difference. I think there's great pleasure in giving. There is.

[00:20:57] But if it's really inefficient, then it doesn't appeal to people like me. There's plenty of people who don't care about efficiency and say- You just don't want a bunch of bureaucrats taking the money along the way to whatever cause- Consultants and people writing reports. Right. Because then you're effectively creating the misery you're trying to get rid of. It's fully LEEDS compliant. It's not LEEDS certified. Why?

[00:21:22] Because there was five consultants who wanted to have a take at certifying it on this or that. I hate that bureaucracy. So I'm not LEEDS qualified. I don't care. I'm happy to know how I built the building and how sustainable it is, or how climate efficient it is. That's all I care about. So I think people have motivation more internal than external.

[00:21:50] And they will do what they have a passion for. And they have much more agency. I think agency is the key to human meaning. You have agency to do what you want and the freedom to pursue it because your basic needs are taken care of. Instead of today's jobs, which are more than 50% of the jobs in this country, servitude to survival, everybody can't wait to retire. Yeah. Yeah.

[00:22:18] I just, I feel like the, first of all, people always shift sort of the minimum that they want to meet all their needs. That will happen. But, you know, I just do think- But groups will be infinitely flexible. The services you need will all be free and infinitely flexible other than token costs and power costs for tokens. Yeah. Yeah. I do think the challenge is, it seems like it's always, we're almost there.

[00:22:43] You know, it's like, I mean, looking back, I think you said in that 24 interview that I'm really latching onto, you said in 15 years, everybody would be able to get the quality of care you could get at Stanford. Yes. Which you still have time, but do you think that's going to come true? Absolutely. Absolutely. 2040? I had projected 2040. Okay. And I think it'll be way before 2040.

[00:23:12] I think this was 2029, but- I wrote a long paper on healthcare called 20% Doctor Included in 2016, which is when I published it. Okay. And I projected 2040. And my thought was- Somebody was like, oh, give yourself a little bit more wiggle room. Yeah. If you had a good advice. No, I assumed three eight-year cycles. Okay. In fact, I called it V1, V2, V3, all the way to V8.

[00:23:42] If there were three-year innovation cycles and it'd be 24 years or roughly 2040, there's not great precision in any of this. In fact, all of what I projected happening much faster. I won't say who. I won't say who. Somebody from the healthcare establishment, one of the better known names, has been critiquing me for the last 10 years for this projection.

[00:24:08] Came to me last year and said, he was wrong. He's a public health policy person. He was absolutely wrong. I was wrong, but I was right. But I was wrong in projecting it take much longer. So, I think he would say today. And we're doing a joint publication, so I won't. Yeah.

[00:24:32] So, publication in a medical journal together saying AI will do doctoring better than doctors. And we're already seeing some studies showing... We've seen lots of examples of that. And my sense is doctors are embracing it. I mean, they're overworked. You know, they want to get the right answers. They want to help people. Doctors are in a different business than they thought they would be when they were in medical school. Yeah. Most doctors got into medicine because they enjoyed taking care of people.

[00:25:03] Right? Today, they're slaves to billing. You're at a hospital. You're meeting billing codes, not what's the best patient care and empathy. Empathy isn't the characteristic that's optimized in medicine today. And most doctors got into it because of empathy for people who needed help. I think that's true of the majority of doctors, not 100%.

[00:25:30] And some percentage have still mostly work on empathy and a sense of service. So, look... Just a direct question. Do you support single-payer healthcare today? Today, absolutely. It'd be a fraction of the cost if we had single-payer healthcare and rapidly declining costs if we did that.

[00:25:53] I think where people are wrong is to assume that single-payer will be too expensive. The cost of healthcare in this country, the $4 trillion could easily cut... You know, and it's on a pretty strong inflationary cycle. That'll become rapidly deflationary if we had single-payer. Why can't you get, you know, the sort of right...

[00:26:21] You know, can you get Keith Raboy on board with that one? Or like, what's the resistance among sort of your fellows in the tech industry? I think there's a general assumption because there's no proof that single-payer will become inefficient and fraud and all those things. The thing is, if a doctor costs a dollar a month for every American, imagine it costs a couple of hundred million dollars a month to give free doctoring to every American.

[00:26:50] And that means primary care. It means gaps in care. It means chronic disease management. It means mental health for a dollar an hour. How much downstream cost would you reduce? Today, all the incentive, all the incentive was every hospital system. Start with Stanford or UCF. Buying small medical clinics. Not because they want to run them well.

[00:27:19] They want to feed off their cardiac patients. Are you going to go all in for AOC and the president? I feel like you're much more to the left than you even sort of see yourself. You support single-payer. You said you want capital gains taxes, income tax, which would be dramatic. Here's the problem with AOC. She has this model of the well-off against the not-so-well-off. That is not- You just don't want to be villainized.

[00:27:48] If you weren't villainized, you'd be okay with it? No. I think what I'm for is what technology enables in this country in the next 10 years. Basic services become so cheap, we don't need to meter them. I was on the board of a telecom, US West Quest, when I argued that it was costing them more to send the paper bill to people's homes than to provide the telecom service and stop charging by the minute.

[00:28:19] 90s, almost everything was built. Every phone call was built by the minute. And today, every medical visit is built by the visit. That should go away. I don't understand. How does this connect to the AOC point? If things are cheap, then you can give them. I still do think AOC doesn't care about economics and there's plenty of tax people. That's just the wrong approach.

[00:28:48] I think what's the right... And AOC and Bernie Sanders all say, let's not let people lay off jobs. Those are all the wrong things. Right. In the intermediate thing... The pro-union lock-in job is not a dynamic economy. It's just a really bad way to run an economy. It'll sink the economy before we get to abundance. If we free up the economy, and I'll come back to this point,

[00:29:16] and protect people against their biggest fear, which is AI will cause my job to be lost, we protect those people, then we can have a very efficient economy. All of corporate America getting much more efficient. The revenue per employee, every CEO I meet, I say, within five years, you could quadruple your revenue per employee.

[00:29:43] And every CEO should be striving for that. And it's very possible... Let's do it. I just really think... I know, we're wandering all over the place. No, no, I love it. No, no, I'm supporting what you're saying. I'm saying we need a Silicon Valley new deal. It's like you get single-payer healthcare. We reform the tax code. We say, oh, there's a minimum quality of life every American can have. But in exchange, we're not going to be super pro-union. We're not going to sort of jerk around companies.

[00:30:09] And it's going to be about sort of minimum government standards instead of sort of distorting capitalism and stopping companies from being able to be dynamic and quick pace. We should destroy capitalism. But we should recognize capitalism is by permission of democracy. You know, we will get a Mamdani elected. My biggest fear is we get a Mamdani elected. AOC runs for president. And there's this anti-AI backlash for fear of jobs that elects AOC.

[00:30:39] You know, look, Trump got elected because people were fed up with the nonsense of everything became a DEI issue. DEI went to extremes that it shouldn't have gone to. I can't use certain words. I can't even... Nothing Americans say more than being told what they can and can't say. Yeah. That kind of nonsense went so far, all because of the extreme left. And there's different constituencies, even in the left.

[00:31:08] But Mamdani, you know, I live in New York City, has taken a very pragmatic approach that to succeed in governing from the left, you actually have to deliver results. You know, he brags about the potholes. He talks about public safety. I really do... He doesn't talk about where he's going to get revenue. Yeah, well, he wants new taxes. I mean, I'm sure you would sign on to your Capital Gains one. He's trying to do get-a-tear taxes. Hold on. Hold on.

[00:31:34] There's rational reasons and irrational reasons to do this. Okay. For the next 15 years, so let's say to 2040, before we get to abundance in 2050, we want a fair economy. What is AI doing? It's disadvantaging capital, boosting the returns on capital. That's why you get so much more investment going in. It's suppressing the results for labor.

[00:32:04] So labor is $16 trillion of the U.S. GDP. That will keep declining. And I think we could... If we let systems be efficient and let every corporation do what they want to do without people embarrassing them against layoffs, we'd have it labor be, say, $5 trillion of the U.S. economy in real terms.

[00:32:31] That much decline in labor is a big, big deal. Right. Okay? Well, you agree with... And guess what? Who gets the benefit? Capital. Right. And what do we do in our tax system? We favor capital over labor. Why are ordinary income taxes much, much higher than capital gains? Because we've cut all these tax breaks.

[00:32:56] Every tax break, and there's 10 of them I could point to, is worth $30, $40 billion to the U.S. economy. My point is the things you're saying right now are things that bring you closer to AOC, Mamdani. And my point is that I believe the thing that turns you off most about them is that they villainize the billionaire class. But actually, you have a lot of policy agreement. No, there's not policy agreement. I'm not saying slow down layoffs like Bernie Sanders is. Right. You're saying smaller companies.

[00:33:26] Smaller, more efficient. Bernie Sanders has been very anti-AI. I feel like AOC, Mamdani, it's a little bit... Anti-AI, anti-efficiency. I'm not talking about Bernie specifically because I agree Bernie is that sort of old school sort of like union. But look, we can't constrain people from laying off people if you're going to have an efficient economy and compete with China. That is our big bugaboo. We have to compete with China globally.

[00:33:52] So, we don't want to constrain capitalistic efficiency. Okay. What we want to do is say all taxes, whether you earn it on capital or income, have the same rate of return and we will cut most of the tax breaks out. If we just did that, we'd have, my estimate, $400 billion a year.

[00:34:20] Don't your fellow venture capitalists want to murder you over this? This is like the number one thing the National Venture Capital Association has to fight. It's like keep carried... So, I'll tell you a funny fact. We should go back to that topic because it's a very important topic. But I've never been a member of the NVCA and they made multiple attempts to get me to be a member of the NVCA.

[00:34:43] And I've said for 20 years, if you will not advocate for carried interests getting special treatment, then I'll join the NVCA. If your principal job is to advocate for carried interests getting special treatment, I'm not joining the NVCA. I'm not a member. We're probably the only major firm that's not a member of the NVCA. I've had that principal for 20 some years. Amazing. I didn't know that.

[00:35:10] So, very few people know that because these principles matter. So, going back to this issue, if we equalize capital gains and ordinary income because AI is favoring capital over labor, then we have $400 billion a year or so to take care of people who are displaced

[00:35:38] by AI to allow people to be displaced, which I think is what capitalist efficiency is about. You need to take care of them. And, you know, COVID costs us about $250 billion a year when there was mass unemployment. That was roughly the annual additional bill for COVID unemployment. It's something like that. It will happen gradually over the next 10 years.

[00:36:07] It's not happening suddenly. And so, between equalizing taxes, I'm saying we have reverse taxation against, you know. I agree. I think many people are listening to this. I think it's a great point. And I get that it's right. One, which is being favored by AI, gets special treatment. Look, MLPs are worth $30, $40 billion a year.

[00:36:32] Step-up basis in estate planning and all this worth $30, $40 billion a year. Without even getting to any inheritance tax issues, there's all these tricks. A dozen of them worth $30 to $40 billion a year. MLPs. What is MLPs? Master Limited Partnerships. So, oil and gas or housing. And like there's all these things that have been engineered through political influence

[00:37:01] to favor people but have nothing to do with capitalism. They have nothing to do with capitalism. There are special perks engineered in because I could afford to make a contribution. The Republicans are never going to fix this issue, right? Your only hope of fixing this issue is aligning with Democrats. Do you agree with that? I think it's an independent who takes a new view. The United States is set up, you know. So, by the way, I've never been a Democrat. I know, I know, I know. I'm an independent.

[00:37:26] You're a free marketer who became more sort of because of environmental policy that moved you a little bit left. I was a Republican most of my life because I thought fiscal was the most important. Then I added climate to this. And the early Republicans, both the Bushes were pre-climate action. It's a long-term risk, just like homeland security is a risk.

[00:37:52] We should spend a certain amount on risk insurance, whether it's for climate or homeland security or national defense. That's my view. But that's not where the Republicans ended up. So, I became an independent. But back to this issue, if we could do equalized taxes, get rid of special breaks, we'd have $400 billion a year.

[00:38:14] Add to that a token tax intelligence, which is what AI is, you get to a trillion dollars a year that you can use to support the people who would be displaced. So, instead of slowing down the economy and saying, stopping displacement, let's let the economy be efficient, but let's find other ways to take care of them and encourage things like micro-entrepreneurship. Do you think OpenAI is going to align with this?

[00:38:44] I mean, they put out with Chris Lehane a policy paper. It was pretty progressive. But then they're also supporting some, you know, I don't know, anti-democratic, you know, efforts. So, I'm sort of confused where they're actually going to spend their energy. Well, I would guess. And hopefully, they'll call me. Sam would support token taxes. I've asked him. I suspect Dario would support token taxes on intelligence because they both see the impact

[00:39:12] it can have and the good impact as well as ways to reduce the negative impact of AI. A subtext of what you're saying for the listener is that we're going to have a new class divide problem, which is intelligence is going to be something you can easily buy. And that reinforces the power of capital even more because you can spend money and get intelligence. And that further advantages people with access to money. So, Aerosene will be happy with this.

[00:39:36] I suspect in this world, a farm worker gets paid the same as an oncologist because they're both upskilled to the level of the AI. So, it's the capital. It's the money holders that are at the biggest advantage? I think it's probably by 2050, we have the greatest opportunity we've ever had to have equality. What happens to a few percent of the population who own the AI? One of the great human problems is, you know, who gets to live on Central Park or whatever.

[00:40:04] I don't know if we saw, just because we have abundance, do we care more about abundance or do we care about relative status? So, I do think in a world of great economic efficiency and AI-based abundance, we start to pay more attention to equality, which we have not done under capitalism, and say, we'll sacrifice

[00:40:28] some minor level of efficiency to get much more equality in society, and society, I suspect, will be happier. More equal societies are happy, both in the rich people and the not-so-rich people. I want to change gears slightly. I mean, and it's sort of already come up, but one of the things you want to animate the United States right now is sort of a rivalry with China. And part of what you're using to justify that is the sense that America is a great democracy,

[00:40:58] we should protect democracy, China, you know, is a totalitarian regime. You know, I agree, but it feels like the American democracy is under threat right now. Like, what is your sense of how, you know, Trump is literally joking, and I know you're anti-Trump, but Trump is literally joking, joking, you know, that he'll be president in six to eight years still. Like, what is your view on how secure American democracy is today? Look, there's a great book on this.

[00:41:26] I'm trying to remember, by Russ, I think it's Russ Feingold was the senator. That is the senator. I'm not sure what book you're talking about. He wrote a book on democracy in peril or something like that. I buy his thesis that we are not as safe in democracy as we'd like to be. Institutions are no longer safe. We have the era of individuals like Trump.

[00:41:56] So I don't think we are safe. I don't think Western philosophy is safe. Western values are safe. And we have to fight for that. And that's why I think winning against China is so key. What political philosophy the world adapts will depend on who wins the economic race, and that will depend on who wins the AI race entirely.

[00:42:22] So that's why it's so critical to win the AI race, why it's important we win the economic race and bring benefits to the rest of the world with our political philosophy. Do you see a distinction between Western values and Western democratic values? Are you drawing a distinction there? Not really. I mean, you can get into endless... Right. Are you still largely... Like, you were pretty opposed to open source a couple of years ago. Is that still true?

[00:42:51] I'm not opposed to open source. So remember, I was at Sun when we started the open source movement. Oh, but AI open source. NFS was the first open source product. I feel like Sun was responsible for making open source happen. And starting with the NFS file system and then open source operating systems and all that.

[00:43:19] And I was a huge investor in GitLab when it was four people. It's open source and... But today, are you comfortable with your portfolio companies using Kimmy and DeepSea? I'm not. Okay. So I'm not because I think China, we are too open to subversion. And when we open source... And I had this debate with some of the executives at Facebook. When they were open sourcing Lama, they were feeding technology to the Chinese.

[00:43:48] And it has helped them catch up. Now, China is a smart population. They'll catch up anyway. But we shouldn't be helping them catch up. And we should stay at least a hair's breadth ahead of them. Because I think our fundamental system is superior. And we have the talent. Now, they have the talent too. But many more of the most brilliant minds on the planet want to move to the U.S. or did before Trump.

[00:44:17] Instead of moving to China. And so I think that was one of our strongest assets in staying ahead. And Trump has killed that, to be honest. It's terrible. The biggest damage he's ever done to this country long term is making it less attractive for the best talent in the world to come to this country and keep us ahead. And it's insane that Elon is one of his biggest, if not biggest, funders. And Elon's companies are full of immigrant founders and is highly dependent.

[00:44:46] He's an immigrant himself. There's hardly a successful company in Silicon Valley that doesn't have immigrant founders. And in some key role in the top management. And it's one of the things that makes America great, in my view. I mean, if I'm really pressed on, I asked you who the heroes are.

[00:45:08] I mean, some of the people I think about are like, you know, Sundar at Alphabet and like Satya Nadella at Microsoft, who have been great stewards of important companies that it feels like through a terrible culture shift, you know, are real stewards of institutions that provide great value to Americans. They've done a great job. And immigrant founders have had, immigrants have had a huge role in the knowledge economy. In fact, immigrant doctors. We'd have a huge shortage if we didn't have immigrant doctors.

[00:45:39] So, no matter where you look. Do you try to talk to Elon about it? You talk to him occasionally or not much? Not a whole lot now. He used to be like, oh, I have this space company or you're on the outs at this point. Look, I'm not afraid to call him when it needs to be called. Look, the nice thing is I can't be fired. So, I don't worry about it. I never worried about a career because it's just done fine for me.

[00:46:05] I've never cared about the things other things fear, which is somehow… I don't imagine you want to be on his list of Twitter people. He just like spices randomly or like you've avoided the mass public sort of Republican animosity. Yeah, I'm probably avoided that. But, you know, look, I admire Elon for what he's done well. As an entrepreneur, he's done a great job.

[00:46:33] We wouldn't have electric cars if we had left it to General Motors. We wouldn't have had space launches, though I'd say Rocket Lab or one of our companies deserves as much credit if it wasn't for Elon and we had relied on Boeing or Airbus or somebody else to do. This is a business question, but do you think the SpaceX IPO is going to take down the whole mania happening right now? I don't know. It just… The expected, you know, multiple on revenue is like ludicrous.

[00:47:03] I don't care about Wall Street. I don't care about prices. The reason I don't care is people think that's reality. It's not reality. When people talk about a bubble, I say a bubble in what? Was the dot-com thing a bubble? Not if you look at real usage of the internet. Look at internet traffic and tell me… They just overpriced where it was going to go.

[00:47:29] Whether investors bounce between fear and greed and that causes change in valuation, that's not reality. It would kill a bunch of good companies probably if the bubble popped. You know, it's just brutal. You can do nothing about investor bouncing between fear and greed, which is what causes bubbles and busts. So, I don't worry much about Wall Street. I don't trade almost ever.

[00:47:59] I just do my thing with companies. And frankly, I don't… I've never in 40 years said I'm in the investing business. I've never called myself a venture capitalist. I know. What's fun is helping entrepreneurs build companies and helping them with mistakes you've already made. I've made more mistakes than anybody on the planet when it comes to building companies because I've been doing it for a long time. And I hope I can help entrepreneurs avoid some of those.

[00:48:29] That's my job. That's how I define what I do. And it's a lot of fun. We're recording this, obviously, as Sam Altman and Elon Musk are fighting in court right now. You were the earliest venture investor in OpenAI. Though, Reid Hoffman was in the round but as an individual, right? Or you guys… Who got the bigger shirt? That's right. Yeah. What is…

[00:48:53] There was a viral tweet that you sort of knocked down about the returns you've seen on OpenAI. Will you say anything about how well you've done? I think the worry is that it's been just so dilutive that what should have been sort of an all-timer has gotten diluted. It's a fabulous return. I won't go into specifics, but it's been a fabulous return by any metric. Have you sold some along the way or…?

[00:49:20] Very tiny amount a long time ago, most of it internal between our funds. So, you know… Do you think it will be one of the all-time great investments? I think so. Now, my all-time great investment was Juniper Networks where we got 2,500 times our money on a $3 million investment in a world where the largest funds were $200 or $300 million. That's not happening.

[00:49:48] It was sort of a fabulous return. Why? Because we bet on TCPIP being the core of the internet when everybody else was betting on another technology called ATM. Because the telcos were thinking ATM and Cisco was thinking ATM. Cisco's CTO told me they would never do a TCPIP router above OC12, which is less than my home service today. So, we just built Juniper to be TCPIP for the public internet.

[00:50:18] So, that's the best investment. But obviously, you'll make far more money off the volume of open-net, but the level of return won't be the same. I hope it's 2,500x. It could be. What do you make of this idea that they stole a nonprofit? Or obviously, you were investing when it was becoming a for-profit. You know, Sam went outside because Elon failed to fulfill his commitment to fund the company. Last minute, there was no real source of funding.

[00:50:46] So, Sam had to go looking and perfectly reasonable for a nonprofit to have a for-profit subsidiary. Look, Nordisk is set up. Many European companies, IKEA is set up that way where a nonprofit owns a for-profit. And has for-profit shareholders. This is not the first time it's been done.

[00:51:11] In fact, the best companies in Europe have this structure of a for-profit subsidiary with for-profit investors. But the majority is still owned by a nonprofit. It depends on how much money they've raised, how fast they've expanded. It's been a pretty fabulous story. I mean… What is the alternative? Right. I'm sympathetic to the idea to build this AI company. You needed the money.

[00:51:38] The only way to get money on that scale is from people trying to make a profit. Therefore, you need to be a for-profit if you believe in the mission of delivering AI. Capital formation is a key function of capitalism. That wasn't possible. Look, Elon is just jealous. It's nothing but jealousy that… It's amazing how much he's just like desperate to be in control. I mean, you know… He wanted to be in control. He wanted to own… There's lots of things I won't go into.

[00:52:07] He didn't have the best intentions. He definitely didn't have nonprofit intentions. He doesn't even care about climate. It worked out well for him at Tesla. That wasn't his pitch originally. Should we still worry about climate? Nobody talks about it anymore. We absolutely need to worry. Just because Trump had made it bad to talk about climate doesn't mean the problem has gone away. We've just buried our head in the sand.

[00:52:35] We are continuing to invest in climate. Now, our view is climate technologies have to be competitive. We have the best steel company. I'll beat a coal-based steel plant any day. I can build a… Beat a coal-based cement plant any day. So these fundamental technologies have to be cost-comparative. You can't defy the laws of economic gravity.

[00:53:02] Our climate philosophy is more about new technologies that are cost-comparative with their fossil competitors. Fusion will be that way. We did fusion, cement, steel… So in pro-energy, you need energy. It's not about sort of being, I don't know, like monks and living without abundance. So this is another place, you know, people… Efficiency legislation consume less. That's the wrong way to approach climate.

[00:53:27] Make it cleaner and cheaper with sustainable technologies. That's the right way. That's what fusion will do. That's what super hot geothermal will do. We work in both those areas. In fact, have the dominant companies in both those areas, the most cost-effective companies. Same with cement. Our cement is cheaper than regular cement. Our steel is cheaper. Now, we have to build large plants to get to those cost points.

[00:53:54] And capital is hard to get in some of these areas in the Trump climate era. But most people would agree with me that that's the way to scale sustainability, not by saying it's okay to pay twice as much. Are you cheering for Anthropic? I cheer for all AI, including Anthropic. What do you make of, you know, I feel like the Trump administration in places has made me more of a libertarian.

[00:54:21] Like, you know, I guess during Biden, I would have been a little bit more comfortable with the Biden administration regulating in some way, at least wanting to be alerted to what's happening with different models. Now, the Trump administration… Nuance matters here. We can't have private values determine what our national policies on national defense are. We can, in fact. Everybody in government, no matter who they are, are required to abide by the law.

[00:54:51] They don't always, but they're required to abide by the law. So, if a government agency is abiding by the law… But it's the Trump administration, an administration that has no history of abiding by the law. The version of abiding by the law is what we can get away with until, you know, they arrest us in 20 years. I agree with that. My fundamental problem with the Trump administration is lack of values. So, why… So, saying let's just defer to democracy is sort of laughable.

[00:55:19] Democracy includes, first, number one, the main branch is supposed to be Congress, which has done nothing to react to AI. So, to say defer to democracy, well, our main branch of government hasn't done anything. Let's look at the principle. A private company can choose to sell or not sell. That's always been true. But that's what… And Theropic was saying, we want… We'll just walk away. The issue was that it was a new contract. My tweet was, I admire Dario for his principle.

[00:55:47] I happen to believe that particular principle was wrong in trying to constrain the defense department. I think building a constitution into law teaching AI to have values. I mean, to go back to sort of technology as a religion, like one of the commendable things they're doing is being thoughtful and intentional about the technology they're building. And in a free market society, they are one company that is allowed to build it in their way. I think they're doing it in a very admirable way. And I don't think we should encourage the federal government to bully them into doing what they want.

[00:56:16] I think the federal government, the Department of Defense or Department of War, whatever it's called now, has one objective. I believe it's set by Congress. So I think the Department of Defense is still… But anyway, go ahead. Not lose to Putin or be subject to Putin. If Putin walks into China, Xi walks into Taiwan, they need to be able to defend us and the Western

[00:56:46] world against that. Okay? I think that discretion shouldn't lie with a private company, but they should have the ability to not sell their technology to the Defense Department or Department of War. Well, that's effectively what they're doing. They're saying we're not going to sell it unless you agree to this contract. And that's why I admire Dario for sticking with his principles. Right. But then you sort of bought into this idea that if you're following the law, if as long as they agree to follow the law, that's sufficient.

[00:57:16] But you agree with the idea that they're not a set of people that can be trusted to follow the law. I agree that's true. I hope it's very temporary. Hopefully, another two years and we are done with this problem of defying the law and we go back to democracy. Corruption matters, right? I mean, as a business person, what do you make of sort of just the, I don't know, rewarding… I mean, the crypto has become purely about… Crypto is huge.

[00:57:45] Does anybody know what happened to TikTok? Senator Cantwell was on the committee that decided to ban TikTok or not. I talked to her multiple times to convince her TikTok should be banned. She was against it and then came around to believing it should be banned. There's a couple of Democratic senators who were key to that. But now it's disappeared into the abyss.

[00:58:13] Nobody knows who made what and who owns it, who's in charge. Fascinating. I haven't even really… You're right. Nobody knows. She was asking… There's so many things it's hard to keep track of. The senator on the committee voting for the ban had no idea. She asked me recently, do I know what happened to it? I have no idea. She has no idea. Maybe I shouldn't have an idea. She should. Insane. There's no accountability left. Corruption is just rampant.

[00:58:43] Are there any politicians that you're optimistic about, whether it's the California governor primary, Democratic primary coming up, independent? Any politician of any stripe that you're like, oh, I have some optimism about them? Let me not name names. I do tell you. Look, in the California primary, I'm supporting Matt Mahan over any of the other candidates. I like him. I had him on this podcast. Nobody watched it.

[00:59:11] You need to be a little bit dynamic. I mean, I like him. I was set up… The problem with mainstream media… But it's the public. It's the public. Yes. Right? Why did Trump get away with saying, Haitian immigrants are eating your pets? Not that he believed it. Not that maybe the people in Ohio believed it.

[00:59:37] It just enforced the idea that immigrants are bad, which was his campaign pledge. So, he's a great communicator in that sense. Even if it sounds irrational to some of us, why would anybody believe that? How can a person be president who says that?

[00:59:56] I mean, but I hope we restore a sense of democracy and law-abidedness in Washington. Yeah. I mean, I like Matt Mahan. I think he's smart. I think he's a little… He's in a weird position where he doesn't want to talk about Trump too much because he needs to somehow get Republicans. But I feel like that's a hard path to win. And it seems unlikely. I mean, yeah, California governor's race.

[01:00:25] Do you think the billionaire attacks… You've been less worried about that than some, right? Look, I think it's really bad policy. I critique Ro Khanna over it. I think he's taken less of a stance in favor of it since. But it's a really bad idea. We've, in essence, reduced our tax collections over 10 years by all this wealth leaving the state. That would have happily paid ordinary income tax as they sold their stock.

[01:00:56] Wealth taxes have lots of problems. I'm not against more taxes, but wealth taxes are a bad idea. Net-net will reduce tax collection in California. The last question I'll ask you, you know, you're investing in all these great startups, like companies that people haven't heard about. By the way, I just want to say, I chose to stay in California and I'll stay in California

[01:01:21] despite the tax because my lifestyle is more important than the 5% the wealth tax might. Well, the real issue is sort of the paper, you know, it's like if it turns out that control over Google is like super, is valued as… Well, there's lots of other problems. Even the goal of tax collection, tax collections over 10 years will go down, not up. Because they're scaring people away. Because they're scaring people away.

[01:01:47] And scaring away the people who created the wealth, who should be rewarded for that wealth. But no, and nobody on the presidential level that you're willing to say you're excited about? I think it's too early. Do you think you'll be a big donor? If there's an independent candidate, for sure. I think right now to win in a Democratic primary, it's not a position I'm likely to support.

[01:02:20] Startups that people should be paying attention to right now. I know you have so many. I know you have so many. Another competitor to OpenAI or Anthropic. I think that business is done. But are there more innovations to be had?

[01:02:51] Absolutely. Huge gains in efficiency. Everybody knows inferencing is really inefficient. We are using the same GPUs to do both. Inferencing and training. Base 10 together, fireworks? Do you have a bet? Or what is your inferencing bet? Yeah. My bet is, so I won't talk about specific investments. Coming down the... Yeah. We have plenty of ways to make GPUs much more efficient.

[01:03:21] Have alternatives to GPUs. Transform models that are alternatives to GPUs. We just announced, so one announcement was Prism ML that reduced an 8-bit model to a 1-bit model so it can run on a phone. Like, that's innovation. Right? A mathematically equivalent model that can run on a phone. That's fun. There's other ways to do those things. So lots of room for innovation around LLMs.

[01:03:51] Lots of room for alternative to LLMs. World models is a really, really interesting area and lots of fun things happening. And a lot of diversity in approaches. In how world models and robotics are being approached. Big area for innovation the next five years. Outside of here, I've talked about some of the climate areas.

[01:04:17] For the first time in a very long time, I'm bullish on quantum. Is there a particular company? Won't talk about it yet. I've looked at a dozen quantum startups over the last 10 years. Most multiple times. I've never had the confidence we'll get to quantum in any reasonable time frame. I think we have finally have a shot. Well, Vinoa Costa, we started talking about religion.

[01:04:46] You're, you know, by 2040, you need to help deliver us material abundance. Make sure Americans have, and everyone around the world has quality health care. No pressure. Excited to watch your startup investments. And I'm sure we'll have you back on the show soon. Thanks for coming on. Thank you. Lots of exciting things to talk about. Yeah, I loved it. All right. Thanks. That's our show. Thanks for listening. Please like, comment, subscribe. We're a new channel. We can use all your support. Go watch some of the old videos.

[01:05:12] You can follow along on the sub stack, newcomer.co. Or if you've got endless time on your hands, go watch the Cerebral Valley Show, my chat show with Max Child and James Wilserman. Thanks for watching. See you next week.