TL;DR

Marc Andreessen and Ben Horowitz believe that the unique value of startups is that they drive innovation, while monopolies hinder innovation. The success of technology and startups is critical to the United States' technological advantage, especially in the fields of blockchain, cryptocurrency, and artificial intelligence.

They discussed the importance of blockchain and cryptocurrency, arguing that these technologies can reshape the way society works, especially in terms of controlling information and data. They believe that blockchain can help achieve a fairer capitalism where creators can receive higher returns.

The Biden administration's regulatory strategy in the cryptocurrency sector has been criticized as a crackdown on innovation. Andreessen and Horowitz believe that these policies are not conducive to the development of startups and call for clearer and friendlier regulations. They expressed concerns about the Biden administration's regulatory measures in the technology sector, believing that these measures may restrict innovation. Especially in the field of artificial intelligence, the Biden administration's policies may hinder the United States' position in the global technology competition.

· They tried to meet with Biden but were rejected, and the recent meeting with Trump gave them a deeper understanding of his personality. Although he is a complex figure, his courage in the face of crisis is impressive. In the current policy environment, Trump is more suitable than Biden to promote the development of technology and start-ups. They emphasized that this is not based on personal preference, but based on analysis of the actual impact of policies.

· They also discussed QSBS and the billionaire tax. QSBS (Qualified Small Business Stock Exemption) is a key tax policy to encourage entrepreneurship and investment. The Biden administration proposed to abolish QSBS, which will have a serious impact on startups and venture capital, and may hinder innovation and economic growth in the United States. Biden's proposal to tax unrealized capital gains was called "the last straw" by Andreessen. This tax policy will force startups to pay taxes on the growth of valuations, which is not feasible for private companies without liquidity. This policy will be a devastating blow to the startup and venture capital ecosystem.

The following is BlockBeats’ compilation of the content of this podcast:

The importance of startups

At the intersection of technology and politics, the challenges facing startups are particularly complex and important. In this installment, Ben Horowitz and Marc Andreessen explore the unique value of startups and why they think Trump is the right choice to support early-stage tech companies.

Ben Horowitz: There are going to be a lot of friends who are going to be mad at me for saying nice things about President Trump. You know, the last thing I want is for our company, our employees, and the companies that we invest in to be involved in this, because it's going to be very emotional. It's hard. But the reality is that the future of our business, the future of technology, the future of new technology, and the future of America is at stake. We're here. For startup tech companies, we think Donald Trump is actually the right choice. I'm sorry, Mom. I know you're going to be mad at me for this, but we have to do this.

All right, today we're going to be talking about the startup agenda, and specifically the impact of the upcoming presidential election on startups, and how we think about it, and how you think about it, because things are about to get very real for startups and the presidency. Sometimes I feel like we focus too much on the negative in politics. It's good to know a little bit about that. Maybe we should just wrap up a little bit.

One of the reasons we do this podcast is that there are people who know more about foreign policy or rights issues for various groups than we do, or economic theory or various other things. But when it comes to startups and the policies around them, there are probably only one or two people in the entire world who are at our level because we do it every day. We work with startups. We see the impacts of policy. And then we do a lot of work in Washington. And that's why we do this podcast. So, listeners, you know, we spend a lot of time with senators and representatives from both parties. Marc mentioned that we met with President Trump. We did meet with White House officials, many White House officials, including Chief of Staff Jeff Zeitz and National Security Advisor Jake Sullivan, Commerce Secretary Gina Raimondo, and others.

We haven't met President Biden. We tried and failed. But we have a pretty good sense of working with the administration. We are really deeply embedded in politics.

We're focused on one issue, the startup agenda. We're pretty bipartisan on that agenda because we donate a lot of money to both parties. We have great people on the Democratic side who are very pro-startup. We have great people on the Republican side who are very pro-startup. We're just focused on that. Marc, maybe you want to talk about your background and your story in politics because it's different from the last time we talked about your experience inventing the browser. You got into presidential politics at a young age. I was going to say, obviously you have to be a Democrat because you have to be a good person. That's the root cause.

But in particular, successful business people can essentially be successful philanthropists. That's the path that Gates and many others have pioneered. And then you can be progressive on social issues, you can be on the right side of these social changes. People were paying attention to that at the time. It all seemed completely obvious and completely, completely easy. So I went down that path, and frankly, it lasted all the way up to 2016. In retrospect, I would say that there was a beginning of anti-tech, anti-business sentiment growing probably in the early 2010s. And one thing that bothers me, by the way, which we probably won't discuss today, is the growth of anti-philanthropy sentiment.

Marc Andreessen: I don’t usually talk about this unless it’s relevant to the presidential campaign that we’re talking about in this podcast. It’s relevant for people to understand where I’m coming from, where we’re coming from, at least in terms of the origin story of thinking about these issues. I’ve had an unusual career and life in many ways. One of them is because of the things that I was involved in early on. The first time I met a president of the United States in a business context was when I was 23 years old, 30 years ago, in the summer of 1994, when Clinton was president. Bill Clinton and I did the first presidential webcast together. I know a lot about technology. I was the third participant in the first presidential webcast, when Bill Clinton was in the White House and I was in an auditorium on the Google campus. I met Bill in 1996, I think. And then Al Gore, I obviously knew Al Gore, I talked about this last time, Al Gore was very important in creating the Internet and making all of my work possible.

I knew him well. I knew them, interacted with you, worked with you. And, they were right. By the way, there were all kinds of political issues along the way. The crypto wars were actually a big issue at the time. There were a lot of interactions with the White House, and ultimately they made the right decision on that. I knew them very well. I actually supported Clinton politically in '96. I supported Al Gore in 2000, and I was very disappointed when he lost. And then, through all of that, I also knew Senator Kerry, John Kerry, and supported him in 2004. I met Obama, and I have Dan Rosenzweig to thank for bringing me to meet the early Obamas. I wrote a beautiful endorsement of him in 2008. You can still find it online now.

I publicly supported Hillary Clinton in 2016, and I've never been as deeply involved in this politics as I am today. But I've always been involved, in and out of it. In the 1990s, I was a Generation X, I came of age in the 90s as an entrepreneur. You take it for granted, like almost everyone I know, myself included, that of course you're a Democrat. Of course you support a Democratic president. The answer is, because that basically solves all the basic formulas. The formula comes down to a simple answer, which is that Democrats in those days were pro-business at the presidential level. They were pro-tech, pro-startups. Pro-America winning in the tech market. They were pro-entrepreneurship. You could start a company. They were pro-business. You could be in business. You could be successful in business. You could make a lot of money. And then you gave that money to charity, and you got tremendous recognition for it. And, that solves all your problems. I was going to say, obviously you had to be a Democrat, because you had to be a good person.

Marc Andreessen: It was an underlying idea. But specifically successful business people can essentially be successful philanthropists. That was the path that Gates and many others had pioneered. And then you could be progressive on social issues and be on the right side of these social changes that people were paying attention to at the time. The whole thing seemed completely obvious, completely simple. So I went down that path and it lasted until 2016. In retrospect, probably in the early 2010s, anti-tech sentiment and anti-business sentiment had begun to emerge. And then one thing that bothered me a lot, which we probably won't discuss today, was the growth of anti-philanthropy sentiment.

Ben Horowitz: But, oh, people who make a lot of money and give it away, get criticized for giving it to charity instead of paying more taxes. I just want to name a specific moment where it was like envy was pushed to the extreme.

Marc Andreessen: The specific moment that made me realize that this was changing was when Mark and Priscilla Zuckerberg founded the Chan Zuckerberg Initiative. They committed to giving 99% of their assets to the Chan Zuckerberg Initiative. But there was a political faction that was very critical of them. The theory was that they should pay taxes and the government should distribute the money. They shouldn't have control over where the money goes. Another criticism was, "Oh, they're just doing this for a tax cut."

Ben Horowitz: That's not true. That can't be true because you can't give away 99% of your assets and get a tax deduction.

Marc Andreessen: You give away 99% of your assets for a tax deduction. It just doesn't make sense. It's like people are talking behind your back and being jealous. That's exactly what it is. The formula is starting to break down. Like a lot of us in tech, over the last eight years, especially over the last four years, it's become a much harder puzzle to solve.

Ben Horowitz: We both have come to realize that it's a lot more complicated than we thought. And I've had a similar experience, or maybe a weirder one. My grandparents were communists. My father, I always call him the Tiger Woods of communism, because he was raised to be like the greatest communist of all time. He was the editor of Ramparts magazine, which was a very left-wing magazine. He did a lot of work with the Black Panthers in the early '70s, and that had a big influence on me culturally. So a lot of my cultural references come from that era. And then I grew up in Berkeley, which was basically a communist town. And all of my friends growing up were of course very, very left-wing. And then my mom is still very left-wing. But my dad eventually turned to the other side, and if you're interested in this kind of extreme right, you can read all about him. I grew up hearing both viewpoints debated very fiercely, and I still do to this day. But it wasn't until this process that I actually really understood all the details.

And there's a huge difference not just between Democrats and Republicans, but there's a huge difference between politicians, especially when it comes to tech. That's one of the big issues. We've met with a lot of forward-thinking people on both sides now, but there are also people on both sides who don't understand anything. And there are people who have, you could almost say, dark agendas that are almost inevitably related to what I call control. Can they control power? When you get into politics, you get into power, and if people pursue power by hook or by crook, that can be very, very dark. Okay, so let's talk about, why is startup tech even important to the United States? We have so many pressing issues, like competition with China, potential war; our war with Russia; our war in the Middle East; we have all these issues about equality and fairness. So why should anyone care about tech, let alone startup tech as an agenda, that should affect how they think about policy or politics or who they should vote for?

Marc Andreessen: Okay, I'll break it down into three parts. First is why is technology important? Second is why is startup technology important? And then which startup technologies are important? So, is technology important? My point is that if you look at the long history, we Americans have been lucky in many ways. And one of the ways we've been lucky is that our country has basically been the premier or dominant country in the world since World War I. It was actually the United States entering World War I that really turned the tide for the Allies. And again, the United States entering World War II turned the tide for the Allies in that war. And also during that period, the United States was also the leader in what's called the Second Industrial Revolution, which is basically the promotion of all of the modern technology that we live with today. Basically everything from cars to railroads to telephones to logistics, and our air conditioning, everything that makes our lives livable and comfortable, came through that era. The United States was far ahead in all of that. And we continued to lead after World War II. We were the dominant country on the planet for the next 50 years. And that proved to be very critical to the Cold War. Because there was a hostile ideology of full capitalist communism incarnated in the Soviet Union. People who didn't live through that time don't feel that way, but it was a life-or-death situation. I grew up thinking that there was a high probability that we would all die in a nuclear war until it was over.

Ben Horowitz: We used to have to do those drills. Remember when we were hiding under the table?

Marc Andreessen: You hide under the table. When the bombs are falling, hide under the table, because it will help. This has been a very real thing for my generation and older generations since the late 80s. Basically, for the last 110-plus years, the United States has been the most important country in the world. You can make a lot of criticisms about what the United States has done. But in terms of basic economic development, technological development, military power, and again, by the way, American foreign policy, you can have a long discussion about this. But look, the last 70 years have been an unprecedented period of global peace relative to any period in history. A lot of that has to do with American technological and military dominance. We're all used to that. And frankly, we take it for granted. We inherited it and take it for granted. But it's not necessarily the case, and it hasn't been the case before. The world used to be more divided and it used to be more violent.

Ben Horowitz: I mean, it didn't last very long. But that's the thing, it could easily change again.

Marc Andreessen: It may change, indeed. We live in this world. We think really, very seriously, that we don't want to live in some other world. We want this to continue. There's a very popular narrative out there that America's best days are behind us. We've lost control and so on. And, so far, that's not true. But, it's still up in the air.

Ben Horowitz: And the United States still has real strengths. People underestimate that. It's a little bit like with politicians. It's easy to see all the things that are wrong with the United States. But we take for granted what's right about the United States.

Marc Andreessen: I call it the triangle. There are basically three areas where we are the best in the world. And those three areas happen to be very important areas. Those areas are technology, economics, and military. If you think of them as a triangle, it's interesting because each one depends on the other two. This even predates Silicon Valley. This actually goes back 100, 110, 120 years to the creation of any modern warfare system. Including places like ancient Japan.

We have had technological dominance, technological advantage, throughout this period. The fact that we have technological dominance essentially gives us economic dominance. The cornerstone of American economic success over the past century has been gained in wave after wave, going back to the second industrial revolution, continuing through the computer age, the internet age, and now the age of artificial intelligence and all the other biotechnologies that go along with it. So technology and economics are deeply intertwined. You can only have the world's leading military if you are the world's leading technological power and the world's leading economy.

Because you need the technology to actually develop the military systems that make you world-leading, and you need the economic power to pay for those. Basically, the greatest success story in our lifetime from a world situation perspective was the defeat of Soviet communism in 1989. Basically, if you read what happened during that period, they basically surrendered. They basically gave up. And this happened in a way that, by the way, shocked almost all the experts, the Soviet Union just showed up one day and they said, we’re done. We’re not going to fight you anymore. We’re done. Basically, what happened was they fell behind technologically. They fell behind economically. And then they concluded that as a result, they were sunk militarily, and it was going to be impossible to win the war, and there was no point in trying to fight anymore. It was one of the most extraordinary events in world history. This global empire with huge ambitions and huge power and a record of massive brutality just gave up.

Ben Horowitz: Hundreds of millions of lives have been saved and improved coming out of that situation.

Ben Horowitz: I know. It's interesting because there's always been this communist movement in America, especially among young people. And the response to that is always, just meet anybody who lives in a communist country and you're cured.

Marc Andreessen: Absolutely. Basically, any criticism of capitalist society or economy is magnified tenfold under communism. Whether it's environmental damage or quality of life, any problem is worse under communism. It's a huge moral victory, but it's due to that triangle, which is technological strength, economic strength, and military strength. That's why technology is important. If you don't have the technological part of that triangle, you don't have the economic part. You don't have the military part.

Ben Horowitz: And, that strength was built through small technology entrepreneurs, like Thomas Edison, Henry Ford, who built these technologies and companies and these manufacturing capabilities that ultimately helped us win World War II and build the economy. Those same entrepreneurs are still around today, or their descendants. Those are the interests that we care about and try to protect.

Marc Andreessen: Exactly. And then that brings us to the role of startups and startup tech. And I would say to that question, and this is probably the point that I disagree with most with my very close friend Peter Thiel, who famously wrote in his book that innovation requires monopolies. And this is probably the point that I disagree with most with him, that monopolies prevent innovation. Monopolies are death to innovation. Because as you like to say, the motto of any monopoly is, "We don't care because we don't need to." And you just look at the behavior of any monopoly company, they just don't innovate. And the answer is because they don't need to. It's hard, it's risky, it's dangerous, you might get fired, people might laugh at you, they just won't do it. So basically, innovation comes from startups, and innovation comes from startups in two ways. It comes directly from startups. As you mentioned the long list, by the way, like Lockheed, Lockheed was a startup. Hughes, all the core technology defense contractors that won World War II, the Cold War, were startups when they started.

When they succeed, they are startups. You have the contribution of startups, but you also have this other important part, which is that the presence of startups keeps the big companies on their toes and incentivizes them to take action. Of course, the classic example that's playing out right now is Google inventing the Transformer in 2017 and then shelving it. Then startups came in and picked up the idea and ran with it. OpenAI, Anthropic, and many others. Now, as a result, Google is now forced to react, and now they're doing amazing technology development, with Gemini, doing brand new breakthroughs, but this is clearly a result of responding to the pressure of startups. So even if you think that all innovation is going to come from big companies, they still have to be incentivized to do it, and they're incentivized to do it through competition from startups. So basically, what this means is that the survival of America's technological advantage depends on whether startups succeed, whether there's fertile soil. Finally, I want to point out, which startups, because one of the responses I heard was, Marc, obviously that's true for something like autonomous software for drones, because it has military applications, but how can you say that about other XYZ things?

What I’m saying here is, look, historically there’s really no way to predict which innovations are going to be really important. My personal example, which we talked about in the last podcast, is that in 1993, no one in the expert class would have said or did say that the Internet would be of strategic importance. If you go back even further, one of my favorite books that I recommend to people is a book written 50 years ago by an MIT professor called “Man, Machine, and the Modern Age.” In it, he goes through the history of new technology and how it’s adopted and pervasive. One of the examples in the book is that 100 years ago, there was a guy named Sims who invented the first naval gun. This was about 120 years ago. The first naval gun that could automatically counteract the rolling of the ship in the waves. This gun would automatically stay aimed even when the ship was rolling. Basically, he was ridiculed, scorned, mocked, and stonewalled at every step of trying to promote this idea.

It’s a spectacular story. But there’s always this story with any of these new things, there’s always this story of thinking this thing can’t possibly be important, can’t possibly be meaningful. And then it turns out that some things may not end up being important, but a lot of things end up being very important. Now there are people who think they can judge in advance which new waves of technology, which new waves of startups are going to be strategically important. That’s a deeply misguided view. They don’t have any ability to predict that. And if they apply that judgment, they’re likely to be wrong. People like them have been wrong on these topics in the past. And now this tendency to apply this kind of, how do you put it, perverted industrial policy, where you say, this sector is going to be allowed to succeed, but we’re going to kill this sector in the cradle, is extremely destructive.

Blockchain, Cryptocurrency, and Stakeholder Capitalism

Ben Horowitz: The strength of America is all of our smart talent coming into play. You're right. Anybody with talent and ambition can try to build something. As opposed to, in Stalin's Russia, if Stalin didn't let you build it, you couldn't build it. That's the problem, when you start applying that, you get all of America's weaknesses, our crazy disconnected culture, all of the depravity, the drug addiction, but without the thing that makes it all work, which is the capacity of all of our talent. We're so talented as a country, talent from all over the place coming into play. That's not the case in top-down governments, communist societies, dictatorships, and fascist societies. In those worlds, talent is destroyed. Interestingly, we spent some time with Malay Yu specifically, but I just met him at a conference in Argentina. The story of Argentina, which is one of the most talented countries in the world. But all of this talent has been completely destroyed by bad government. That's something we worry about at the startup tech agenda.

So should we get into the subject and compare and contrast the two campaigns? We certainly know more about Biden because he's been developing policy for the last four years. But we do know what Trump's platform is and what he did before. Some technologies don't stand out, but some are important. Okay, the first topic is blockchain and cryptocurrency, which is one of the areas where we interact with the government the most. Let me start by explaining why we think this is a very important technology. Going back to the internet, one of the reasons the internet is so special and so important is what's called permissionless innovation, or the network that nobody owns. Microsoft doesn't own it. AT&T doesn't own it. Nobody owns it. It's the community that really runs the network, runs the internet. Anyone can participate, build a node, and go ahead. The result is that anyone can build a business on it, anyone can create on it without having to pay a lot of taxes to some corporation. I remember all the companies that were built. But I also remember Prince putting his work on the internet. It was an amazing thing because he was able to, he changed his name to get out of his record deal and did all these things. It was an exciting place where everyone could participate without having to go through the big industry. You could do your thing. If you had a talent or an idea, you could publish your idea. You can write and put it on the Internet. You don't have to get a job at the New York Times, etc.

But then what happens is that the history of technology is a little bit like the history of networks and servers. Eventually, servers were owned by big companies. There are a lot of technical architectural reasons for that. So the companies that won ended up owning all the data in the discovery layer, like Google, or the social network layer, or the commerce layer. And so those companies became extremely powerful, more powerful than most governments. Google, I think we would all agree, is more powerful than 95% of the countries in the world in terms of control of information, ownership of money, and so on. And that has adverse effects on people in a number of ways. One is to build a business, you are now dependent on them, and they can wipe out your business by making you unfindable. If you’re an artist, the percentage of your work on Instagram or TikTok is very high. You upload content, they take 99% of the proceeds. There’s nothing else in the world that does that. Even a company like Apple, you build an app, they take 30% of the proceeds. That’s just a very high concentration of economic power, control of information, and so on.

We haven't had the technology to handle this. But now we have blockchain. It's an amazing breakthrough, you can build these same services. But instead of companies owning the computers, the community owns the computers. Just like the community owns the network. It's a really remarkable breakthrough, and it's really important for society, for the functioning of a healthy society. We hear words like "stakeholder capitalism." It's really stakeholder capitalism, where the customers participate, the people in the community who run the servers participate. Everybody participates. If you look at the early services, the creators get 90% or 95% of the revenue, not 1%. It's really an amazing potential.

Also, it solves some real societal problems that we face, like who controls the truth. This comes up with deepfakes, right? You have AI, and now AI can make a video that looks like Joe Biden. How do we know if this is real or fake? Who tracks this? Who has the source of truth database for this? Is it the Republicans? Is it the Democrats? Is it Google? Who do you trust, or should it be a community-run utility? That's the problem that blockchain solves. We think this is probably the most important technology that we deal with for our country, for our society. It needs, we work with the government right away, because it does need regulation because of the way that communities pay for the electricity that's used to run computers. There are about 10,000 people running the Ethereum service, which is one of these blockchains, and they get these tokens. These tokens are their compensation. And the thing about a token is, it could be a stock certificate, it could be a Pokémon card. It could be a lot of things. It's a computer structure. There needs to be a regulation that says, in this case, it's a stock certificate, in that case, it's a Pokémon card. This is how the industry can function and it eliminates the problem of scammers or people who try to steal your money.

Then we went to Congress and said, maybe we can pass a law to clarify when something is a security and when it is a commodity. The government, in the form of the SEC, specifically, the White House appointees, have been working against us every step of the way, using very vicious tactics. Just to name a few. First of all, they refused to issue any guidance. And then they went after companies without a law, without guidance. So they sued over 30 of our companies. They sent out Wells notices, and in our entire career, we have never seen a Wells notice, which is a form of the SEC announcing that they are investigating and going after any private company. These are private companies. Specifically targeting this industry. They lost almost all of these lawsuits, but the thing is, when you are a startup, you don't have the money to fight the US government. They are destroying the industry that way. They also made it impossible to have a bank account, even though it was completely legal to build this technology as a company. The FDIC said to the banks, look, if you bank these companies, we will not let you borrow at the federal window. We will attack you, and this is a similar technology that you can read about that they used with the marijuana industry, why you have to pay with cash at a marijuana store, and it's the same thing they're doing in the crypto industry. You can read that if you're interested. They call it Choke Point 1.0 for cannabis and Choke Point 2.0 for crypto. Going beyond the law using the administrative state to basically try to destroy an industry.

And then we tried to pass laws. There was this thing called SAB 121, which was a crazy rule that the SEC wrote that said if you have custody of someone else's crypto assets. If I have some Bitcoin and I give it to you for safekeeping, then as a bank, if the price of Bitcoin goes down, you're liable for that, even though it's my Bitcoin. You can imagine, oh, banks store gold, if the price of gold goes down, their customers own it, not them, they have to be liable. Well, that meant that banks couldn't have custody of crypto. So the Senate, Democratic senators like Chuck Schumer said, come on, this is ridiculous. They voted to get rid of the rule. And then Joe Biden overruled the Senate's bipartisan conclusion to keep the rule. That's how aggressive they were. It was really very frustrating and difficult for us and the industry as a whole, and this is such an important technology.

This has been one of our biggest battles. To give you an example of how devastating this has been. We have a company, Worldcoin, that's solving a big AI problem, which is that you can prove you're human without revealing your identity, without revealing your social security number, without revealing your driver's license, like hard password strength, proving you're human and not a robot. They've launched their product all over the world, except the United States, because of this behavior and these laws. To give you an idea of ​​how important this product is, I mean, this is something that the Malaysian government endorses. This is something that half the citizens of Buenos Aires use regularly. This is a really great, useful, important product that can make everyone's life better, and we just banned it in the United States. This is a huge blow to startup tech.

Challenges facing the crypto industry

Ben Horowitz: We should go ahead and explain this. We were trying to meet with Gary Gensler, who was the chairman of the SEC. He was running this campaign against cryptocurrencies. We were the largest cryptocurrency investor or the largest blockchain investor in the world. We requested at least six meetings with him. I even reached out to his office at MIT, and he said, “Gary has to meet with you. This is so important, he has to meet with you. You know all these things.” But he couldn’t schedule a meeting.

It was said that his appointment was not made by Joe Biden, but by Elizabeth Warren, who also refused to meet with us. She was the only senator who refused to meet with us. And of course, we couldn't meet with the president himself. It was really, really frustrating and difficult. By contrast, when we were pushing this legislation, the Market Structure Act, which determines what is a commodity and what is a security. We've been working with Congress, and by the way, we've had some wonderful Democrats in Congress who have been helping to support us. Richie Torres has been an absolute hero. Chuck Schumer has been great. We've been working with them. While we were doing all this, Donald Trump came out with his cryptocurrency stance, which you can read about, but the contrast is just crazy now.

Marc Andreessen: This was out before we met with him, and it's such a stark contrast. But he's talked about it publicly, and he's actually going to speak at the Bitcoin conference after the convention. He'll have more on the record on this. But one of the things that Trump did was he rewrote the Republican National Committee platform for this convention, and he rewrote it himself. It's a very clear, simplified version of the statement of intent. You can download it and read it, it's worth reading. I'm going to read a paragraph straight out of it. It's chapter three, building the greatest economy in history. Point five. We all support this. We all support this. Point five, supporting innovation. Those words are only coming out of his side now. And then, paving the way for future economic greatness, leading the world in emerging industries. The first industry is cryptocurrency. We will end illegal and un-American cryptocurrency suppression. We will oppose the creation of central bank digital currencies, which is another section. We will defend the right to mine Bitcoin. We will ensure that every American has the right to self-custody digital assets and to trade them without government surveillance and control. It's like a full-throated endorsement of the whole space. Like a completely unanimous embrace of the whole thing. He's talked about this on other occasions and confirmed this. This is a complete 180 degree turn from what we have experienced.

Ben Horowitz: I would also say, look, our experience with this is just personally shocking to me. This idea. And, again, I come from an era where, you know, the Clinton administration would never have done this. I don't think the Obama administration would have done this. Absolutely not. This is new behavior.

Marc Andreessen: Well, whether you like Obama or not, he was always interested in the opinions of business and technology experts, whether he agreed with them or not. He wanted to learn. He was not someone who wouldn't listen, and that's the problem we're facing right now.

Marc Andreessen: It's tough. It's terrible. It's a brutal blow to the industry, and I've never experienced anything like it. I'm completely shocked that this could happen. It's really frustrating that we can't make progress on this with the White House.

Ben Horowitz: This is a completely intolerable situation. And it's really interesting because our old friend Reid Hoffman wrote an article defending the Biden administration. He said that business people should support Joe Biden because the most important thing for business is the rule of law. And I thought, wow, that's really ironic because they're basically subverting the rule of law to attack the cryptocurrency industry. I mean, we're dealing with this right now. It's probably the most emotional topic. Let's move on to artificial intelligence. Why is artificial intelligence important to the United States?

Marc Andreessen: Look, this is actually an issue that nobody is arguing about. The cryptocurrency people will argue about whether it's important or not. And obviously, we are very adamant that it is important. With regard to AI, there is no question that it is going to be a cornerstone technology for economic growth. If it is allowed to develop the way it should, what is going to happen next is at least equivalent to the dot-com boom of the '90s, if not the computer boom of the computer industry since the '50s, which is equally as huge for the United States. And then look, AI is powerful enough that it may be bigger than both of those. It may be the biggest technology boom ever. So it is extremely important in terms of the economy. And then look, the whole military doctrine, military warfare, everything is going to change. All of that is going to change around AI. And by the way, this is already happening. The Pentagon has classified AI and autonomy as what they call the third offset, which means basically the next generation of cornerstone technologies. There have only been three offsets in the history of the post-World War II era.

The first was nuclear weapons. The second was mobile warfare and related things. The third was artificial intelligence and autonomy. So there's going to be a complete revolution in military affairs, a complete reshaping of the way the U.S. military and the way the global military operates. The Department of Defense, yes, they've confirmed this. We're working with them a lot on this. They're pushing very hard on this. We've funded a lot of companies. So this is going to be very important. And then China has announced the exact same thing. So they're completely reshaping their military in the same way. This is a huge national security priority for them. They have complete control over their domestic tech industry. So they're using the capabilities of their smart people to do this in a very targeted way. This is the technology race of our time. This is going to determine the world for hundreds of years to come. This is going to determine how government operates. This is going to determine how surveillance operates. This is going to determine who wins wars.

Marc Andreessen: You can see it. Exactly. I don't know why you're so worked up about this. Well, look, you can already see it. You can see it in Ukraine. Right, I won't go into detail here, but it's a huge shift. It's all happening right now in Ukraine. And then obviously, this has major implications for what's happening in the Middle East and elsewhere as well. This is a core and foundational technology, as you would expect. The United States absolutely has to win this race.

The Biden Administration’s Regulatory Storm

Tax policy is a key factor when it comes to the growth of the tech industry. Marc and Ben discuss some of the Biden administration’s tax proposals and their potential impact on startups.

Ben Horowitz: So what have we seen so far from the Biden administration, and specifically the White House? I would start by saying that Congress has been working in a bipartisan way to try to find answers. The Biden administration is doing it a little bit differently. Maybe you could just give a brief rundown of their actions so far.

Marc Andreessen: There are different branches. The Secretary of Commerce that you mentioned is one example, someone that we've been talking to.

Ben Horowitz: Gina Raimondo, by the way, has done an excellent job thinking about AI policy, particularly how it relates to issues like China and CFIUS, as has National Security Advisor Jake Sullivan. When we talk about the White House, we're really talking about executive orders from different parts of the administration.

Marc Andreessen: And there are also a lot of very smart people working in the Department of Defense and the intelligence agencies who, I want to emphasize, are doing a lot of good work. There are a lot of positives, a lot of people who are really working hard to solve this problem.

Ben Horowitz: Okay. We were trying to figure out this idea in the executive order, which was basically about how much math you were allowed to do in your life, limited to 10 to the 26th power of floating point operations (flops). You and I did a lot of research. We talked to advisers in the White House. We talked to policymakers in the White House. And it came down to a belief that we thought was very wrong, basically, if I can summarize it, that there were probably only three or four base models in the world that were relevant because only a few companies could afford the number of GPUs needed to make relevant models. There were a couple of problems with that idea, first of all, there are a lot of techniques that show that you can make a really good model on top of someone else's model. There's a technique called distillation, which we talked about before. Second, it seems that labeled post-training data is more effective than adding GPUs. Scale may be the only thing that matters, or it may not be. But there are a lot of very important base models that are not big right now, but are getting a lot of use, that are not in those three typical models. The idea is that if there are only three models, and they're all from the largest companies in the world, then you can put some hurdles in front of them and they'll jump over them. But for startup tech, by the way, almost all of the underlying models that we invest in are growing very fast and performing very well. Those are going to be killed by this type of regulation. Maybe you can talk about why the floating point limit is so weird.

Marc Andreessen: For those who don't know, a flop is a technical term in computer science that stands for floating point operations per second. I know that doesn't explain much. What it means is, basically it's a way of measuring the speed of a computer, how many operations a computer can do per second, how many math problems a computer can solve per second. Computers are so fast at doing math right now, they can solve millions of these operations per second. The idea of ​​this 10^26 floating point operations in the executive order is that, as you said, only the largest models can reach that level. The problem is that there's a 50-year-old general phenomenon in our industry that's probably the clearest thing we know about in terms of predicting the future of our industry, and that's called Moore's Law. It's a well-known phenomenon that computing power gets faster and cheaper every year.

I'll also say one thing about Moore's Law, companies like Nvidia, which is now one of the most valuable companies in the world, about three trillion dollars. It's an incredible success story. The reason the company exists is that their chips are now the standard chip. That's now the biggest profit incentive in the history of the global chip industry, and you have to find a way to apply more chips to compete with Nvidia, lower prices, and increase production. Even Nvidia's CEO, Jensen Huang, who is a good friend of ours, is pushing this hard. He predicts that the speed of these chips will increase dramatically, and the cost will decrease dramatically from now on. So, over the next decade, the cost on the chip side will drop dramatically. So, things that seem very expensive and difficult today will become very easy and very cheap. There are a lot of historical precedents. I'll give you one of my personal favorite examples, which is the original Sony PlayStation, which is a game console.

Ben Horowitz: I have one, playing Call of Duty.

Marc Andreessen: People knew about that. When those first shipped, they were export controlled. So Sony wasn't allowed to ship those to countries that might have had their own nuclear programs, like Iraq, because those chips were considered too powerful and could be used to build nuclear weapons. That illustrates a fundamental question of why that was, because previously someone had set floating point limits for supercomputers. And then Moore's Law took effect, and the Sony PlayStation came out as a $300 consumer device with that floating point power. All of a sudden, you're export controlled. The beauty of GPUs. The PlayStation was one of the first GPUs, it had a GPU in it. These restrictions that seemed reasonable, quickly became unreasonable. I'll give you another example of how this deflation happened. When OpenAI first released GPT-2, which was probably the first working LLM, they refused to release it at first because they had all these warnings about how dangerous it was and how safe it was. Today, GPT-2 is now open source.

Ben Horowitz: Andrej Karpathy was the co-founder of OpenAI, he's left, and now he has an open source version of GPT-2, and you can train your own GPT-2 from scratch for less than $100 in compute cost.

Marc Andreessen: It's like, I don't know, they went from like $10 million to $100. Right, that kind of order of magnitude drop. Things that seem very expensive and very exotic today are going to become mainstream and then become cheap and then everybody will have them. And by the way, every computer science student will have a model of this capability on their laptop. You'll have a model of this on your phone. In five years, you'll have a model of this scale on your phone.

Ben Horowitz: This deflationary effect is going to be very significant. These things are going to come out, and these are going to be like microprocessors.

Marc Andreessen: These are going to be universal products that are basically going to be built into all devices, and everybody is going to have it in every device. And basically the regime that the government is threatening to put in place now, by using this restriction, is going to basically establish the two or three companies that they think are the only ones that matter. Sort of like a perpetual monopoly. They're going to basically destroy the startup ecosystem underneath. They're going to establish an OPEC of AI from the beginning. This is one of those things that I have a really big question about: Is this ignorance or is this malice? Right, do they not understand what they're doing, or are they doing this on purpose to set up a cartel? In either case, I find this completely unacceptable.

Ben Horowitz: This kind of ties into something we talked about before, which is the idea that you can predict the future, kind of like the precautionary principle. The trouble with predictions, as Yogi Berra said, is that they're hard, especially predictions about the future. And if they could predict the future, obviously the future is unpredictable, because if they could, they'd be richer than Elon Musk or anybody. Of course, nobody can predict the future, especially about these things. We know we can't predict the future, and our whole job is to try to predict the future. But it's very dangerous to make policy before the future arrives. We've seen this in Europe, where they've basically eliminated almost all innovation through this idea of ​​the precautionary principle, trying to stop the future before it happens. That's probably the one that I'm most concerned about.

Marc Andreessen: I need to emphasize that AI is really just math. I mean, like, just math. They're just big math models. Huge equations. Specifically, it's an area of ​​math linear algebra, and there's a key algorithm called gradient descent. And, these things, by the way, these things are taught in every university. These things are all over the world. These things are in YouTube videos. These things are all over the internet. They're talking about standard textbooks. This is not rocket science. This is hard to hide. Nuclear physics. This is being done in the open. This idea that this is going to be shackled, by the way, this again goes into the relationship between the nation and global policy. The idea that we're deliberately shackled ourselves through cumbersome regulations while the rest of the world is lighting up this, and China is lighting up this, there are very few secrets in this area. So basically any restrictions that we impose on ourselves are only going to put the United States at a disadvantage relative to the rest of the world.

Ben Horowitz: It's kind of like something we talk about a lot, you go into these laws. There are laws and there's enforcement. If you make a law that's impossible to enforce, you end up in what's called anarchic despotism, which is anarchy for people who don't follow the law. They can do whatever they want. And then despotism for law-abiding citizens. Because no matter how carefully you follow the statutes of a Byzantine government, you're going to screw up something and get punished. That's a very bad policy idea. That's another thing we worry about when we're dealing with these things and how we come to these policy conclusions.

Marc Andreessen: If I hadn't seen what blockchain was doing, I wouldn't necessarily feel this strongly. But I see where this path is going. You can see this in adjacent industries right now in real time. And it's already catastrophic. We can't let this happen again in our country.

Ben Horowitz: We're still very early in the industry. We're seeing really important innovation across the industry. People from all walks of life and all countries are doing amazing things in AI. It would be very irresponsible to say right now that I know where the industry is going to go in 20 years. My biggest concern is the foreshadowing of what we're seeing in cryptocurrency. This is going to happen in AI. This will unleash the administrative state to do their thing. While I have no confidence that they will stand with us, some of them are not the entire administrative state, but certainly some of them will affect us.

Marc Andreessen: But if the government cracks down on AI like it did on blockchain, then the country will be in deep trouble.

Ben Horowitz: That's the default path. That's totally unacceptable. That's very concerning. That's a very concerning meeting. It's not set in stone. As I said, there are a lot of people in the White House, and we'll have to continue to fight, and we'll see. But it's very concerning. And then let's talk about what Trump is proposing. We actually talked about this when we had dinner together. We talked about all of these topics and confirmed this.

Marc Andreessen: Including building the greatest economy in history, supporting innovation, and supporting artificial intelligence. We will repeal dangerous executive orders that hinder AI innovation and impose radical ideas on technology development. We will support the development of AI based on free speech and human flourishing.

Ben Horowitz: I think it sounds like a good plan. When we met with him, I thought his comments were very insightful and funny. I would contrast the Biden administration's approach, especially the core of the White House, with Trump's approach. The White House has a very sophisticated model. I would say they think they know a lot. They know that startups are not going to be important, that only a few companies will be able to deliver large models. They know all the things that we don't know, they've never heard of distillation, or how AI actually works. But it's a very sophisticated view of the world. Whereas Trump has a very simple view. He said to us, AI is terrible, but we absolutely have to win, because if we don't win, then China will win. It's a very bad world. That's actually a more correct view, that's the basic truth. Look, when things start happening they do need to be regulated, and we should regulate them. But anticipating it, the car is coming, and we think someone is going to build a car that goes 500 miles an hour and no one can control it. We're going to ban cars now. I mean, it's a little bit like the attitude towards AI, we think there's going to be a sentient model in the future. Right now, it doesn't seem like anyone has built anything that's heading towards sentience. Doing that, we have these great things that can tutor kids. No, you can't tutor a child because someone might come up with an idea and make a sentient AI. We have to cut off the tutoring. That's a very scary way of thinking, I would say that.

Marc Andreessen: One of our arguments is because we are not against regulation, as we described in the blockchain section, obviously there is regulation.

Ben Horowitz: Or a model that can self-construct into a super genius, and nobody knows how to do that.

Marc Andreessen: We are actually trying to pass a regulation law. It's Act 21. We are actually trying to create a regulatory structure because the government won't do it, we are actually trying to create new regulations. One of the arguments we made in meetings with the White House to discuss AI policy was that problems would come from AI, but the assumption was that they should be regulated. Regulation should happen at the application level, not at the technology level. The reason is basically because regulating AI at the algorithmic level is regulating mathematics. It doesn't make any sense. Of course, we're not going to do that, do you remember what they said? No, actually, we can classify mathematics. Literally, they did classify entire areas of physics in the nuclear age and made them state secrets, like theoretical physics, science, physics. We completely classified it as a state secret, and that research disappeared. We have every ability to do that again for AI. We will classify any area of ​​mathematics that we think is going in a bad direction and we will terminate it.

Ben Horowitz: That's interesting too, because now with this thing called the internet, it's very hard to keep math secret. There's no such thing as information secrets anymore.

Marc Andreessen: Some people might say, well, that's crazy, and of course they wouldn't do it. But the problem is, I've seen so many crazy things so far. I no longer believe, I feel like I can no longer predict which of these crazy ideas will be implemented. I personally cannot tolerate the idea that literally we might make linear algebra a state secret. That's out of bounds.

Ben Horowitz: You're getting excited. Okay, I just want to briefly mention fintech because fintech is almost a subcategory of crypto in some ways, from a regulatory approach. But I want to tell a story because it really impacted me personally. We invested in a company called EarnIn, and the founder is Ram, and I didn't invest in it, but I talked to Ram almost every month. We had dinner together every two weeks because, first of all, what he was doing was a mission for him, and it was very closely tied to my life. The fact of the American financial system is that it's extremely bad for poor people. If you're poor, you pay more for banking services than rich people. You and I don't pay anything for overdraft protection. And people in the lower income brackets spend more on overdraft fees than vegetables. It's very bad.

There's a huge portion of the population in the United States that doesn't have a bank account. If you don't have a bank account, you're usually in the underground economy, and you don't have credit. My wife's family is in that category. The only credit that a lot of people have access to in that situation is payday loans, which have extremely high interest rates. There's no business that can grow with payday loans or credit card loans. That basically means you lose your freedom. You don't have credit. If you look at countries that don't have credit, they don't have job growth. Like Argentina, which is a big problem that Malay is working on, because there's no consumer credit, they haven't had job growth in a decade. It's a very bad thing for certain segments of society. A lot of black and Hispanic people are in that category, and they can't get out of this, they don't have a bank account. Ram came up with this most amazing solution, where if you're in an hourly job, they can track if you go to work and you need a loan because you haven't been paid for two weeks. Let's say you don't get paid until Friday and your daughter's birthday is Saturday and you want to have a birthday party. He'll lend you the money interest-free. You can choose to tip him or not.

So you start building credit, and then he’ll bank you. And his interest is much higher than normal. This is finally a gateway out of the underground economy and giving opportunity to people who have been systemically oppressed by the financial system, and it’s a great mission, a great guy. His story, he’s doing this. He’s providing these loans, everyone from police officers to McDonald’s employees and so on. No company is more loved by customers than EarnIn. And then COVID comes, and this is what startups are like. You have this great mission, you’re growing like crazy, everything’s great, you hired all these people. COVID comes, the government sends these stimulus checks, and people don’t need to work. Everything disappears, and he has to lay off people, and it’s very difficult. His dreams are shattering and so on. But eventually, it’s over, and he’s back on his feet. Like he’s now ready to grow the business, but he’s being attacked by the administrative state, specifically the CFPB. They send him threatening letters demanding that he waive attorney client privilege. How can that be legal? And then they demand that he submit to the jurisdiction of the CFPB or they’re going to sue him. I’ll never forget this, we asked, sue us for what? We won’t tell you. Like that kind of harassment. It's just crazy.

We talked about, and I feel like it’s important to point out, this isn’t the Democratic Party. This isn’t all Democrats. This is really a far-left, totalitarian branch of the Democratic Party that’s taking some of these things under Biden. Part of it is because Biden delegated it. I’ll put it mildly, he delegated it, but it has real implications. Anyway, then, we have this situation in biotech as well. Biotech is in an incredible position with AI, where all of a sudden we have the potential to cure all kinds of diseases that were previously untreatable. But like these limits on AI, these arbitrary limits, 26. How many floating point operations is that? 10 to the 26th power. If you exceed that, then what if that’s the size of the model you need to cure cancer? It’s a similar struggle.

Another problem we have in this administration is that it's almost impossible to cure rare diseases from a venture capital perspective because it takes a billion dollars to get a drug to market. And they don't always work. Multiply that amount. And there aren't that many patients for rare diseases. You can never get your money back. But we had a company with an incredible drug platform. They cured a rare disease, got it through Phase 1 clinical trials, cured it. Sold the drug to Sanofi so that they could get some money back to develop more things. The Biden administration sued them and scuttled that deal. It basically makes it almost impossible for us to fund any company that's trying to cure rare diseases. It's an arbitrary clampdown on startups. Maybe you could talk a little bit about the dynamism of the United States. We do a lot of work in the defense space, as you mentioned, my job. Maybe talk a little bit about why the way we're building our military right now is unlikely to be competitive if it's not upgraded. And then what some of the barriers to that are.

QSBS is the lifeline of innovation

Marc Andreessen: Exactly. The most important part of the tax code for startups and innovation is called the QSBS, the Qualified Small Business Stock Exemption. It's been around for quite some time. It's a bipartisan measure, one of the few that has been passed and signed into law by both Democratic and Republican administrations over the years. It basically says that if you start a company from scratch and hold shares in that company for at least five years, then when you sell those shares, the gain on the sale will be tax-free or taxed at a very low rate. That's the fundamental incentive for people to start companies, invest in companies, hold those investments, and then eventually sell those investments and roll the proceeds into new companies. It's the reason we have the Silicon Valley that we have today. It's the reason we have the entrepreneurial culture that we have. It'