Podcast Episode Summary
The podcast episode features Mark Andreessen, co-founder of Andreessen Horowitz (A16Z), discussing various topics related to venture capital, entrepreneurship, and technology.
Main Topic: The discussion revolves around the evolution of venture capital, the role of startups in driving innovation, and the potential for new technologies to disrupt traditional industries.
Key Points:
Takeaways:
Participants:
Parting Thoughts:
The podcast episode provides valuable insights into the world of venture capital and entrepreneurship. Mark Andreessen's perspectives on the evolution of venture capital, the role of startups, and the potential for innovation to disrupt traditional industries are thought-provoking and informative. The discussion also highlights the complexities and nuances of the venture capital industry, including the potential for overfunding and overstaffing.
My podcast with the brilliant Marc Andreessen is out!
We discuss:
* how AI will revolutionize software
* whether NFTs are useless, & whether he should be funding flying cars instead
* a16z's biggest vulnerabilities
* the future of fusion, education, Twitter, venture, managerialism, & big tech
Watch on YouTube. Listen on Apple Podcasts, Spotify, or any other podcast platform. Read the full transcript here. Follow me on Twitter for updates on future episodes.
Timestamps
(0:00:17) - Chewing glass
(0:04:21) - AI
(0:06:42) - Regrets
(0:08:51) - Managerial capitalism
(0:18:43) - 100 year fund
(0:22:15) - Basic research
(0:27:07) - $100b fund?
(0:30:32) - Crypto debate
(0:43:29) - Future of VC
(0:50:20) - Founders
(0:56:42) - a16z vulnerabilities
(1:01:28) - Monetizing Twitter
(1:07:09) - Future of big tech
(1:14:07) - Is VC Overstaffed?
Dwarkesh Patel
Okay. Today, I have the great pleasure of speaking with Mark Andreessen, which means for the first time on the podcast, the guest and the host, playback speed will actually match. So, Mark, welcome to the Lunar Society.
Marc Andreessen
Good morning, and thank you for having me. It's great to be here.
Dwarkesh Patel
My pleasure. So, have you been tempted any time in the last 14 years to start a company? Not A16Z, but another company?
Marc Andreessen
No. I mean, we did. So we started our venture firm in 2009. It's given my partner, Ben, and I a chance to fully exercise our entrepreneurial ambitions and energies to build this firm. We're over 500 people now at the firm, which is small for a tech company, but it's big for a venture capital firm. So it's let us fully get all those urges out.
Dwarkesh Patel
But there's no product where you think, "Oh, God, this needs to exist and I should be the one to make it happen"?
Marc Andreessen
I think a lot. We look at this through the lens of, "What would I do if I were 23 again?" So I always have those ideas. But starting a company is a real commitment. It really changes your life. My favorite all-time quote on being a startup founder is from Sean Parker, who says, "Starting a company is like chewing glass. Eventually, you start to like the taste of your own blood." That quote always gives people a queasy look on their face when I roll it out. But it really is intense.
I always tell people, whenever anybody asks me if they should start a company, the answer is always no. It's such a gigantic, emotional, irrational thing to do. The implications of that decision are so profound in terms of how you live your life. There are plenty of great ideas and plenty of interesting things to do, but the actual process is so difficult. It gets romanticized a lot, and it's not romantic. It's a very difficult thing to do. I did it multiple times before, so at least for now, I don't revisit that.
Dwarkesh Patel
But being a venture capitalist is not like that. When you're in the 30th pitch of the day, you're not wondering if chewing glass might not be more comfortable?
Marc Andreessen
No, it's different. People are wired to respond to stress in different ways. There are people who are wired to be extremely productive and actually get very happy under extreme levels of stress. I have a different inclination. I'm fine with stress. I'm inclined towards it. If I don't have any, I seek it out. But I don't pass a certain level. I don't really enjoy it. It degrades the quality of my life, not improves it.
There's stress in everything, and there's certainly stress in being an investor, but it's a completely different kind of stress. When you're a startup founder, it's all on you. Everything that happens is on you. Everything that goes wrong is on you. When there's a crisis in the company, it's on you to fix it. You're up at four in the morning worrying about things. As investors, there's a layer of buffer. We have no end of problems and we help our portfolio companies as best we can, but some crisis inside a company—it's not my company. Not everything is my fault. So it's a more diffuse kind of stress, and honestly, it's easier to deal with.
Dwarkesh Patel
Got it. That makes sense. Why did you stop your blog? Would you ever start it again?
Marc Andreessen
I write intermittently. I stopped the original blog around 2007 to 2009. Then we started the firm, and that was like having a new baby that soaked up all my time. I write intermittently, and I do social media intermittently. Part of it is I have a lot to say and a lot that I'm interested in, but also I like to experiment with new formats. We live in a fundamentally different world as a result of social media and the internet and blogging and Twitter. So I try to keep my hand in it and experiment. But I rotate both how I spend my time and rotate what I think makes sense.
Dwarkesh Patel
Before AWS, deploying applications was probably the bottleneck on new software. What is the biggest bottleneck today? At what layer of abstraction do we need new tools?
Marc Andreessen
Sitting here today, I think overwhelmingly it's the impact AI is having on coding. I think there's a real possibility that basically every application category gets upended in the next five years. The whole model of how applications get built across every domain might just completely change. The old model without AI—you typically have some sort of database, some sort of front end, forms, known user interaction models, mobile apps. We got to a pretty good shared understanding of how humans and machines communicate in the windowing era, the mobile era, the web era.
I think AI might just upend all that. The future apps might be much more of a dialogue between computer and machine. That's either a text dialogue, a spoken dialogue, or some other form of dialogue. The human is guiding the machine on what to do and receiving real-time feedback. There's a loop, and the machine just does what it does and gives you the results. We're potentially on the front end of that. The very fundamental assumptions about how software gets built might just completely change. So the tools for that are at the very front end. There's an entirely new stack that needs to get built to do that. That's probably the big thing.
Dwarkesh Patel
Is there a reason, though, AI is not one of your focus areas? Or as far as I know, you guys don't have an AI fund dedicated to the technology specifically?
Marc Andreessen
We look at it as it's all of software. It is the core business. Software is the core of the firm. We've been public on that for a long time. The core venture fund is the core software fund. AI basically is the next turn on software. I view it kind of as the opposite of what you said. It is the most integral thing that we're doing. Separate funds get created for new areas that are structurally different in terms of how industries work. But AI is basically the future of software, so it's the future of the core of the firm.
Dwarkesh Patel
Got it. Now let's talk a little bit about your past. You sold Netscape for $10 billion, but today Chrome has something like 2.7 billion users. Opsware was sold for $1.7 billion. AWS is going to make close to $100 billion in revenue yearly. In retrospect, do you think if these companies had remained startups, they would have ended up dominating these large markets?
Marc Andreessen
I spend virtually no time on the past. The one thing I know about the past is I can't change it. So I spend virtually no time revisiting old decisions. People I know who spend a lot of time revisiting old decisions are less effective because they mire themselves in what-ifs and counterfactuals. I really don't have theories on it.
The big thing I would say is reality plays out in really complicated ways. Everything on paper is straightforward; reality is very complicated and messy. The technical way I think about it is every startup is charting a path-dependent course through a complex adaptive system. People used to have this obsession with chaos theory. We're used to thinking about systems as if they're deterministic—you start at point A, you end up at point B, and you can do that over and over again. In the real world, with 8 billion people interacting and trying to start companies and intersect in markets, there are random elements all over the place. There's path dependence as a consequence. You run the same scenario starting at point A. One time you end up at point B, one time you end up at point Z. There are a million reasons why the branches fork. You just can't revisit old decisions. The world is too complicated and messy. So you take whatever skills you think you have and you just do something new.
Dwarkesh Patel
Makes sense. Aren't venture capitalists part of the managerial elite? Burnham says that the rise of the finance capitalist is a decisive phase in the managerial revolution. What would he think about venture capitalists?
Marc Andreessen
I actually think about this a lot. James Burnham famously said there are basically two kinds of capitalism. We call them both capitalism, but they're very different in how they operate. The old model is bourgeois capitalism. That was the classic model where the owner of the business—often with their name on the door, like Ford Motor Company, Andersen Horowitz, Disney Company—owned the business, often 100%, and ran the business. These are the people the communists hated. The bourgeois capitalist, company owner, builder, CEO, all as one person. There was a direct link between ownership and control. The person who owns it controls it. The person who controls it runs it. There's a proprietor of the business.
What he said is that as of the middle of the 20th century, most of the economy was transitioning to a different mode called managerial capitalism. That transition is now complete. In managerial capitalism, you have a separation of ownership and management. You have one set of owners—dispersed shareholders, a million of them for a big company, who aren't paying attention and have no ability to run the company. Then you have a professional manager class that steps in and runs the company. As a consequence, the managers end up in control, even though they might own only 1% of the company. He said it doesn't matter if you think this is good or bad; it's inevitable because of scale and complexity. Modern industrial and post-industrial organizations are going to be so big, so complex, and so technical that you need this professional managerial class to run them. This is exactly what's played out.
A consequence of that is that managerial capitalism has a big advantage: managers are often very good at running things at scale. We have giant industries and sectors of the economy—healthcare, education—running at giant levels of scale, which was new in the 20th century. But the consequence is that managers don't build new things. They're not trained to do it. They don't have the background, the personality, the temperament, or the incentives to do it. The number one job if you're a manager is not to upset the apple cart. You want to stay in that job as long as possible, get paid your annual comp, and not introduce risk. So managers can't and won't build new things.
Specifically to your question, the role of startups—the role of entrepreneurial capitalism—is to bring back the old bourgeois capitalist model enough. It's a rump effort because it's not most of the economy today, but bring it back enough to at least be able to build the new things. What we do is fund the new bourgeois capitalists, who we call tech founders. There are basically two layers of finance that enable bourgeois capitalism to resurface a little bit within this managerial system. Venture capital does that at the point of inception, and private equity does that at a point when a company needs to transform. I view it as we're an enabling agent for at least enough of a resumption of bourgeois capitalism to get new things built, even if most of the companies we build ultimately end up being run in the managerial model. As Burnham would say, that's just the way of the modern world.
Dwarkesh Patel
But you guys get preferred shares and board seats, and rightfully so. Wouldn't Burnham look at this and say you're not the owners, but you do have some amount of control over your companies?
Marc Andreessen
He would say we're like a hybrid—a managerial entity that is in the business of catalyzing and supporting bourgeois capitalist companies. He would clearly identify the startups we fund as the old model. You can draw a straight line from Thomas Edison and Henry Ford to Steve Jobs, Larry Page, and Mark Zuckerberg. That's the model: a founder, a CEO, at least when they start out owning 100%. They do have to raise money most of the time, but they're throwbacks to this older model of bourgeois capitalism.
So I think he would view us as a managerial entity, but one that is in the business of causing new bourgeois capitalist institutions to be created. I think he would credit us with that. Then he would say what I also said: our fate is that most of the companies we fund and most of the founders we back end up over time handing off control to a managerial class. When our companies get to scale, they get pulled into the managerial orbit, into the managerial matrix. That's when they stop being able to build new things, which causes the smart and aggressive people at those companies to leave, come back to us, raise money, and start a new bourgeois capitalist company.
I view it as the economy is 99% managerial. If we can just keep the 1% of the old model alive, we'll keep getting new things. If venture capital ever gets snuffed—outlawed or it just fails—and there are no more tech startups, then the economy will be 100% managerial. At that point, there will be no innovation forever. People might think they want that. I don't think they actually want that. I don't think we want to live in that world.
Dwarkesh Patel
Will this trend towards managerialism also happen to A16Z as it scales? Or will it be immune? What happens to A16Z in five decades?
Marc Andreessen
At a certain point, this becomes the succession problem. As long as Ben and I are running it, our determination is to keep it as much in the bourgeois model as possible. As you pointed out, it's literally our names on the door. Ben and I control the firm. There's no board of directors. It's just Ben and me running it. It's a private entity with no outside shareholders. So as long as Ben and I are running it in the way we're running it, it will be in the bourgeois model as much as any investment firm could be.
Someday, there's the succession challenge. For tech companies, the succession point is usually when that transformation happens—when it goes from the bourgeois model to the managerial model. The general thing that happens there is that great founder CEOs, when it comes time to hand it off, have basically two kinds of people they could hand it to. They could hand it to somebody like them—a mercurial, idiosyncratic, high-disagreeableness, ornery, entrepreneurial personality. Or they could hand it to somebody who knows how to run things at scale. Almost always, they hand it to somebody who can run it at scale.
There are two reasons for that. The theoretical reason is that it is at scale at that point, and somebody does need to run it at scale. The other is they often have what I call the "long-suffering number two." You've had this high-octane founder CEO who breaks a lot of glass, and then there's the chief operating officer who keeps the trains running on time and keeps everybody from quitting. That long-suffering number two has been in that job for 10 or 15 years and is literally long-suffering. They've always been the underling. Then it's like, "Okay, they now deserve the chance to run the company themselves." That's the handover.
Those founders often end up regretting that decision. In later years, they'll tell you, "Boy, I wish I had handed it off to this other person who was deeper in the organization, younger, more like I am, and maybe would have built new products. Maybe that was a mistake." The fact that they do this over and over again illustrates why the Burnham theory is correct: large, complex organizations ultimately do end up getting run by managers in almost all cases.
The only optimistic view on that is that it's the transition from these companies being in the bourgeois capitalist model to the managerial model that creates the opportunity for the new generation of startups. In the counterfactual, if these companies remained bourgeois capitalist companies for 100 years, they would be the companies to create all the new products. We wouldn't necessarily need to exist because those companies would just do what the startups do—build all the new stuff. But because they won't do that in that model, almost without exception, there's always the opportunity for the next new startup. I think that's good. It keeps the economy vital even in the face of this overwhelming trend towards managerialism.
Dwarkesh Patel
If you had a fund with a 100-year lock-in, what would you be able to invest in that you can't invest in right now?
Marc Andreessen
Our base lockup for venture is 10 years. We have the ability to push that to 15, and for really high-quality companies, we can push that to 20. We haven't been in business long enough to try to push it beyond that. The question if you could push it to 100 years is: is it really time that's the bottleneck? The implication would be that there are more ambitious projects that would take longer that you're not funding because the timeframe is too short.
The problem with a 100-year timeframe, or even a 50-year or 20-year timeframe, is that new things don't tend to go through a 20-year incubation phase in business and then come out the other end and be good. They need milestones. They need points of contact with reality. Every once in a while, a very special company will get funded with a founder who says, "I'm going to do the long-term thing." They go into a tunnel for 10 or 15 years while they're building something, and the theory is they'll come out the other side. These have existed and do get funded. Generally, they never come out with anything. They end up in their own private Idaho—their own internal world. They don't have contact with reality. They're never in the market. They're not working with customers. They become bubbles of their own reality.
Contact with the real world is difficult. Every single time, the real world is a pain in the butt. To mark-to-market your views of what you're doing with the reality of what anybody's actually going to want to pay for requires you to expose yourself to that. It's really hard to do that in the abstract or to build a product that anybody's going to want to use. So this thing where people go in a tunnel for 10, 15, or 20 years doesn't go well. I think 100 years would be an even more degenerate version of that. Best case, it would be an unbounded research lab that would write papers, and maybe something comes out the other end in the form of some open-source thing. But they're not going to build an enterprise that way. Having some level of contact with reality over the first five to seven years is pretty important.
The other way to get at your underlying question is: what if you just had more zeros on the amount of money? What if instead of funding companies for $20 million, you could fund them for $2 billion or $20 billion? They'd operate on whatever five or 10-year timeframe, but you could fund them with $20 billion of venture financing instead of $20 million. I think that's a more interesting question. There are pretty big, fundamental things that could be built with larger amounts of money in this entrepreneurial model. Every once in a while, you see these giant, world-changing things like Tesla and SpaceX that took a lot of money and had really big impact. So maybe that's something the venture ecosystem should experiment with in the years ahead. I would be more focused on that as opposed to elongating the time.
Dwarkesh Patel
But what about basic research? I think you have spoken about the dysfunctions of the academic government research complex. Within the next internet, the next thing that the Andreessen from 10 years from now is building on top of—maybe there needs to be some sort of... if the government effort is broken, maybe you just need to bootstrap something yourself? Have you considered that?
Marc Andreessen
Yeah. So the strong version of this argument is from a guy named Bill Janeway. He was a legendary VC. Janeway is a great, wonderful guy. If people haven't heard of him, he was a PhD in economics. I think he's a student of a student of John Maynard Keynes, so he comes from a highly pedigreed economic theory background. Then he was himself a legendary venture capitalist in his career. He became a hands-on investor at the firm Warburg Pincus and funded some really interesting companies. He's one of these rare people who's both theoretical and practical on this kind of question.
He wrote this book which I really recommend, it's called *Doing Capitalism*, where he goes through this question. The argument he makes is a little bit pessimistic. He argues that if you look at the history of professional venture capital, which is now a 60-year journey, or maybe 50 years from the late 60s, early 70s in its modern form, the big category that's worked is computer science. Then there's a second category that's worked: biotech. And he said, at least at the time of writing, everything else didn't work. All the money people poured into cleantech and all these other areas that venture capitalists tried to fund, from a return standpoint, they just didn't work. You just burn the capital.
When he wrote the book, he ran the numbers and it's like computer science worked twice as well as biotech or something like that. He said this is a direct result of federal research funding over the previous 50 years. Computer science based venture capital was able to productize 50 prior years of basic research in computer science—information science, information theory, communications theory, algorithms. All the stuff that was done in engineering schools from 1940 through 1990. We are productizing that. Then biotech was productizing the work that the NIH and others put into basic research in the biological sciences. That work really started kicking in in the 60s and 70s, a little bit later.
He said the problem is there aren't other sectors that have had these huge investments in basic research. There has been no huge backlog of basic research into climate science or any of the other sectors where people burn a lot of money. He says if you want to predict the future of venture capital, you just look at where the previous 50 years of basic research and federal research funding has happened. His strong form of it is there are no shortcuts on this. If you're trying to do venture capital in a sector that doesn't have this big install base of basic research, you're basically just tilting at windmills.
I think there's a lot to his argument. I'm a little more optimistic about a broader spread of categories. A big reason is that computer science in particular now applies across more categories. This is the underlying point of the "software is eating the world" thesis. Computer science used to be just an industry where people made and sold computers, but now you can apply computer science into many other markets—financial services, health care, and many others—where it can be a disruptive force. I think there's a payoff to computer science and software that can apply in those sectors. Maybe some of the biological sciences can be stretched into other sectors.
Look, there are a lot of smart people in the world. There are niche research efforts all over the place in many fields doing interesting work. Maybe you don't get a giant industry out the other end in some new sector, but maybe you get some very special companies doing special things. Like SpaceX is a massive advance in aeronautics. It took advantage of a lot of aeronautics R&D. It's not like there's some huge aeronautics venture industry, but there is a big winner, at least one. I think more to come. So I'm a little bit more optimistic and open-minded. I think Bill would probably say that I'm naive.
Dwarkesh Patel
You mentioned earlier being able to write potentially nine or ten figure checks to these companies like SpaceX or Tesla who might require the capital to do something grand. Last I checked, you guys have $35 billion or something under management. Do we need to add a few more zeros to that as well? Will a16z's assets under management just keep growing or will you cap it at some point?
Marc Andreessen
So we cap it as best we can. We basically cap it to the opportunity set. Our entire model—it's not a single chunk of money. It's broken into various strategies, and we apply different strategies in different sectors at different stages. It's decomposed. We have like six primary investment groups internally in different stages, so that money is broken out in different ways.
We cap it as best we can to the opportunity set. We always tell LPs the same thing: we're not trying to grow assets under management. That's not a goal. To the best of our ability, we're trying to maintain whatever return level we're maintaining. We are trying to eat market share. We'd like to eat as much market share as possible and go fully exploit the available opportunities. We'd like to fund all the really good founders and back all the interesting new spaces. But we wouldn't want to double assets under management in return for 5% lower returns. That would be a bad trade for us.
To put another zero on that, I think we would need a theory on a different kind of venture capital model, which would be basically trying to back much larger scale projects. Again, I think there's a really big argument you could make that that's precisely what firms like ours should be doing. There are these really big problems in the world, and maybe we just need to be much more aggressive about how we go at it. We need founders who are more aggressive, and then we need to back them with more money.
The counter argument on the Tesla and SpaceX examples is that they didn't need it. They raised money the old fashioned way, round by round in the existing venture ecosystem. So for whatever limitations you think the existing ecosystem has—maybe it's not ambitious enough or whatever—it did fund Tesla and SpaceX. So maybe it works.
This goes to the underlying question. The underlying question underneath all this is not the money part. The underlying question is: how many great entrepreneurs are there? And then how many really big ideas are there for those entrepreneurs to go after? That goes one level deeper: what makes a great entrepreneur? Are they born? Are they trained? What made Elon, Elon? What would you need to do to get 10 more Elons? A hundred more Elons? A thousand more Elons? Are they already out there and we just haven't found them yet? Could we grow them in tanks?
Dwarkesh Patel
Did you say testosterone in the water supply?
Marc Andreessen
Yeah. Or do we need a different kind of training program? Does there need to be a new kind of entrepreneurial university that trains entrepreneurs? Those are the underlying questions. If you show me 10 more Elons, I'll figure out how to fund their companies. We work with a lot of great founders, and we also work with Elon. He's still special. He's still highly unusual, even relative to the other great entrepreneurs.
Dwarkesh Patel
Let's talk about crypto for a second. When you're investing in crypto projects, how do you distinguish cases where there's some real new good or service that a new technology is enabling and cases where it's just speculation of some kind?
Marc Andreessen
What we definitely don't do is the speculation side. We just don't do that. I mean that very specifically: we're not running a hedge fund. What we do is apply the classic venture capital one-on-one playbook to crypto. We do that the exact same way we do with every other venture sector we invest in, which is to say we're trying to back new ventures.
In crypto, that venture might be a new company or it might be a new network, or it might be a hybrid of the two. We're completely agnostic as to which way that goes. We write our crypto term sheets where even when we're backing a crypto C-corp, we always write in the term sheets that they can flip it into being a tokenized network anytime they want to. We don't distinguish between companies and networks.
We approach it with the Venture Capital 101 playbook. We're looking for really sharp founders who have a vision and the determination to go after it. Where there's some reason to believe there's some sort of deep level of technological economic change happening—which is what you need for a new startup to wedge into a market. There's a reason for it to exist. There's a market for what they're building. They're going to build a product and there's going to be an intersection between product and market. There's going to be a way to make money. The core playbook.
We go into every crypto investment with the same time frame we go into venture investing: at least a five to 10 year time frame, if not a 15 to 20 year time frame. That's what we do.
The reason that's not necessarily the norm in crypto is an artifact of the fact that crypto tokens tend to publicly float a lot sooner than startup equity floats. Let's say we're backing a new crypto network. It goes ahead and floats a token as one of the first steps. It has a liquid thing years in advance of when a corresponding normal C-corp would.
There's a thing in behavioral economics: when something has a daily price signal and you can trade it, people tend to obsess on the daily price signal and they tend to trade it too much. There's all this literature on this. It's part of the human experience. We can't help ourselves. It's like moths to a flame. If I can trade the stock every day, I trade the stock every day. Almost every investor in almost every asset class trades too often in a way that damages their returns.
As a consequence, a lot of the investment firms that invest in crypto startups are actually hedge funds. They're structured as hedge funds. They trade. They have trading desks. They trade frequently. They have a public book in hedge fund land. They have these crypto assets they're trading frequently, and then they'll back a startup and trade that startup's token just like they trade Bitcoin or Ethereum. In our view, that's the wrong way.
There's an incentive issue. They pay themselves on a hedge fund model. They pay themselves annually based on the markup of projects that might still be years away from realization of ultimate underlying value. There's a big issue of misalignment between them and their LPs. That's led to these new crypto projects where the tokens are traded too aggressively. In our model, they just shouldn't be. They're not ready for that yet.
We anchor hard on the venture capital model. We treat these investments the exact same way as if we're investing in venture capital equity. We basically buy and hold for as long as we can, and try to have a real focus on the underlying intrinsic value of the product and technology being developed. Basically no speculation. If by speculation you mean daily trading or trying to look at prices and charts, we don't do that.
Dwarkesh Patel
Or separately, another category would be things that are basically the equivalent of baseball cards where there's no real good or service being created. It is something you might think might be valuable in the future, but not because the GDP has gone up.
Marc Andreessen
Oh, baseball cards are a totally valid good and service. That's a misnomer. I would entirely disagree with the premise of that question.
Dwarkesh Patel
But are they going to raise median incomes even slightly?
Marc Andreessen
Sure. There are people who make their living on baseball cards. Look, art has been a part of the economy for thousands of years. Art is one of the original things that people bought and sold. Art is fundamental to any economy. Would you really want to be part of an economy where they didn't value art? That would be depressing.
Dwarkesh Patel
But there's a question of do they value art versus are they speculating on art, and how much of the effort is being spent on speculating on the art versus creating the art.
Marc Andreessen
This gets into an old cultural taboo. It depends what you mean by speculation. If what you mean by speculation is obsessing on daily price signals and buying and selling and churning a portfolio, being a day trader—that's the bad form of speculation. That's the non-productive form.
If by speculation, on the other hand, you mean there are different kinds of things in the world that have different possible future values, and people are trying to estimate those future values and figure out utility and aesthetic value—look at how the traditional art market works. Is somebody supporting a new contemporary artist speculating or not? Yes, maybe from one lens they are. Maybe they're buying and selling paintings, and if it doesn't start going up in price, they flip it and buy something else. But also maybe they're supporting a new young artist. They build a speculative portfolio of new young artists. As a consequence, those artists can afford to be full-time artists. And it turns out they're the next Picasso. That kind of speculation is good and healthy. I think it's core to everything.
I don't know that there's actually a dividing line between that form of speculation and what people call investments. Even when people make investments, look at the institutional bond market. US government debt. People in the bond market today are trying to figure out what that's worth. Is the debt ceiling going to get raised? Even that's up for grabs. That's not speculation in the bad sense. That's a market working properly. People are trying to estimate.
Ben Graham said financial markets are both a voting machine and a weighing machine. In the short term, they tend to be a voting machine. In the long run, they tend to be a weighing machine. What's the difference? Some people would say they're very different. Maybe it's actually the same thing. Why do prices go up? Because there are more buyers than sellers. Why do prices go down? More sellers than buyers. The way markets work is you get individuals trying to make these estimations, and then you get the collective effect.
There's this dirty interpretation of any kind of trading or people trying to do the voting and weighing process. I think it's this historical ancient taboo against money. It's like in the Bible, Jesus kicking the money changers out of the temple. This old taboo against charging interest on debt. Different religions and cultures tend to have some underlying unease with the concept of money, the concept of trade, the concept of interest. I think it's like superstition, resentment, fear of the unknown. But those things are the things that make economies work. I'm all in favor.
Dwarkesh Patel
I don't mean to get hung up on this, but if you think of something like the stock market or the bond market, you can tell a story where what stockbrokers or hedge fund managers are doing is valuable because they're deciding where capital should go. Should we build a factory in Milwaukee? Should we build it in Toronto? Where should capital go? Whereas what is the story for an NFT? What is the NFT helping allocate capital towards? Why does it matter if the price is efficient there?
Marc Andreessen
Because it's art. Let's take the pure form. NFT is a very general concept. It's basically a form of digital ownership. There will be many kinds of NFTs in the future. Many of them will represent claims on real underlying property. I think a lot of real assets are going to get wrapped in NFTs.
But let's specifically take the form of NFT that everybody likes to criticize: a creative project, an image or a character in a fictional universe. That's just digital art. Every criticism people make of that is the same criticism you would make of buying and selling paintings, buying and selling photographs, buying and selling sculpture.
I always like to push this. What's the Mona Lisa worth? One way of looking at it is it's worth the cost of producing it—the canvas and the paint. You can create a completely identical reproduction of the Mona Lisa with $25 of canvas and paint. So the Mona Lisa is worth $25. Or you could say the Mona Lisa is a cultural artifact worth probably a billion dollars or $10 billion. What explains the spread between $25 and $10 billion? Because people care. Because it's art, because it's aesthetic, because it's cultural. It's part of what we've decided is the cultural heritage of humanity. The thing that makes life worth living is that it's not just about subsistence. We're going to have higher values and we're going to value aesthetics.
Dwarkesh Patel
Do you see a difference between funding the flying cars and the SpaceX's and Tesla's versus this? Sure, maybe it improves the aesthetic heritage of humanity, but does one of them seem a different category than the other to you? Or is it all included in the venture stuff you're interested in?
Marc Andreessen
It's a little bit like saying should we fund a Thomas Edison or Beethoven? If push comes to shove and we can only fund one of them, we probably should fund Edison and not Beethoven. Indoor lighting is probably more important than music. But I don't want to live without Beethoven. I don't want the world to.
The point of the world—this is a very important point. The point of human existence. People have lots of views on human existence, religions and philosophies and so forth. What they all have in common, other than maybe Marxism, is we're not just here to get up in the morning, work in a factory all day, go home at night, be depressed and sad, go to bed. We're not just material. Whatever this is all about, it's not just about materiality. There are higher aspirations and higher goals.
We create art, literature, paintings, sculptures, aesthetics, fashion, music, fiction. Why does fiction exist? Why is a fake story worth anything? Because it enhances your life to get wrapped up in a fake story. It makes your life better that these things exist. Imagine living in a world where there's no fiction because everybody's like, "Fiction's not useful, it's not real." No. It's great. I want to live in a world where there's fiction. I like nothing more at the end of the day than having a couple hours to get outside of my own head and watch a really good movie. I don't want to live in a world where that doesn't happen. Funding movies, as another example of what you're talking about, is a thing that really makes the world better.
Here's the other thing. The world we live in is actually the opposite of the world you're alluding to. The world we live in is not a world where we have to choose between funding flying cars and funding NFTs, or funding Edison versus funding Beethoven. The world we live in is the opposite. We have a massive oversupply of capital and not nearly enough things to fund. The nature of the modern economy is what Ben Bernanke called the global savings glut. We have this massive oversupply of capital generated by the last 200 years of economic activity. And there's only one Elon. There's this massive supply-demand imbalance between the amount of capital that needs to generate a return and the actual number of viable investable projects and great entrepreneurs to create those projects.
If anything, we don't have enough flying car startups. We also don't have enough art startups. We need more of all of this. I don't think there's a trade-off. We need more of all of it.
Dwarkesh Patel
Have we reached the end of history when it comes to how venture capital works? For decades, you basically get equity in these early stage companies. You invest more in rounds. It's a two and twenty structure. Is that basically what venture is going to look like in 50 years, or what's going to change?
Marc Andreessen
I think the details will change. The details have changed a lot, and they will change a lot. If you go back to the late 60s, early 70s, the details were different then. The details were different 20 years ago. They're changing again right now in a bunch of ways.
Having said that, I think there's a core activity that seems very fundamental. The term I use, which I borrowed from Tyler Cowen who's talked about this, is "project picking." When you're doing new things—new tech startups, making new movies, publishing new books, creating new art—when you're doing something new, there's this pattern that just repeats over and over again. If you look back in history, it's been the pattern for hundreds or thousands of years. It seems like it's still the pattern.
You're going to do something new. It's going to be very risky. It's going to be a very complex undertaking. It's going to be some very complicated effort that involves a path-dependent journey through a complex adaptive system. Reality is going to be very fuzzy and messy. You're going to have a very idiosyncratic set of people who start and run that project. They're going to be highly disagreeable, ornery people, because that's the kind of people who do new things. They're going to need to build something bigger than themselves.
They're going to need to assemble a team and a whole effort. They're going to run into all kinds of problems and issues along the way. And then there's just this role. Every time you see that pattern, there's just this role where there's somebody in the background who's like, okay, this one, not that one. This founder, not that founder, this expedition, not that expedition, this movie, not that movie.
And those people kind of play a judgment and taste role. They play an endorsement, branding, marketing role. And then they often play a financing role. And then, by the way, they often are very hands-on and they try to contribute to the success of the project.
The historical example of this I always use is that the current model of venture capital is actually very similar to how whaling expeditions got funded 400 years ago. To the point of the term that we actually have, which is carried interest or carry, which is sort of the profit sharing that VCs get on a successful startup. That term actually goes back to the whaling industry 400 years ago, where the financiers of whaling journeys, literally out of Moby Dick, they go hunt a whale and bring its carcass back to land. The carry was literally the percentage of the carried amount of whale that the investors got. It was called carry because it was literally the amount of whale that the ship could carry back.
So if you go back to how the whaling journeys off the coast of Maine in the 1600s were funded, there were a group of what we—they didn't call themselves venture capitalists at that time—but there were a group of basically capitalists. And they would sit in a tavern or something and they would get pitches by whaling captains. And you can imagine the whaling captains, right? A third of the whaling journeys never came back. A third of the time the boats got destroyed and everybody drowned. So it's like, okay, I'm the captain who's going to be able to not only go get the whale, but I'm going to be able to keep my crew alive. And by the way, I have a strategy and a theory for where the whale is.
Maybe one guy's like, look, I'm going to go where everybody knows there are whales. And another guy's going to be like, no, that place is overfished. I'm going to go to some other place where nobody thinks there's a whale, but I think there is. And then one guy's going to say, I'm better at assembling a crew than the other. And the other one's like, well, no, I don't even need the crew. I just need a bunch of grunts, and I'm going to do all the work. And then another guy might say, I want a small fast boat. Another guy might say, I want a big slow boat.
So there's a set of people, imagine in the tavern at candlelight at night, debating all this back and forth, saying, okay, this captain on this journey, not that captain on that journey. And then putting the money behind it to finance the thing. That's what they did then. That's still what we do. So what I'm pretty confident about is there will be somebody like us who's doing that in 50 years, 100 years, 200 years. It will be something like that. Will it be called venture capital? That I don't know. Where will it be happening? I don't know. But that seems like a very fundamental role.
Dwarkesh Patel
But will the public-private distinction that exists now, will that exist in 50 years? That's really public and private market companies, right? Like companies going public? And just the fact that there's different rules for investing in both, just kind of a separate category. Is that going to exist?
Marc Andreessen
Yeah, so that's already there's already shades of gray. So I would say that's already dissolving. There's very formal rules here, but there's already shading that's taking place. In the last 20 years, it's become much more common for especially later stage private companies to have their stocks actually trade, to be semi-liquid, either through secondary exchanges or tender offers. That didn't used to happen in the 1990s. Then it started happening in the late 2000s. And then you've got lots of people with different kinds of approaches to have different kinds of private markets and new kinds of private liquidity.
And then look, you've got these new mechanisms. You've got crypto tokens, entirely new mechanisms popping up, representing underlying value. And then you have big arguments and debates all the time in public with regulators and in the newspapers about what counts as this, who can invest, this whole accredited investor thing. A lot of this is around protecting investors. There's this concept that high net worth investors should be allowed to take more risk because they can bear the losses, whereas normal investors should not be allowed to invest in private companies. But then there's a counter argument that says you're cutting off growth investing as an opportunity for normal investors and you're making wealth inequality worse. So that debate will keep playing out. It'll kind of fuzz a bit. I'd expect probably both sides will moderate a little bit.
In other words, public companies will get to be a little more liquid over time. The definition of what it means to be public will probably broaden out. I'll give you an example. You can have this interesting case where you can take a company private, but yet it's still effectively public because it has publicly traded bonds. It ends up with publicly filed financials on the bond side, even though its stock is private. So it's effectively still public because of information disclosure. And then the argument is, well, if you already have full information disclosure as a result of the bonds trading, you might as well take the stock public again because you're not losing anything. So anyway, it'll fuzz out somewhere in there.
Dwarkesh Patel
Okay, so there's a clear pipeline of successful founders who then become venture capitalists like yourself, obviously. But I'm curious why the opposite is not more true. If you're a venture capitalist, you've seen dozens of companies go through hundreds of different problems. You would think that this puts you in the perfect position to be a great entrepreneur. So why don't more venture capitalists become entrepreneurs?
Marc Andreessen
I think one is it's just harder. It is harder to build a company. It flat out is. It's not easy to be a VC, but it's harder to build a company. And it requires a level of personal commitment. People get to a point, successful venture capitalists do get to a point in life where they start to become pretty comfortable. They make money and they start to kind of settle into a fairly nice way of living. So going back to the 2am chewing glass thing is maybe a little bit of a stretch for how they want to spend their time. So that's part of it.
I think the other part of it is the activities are pretty different. The way I describe it is actually starting and running a company is a full-on contact sport. It's a hundred decisions a day. Anybody who's running a company has to have a bias to action. You're faced with a hundred decisions a day. You don't have definitive answers on any of them. And you have to act anyway because if you sit and analyze, the world will pass you by. A good plan executed violently is much better than a great plan executed later. So it's a mode of operating that rewards aggression, contact with reality, constantly testing hypotheses, screwing up a lot, changing your mind a lot, revisiting things. Thousands and thousands of crazy real-world variables all intersecting.
Being an investor is different. It's much more analytical, clinical, outside-in. The decision cycles are much longer. You get a much longer period of time to think about what you should invest in. You get a much longer period of time to figure out when you should sell. You generally don't want to trade frequently if you're doing your job right. So you actually want to take a long time to really make the investment decisions and then make the ultimately sale decisions. VCs help along the way when companies have issues, but fundamentally it's a much bigger level of watching, observing, learning, thinking, arguing in the abstract, as opposed to day-to-day bloody combat.
It's a little bit like, why don't the great football broadcasters go get on the field and try being a running back for a season? It's a little bit like that, to be totally honest.
Dwarkesh Patel
Got it. How soon can you tell whether somebody will make for a good CEO of a large company specifically? Can you tell as soon as they've got a new startup that they're pitching you? Or does it become more clear over time as they gain more and more employees?
Marc Andreessen
Well, sometimes they've done it before. I'd say this. The big thing with being able to run things at scale, there's actually a very big breakthrough that people either make or they don't make. The very big breakthrough is whether they know how to manage managers. Because running a big company, say you're running a company with 100,000 employees, you don't have 100,000 direct reports. You still only have like eight or 10 direct reports. And then each of them have eight or 10 direct reports, and then each of them have eight or 10 direct reports. So even the CEOs of really big companies are only really dealing with like eight or 10 or 12 people on a daily basis.
The key breakthrough is how you become trained as a manager. Initially, you manage a team of individual contributors. So I'm an engineering manager. I have eight or 10 coders working for me. And then the breakthrough is, am I trained in how to become a manager of managers? If I'm early in my career, I start out as an individual contributor, let's say an engineer. I get trained on how to be a manager of individual contributors. That makes me an engineering manager. And then if I get promoted to what they call engineering director, now I'm a director and I'm managing a team of managers. Anybody who can make that jump now has a generalizable skill of being able to manage managers. And what makes that skill so great is that skill can scale. Then you can get promoted to be VP of engineering. Now you have a team of directors who have teams of managers, who have teams of ICs. And so forth.
At some point, if you keep climbing that ladder, you get promoted to CEO and then you have a team of managers who are the executives of the company and everything fans out from there. So if you can manage managers, at least in theory, you have the basic skill and temperament required to be able to scale all the way up. Then it becomes a question of how much complexity can you deal with? Can you learn enough about all the different domains of what it means to run a business? Are you going to enjoy being in the job and being on the hot seat?
I think most of the people we back, let's put it this way. I think 100% of the people we back have the intelligence to do it. I think maybe half of them have the temperament to do it. And then maybe half of those have the intelligence and the temperament and they really want to do it. By want to do it, I mean 20 years from now, they still want to be running their company. So enough of them where we get the success cases. But as an entrepreneur, you have to really want that. You have to be smart enough and you have to have the temperament and you have to actually want to learn the skills. And not everybody is able to line those up.
Dwarkesh Patel
Managing the manager of revolution.
Marc Andreessen
Yes, exactly. The best case scenario is a bourgeois capitalist entrepreneurial CEO managing a team of managers who are doing all the managerial stuff required at scale. That's the best case scenario for a large modern organization. They're able to harness the benefits of scale and they're able to still build new things. The degenerate version of that is a manager running a company. In theory, people who can build new products, but if the CEO is a manager in the Burnham sense running a team of people who want to build new products, that company probably will not actually build new products. Those people will probably all even start their own companies.
Dwarkesh Patel
Now, as unlikely as this may be, just humor the hypothetical. Let's say A16Z for the next 10 to 20 years has mediocre returns. If you had to guess, looking back, what would be the most likely reason this might happen? Would it have to be some sort of macro headwind? Would it have to be betting on the wrong tech sectors? What would it have to be?
Marc Andreessen
20 years is a long enough time where it's probably not just a macroeconomic thing. The big macro cycles seem to play out over seven to 10 year periods. Over 20 years, you'd expect to get two or three big cycles through that. So you'd expect to get at least some chance to make money and harvest profits. So probably it wouldn't be a macro problem. I mean, you can imagine if we have the Black Plague return, a real pandemic happens. If something horrible happens, then you could be in a ditch for 20 years. But if things continue kind of the way that they have for the last 50 years or 80 years, there'll be multiple cycles and there'll be a chance to make money for people who make good investments. So it's probably not that.
Then there would be the micro explanation, which is we just make bad investments. We invest the money, but we just invest in the wrong companies and we screw up. That's always a possibility and probably the most likely downside case.
The other downside case would build on what I was mentioning earlier from Bill Janeway. It would just be that there's just not enough technological change happening. There wasn't enough investment in basic research in the preceding 50 years in areas that actually paid off. There wasn't enough underlying technological change to provide an opportunity for new entrepreneurial innovation. The entrepreneurs started the companies and they tried to build products and we funded them, and for whatever reason, the sectors in which everybody was operating just didn't pay off. If we hit like five clean tech sectors in a row or something like that, then the whole thing just doesn't work.
That's the scariest one because that's the one that's most out of our control. That's purely exogenous. We can't wish new science into existence. So that would be a scary one. I don't think that's the case. In fact, I think quite possibly the opposite is happening, but that would be the downside scenario.
Dwarkesh Patel
How vulnerable is A16Z to any given single tech sector not working out, whether it's because of technical immaturity or regulation or anything else? If your top sector doesn't work out, how vulnerable is the whole firm?
Marc Andreessen
Innovation could just be outlawed. That is a real risk because innovation is outlawed in big and important areas. Nuclear, for example. I always love meeting with new nuclear entrepreneurs because it's so obvious that we should have this big investment in nuclear energy and there are all these new designs, but the Nuclear Regulatory Commission has not authorized a new nuclear design since its inception nearly 50 years ago. So it's just illegal to build new nuclear in the US. By the way, there are all these fusion entrepreneurs that are super geniuses. The products are great. It looks fantastic. I don't think there's any prospect of nuclear fusion being legal in the US. I think it's just impossible. It can't be done. So maybe it's just all outlawed, in which case, at a societal level, we'll deserve the result. But that would be a bummer for us.
Dwarkesh Patel
Let's say crypto gets regulated or it's just not ready yet or something. What happens to A16Z as a whole?
Marc Andreessen
It's up to our LPs. We raise money on a cycle, so our LPs have an option every cycle to not continue to invest. Logically, the firm is somewhat diversified now. We have like six primary investment domains now. So at least in theory, we have some diversification across categories. We could lose a category or two and the investment returns could still be good and the investors would still fund us. The downside case from there would be that those categories are actually more correlated than we would want them to be. As a firm, we have a big focus on software. We think software is a wedge across each of those verticals. Maybe AI turns out for whatever reason not to work or gets outlawed or something happens, or just fundamentally makes economics worse. Then you can imagine that hitting multiple sectors. Again, I don't think that's going to happen. But I guess it's a possibility.
Dwarkesh Patel
What did the old management of Twitter fail to see about the potential of the platform?
Marc Andreessen
First, I'd say I have a very hard time second-guessing management teams. My belief is that it's so easy to criticize companies and teams from the outside. It's so hard to run these companies. There are always a thousand factors that are invisible from the outside that make it really hard to make decisions internally. The histories are always really screwed up because what you almost always find is that in the histories of the great companies, there were moments early on where it was really tenuous and it could have easily gone the other way. Netflix could have sold out to Blockbuster early on, and Google could have sold out to Yahoo, and we never would have even heard of those companies. So it's really hard to second-guess.
I guess I'd just put it this way. I've always believed, and I was an angel investor on Twitter back when it first got started, I've always believed that the public graph is something that should be just titanically valuable in the world. In computer science terms, Twitter is what's called publish-subscribe, the idea of a one-way public follow graph. That ought to be absolutely titanically valuable. That ought to be the most valuable intent, loyalty, brand signal in the world. That ought to be the most complete expression of what people care about in the world. That ought to be the primary way that every creator of everything interacts with their customers and their audience. This ought to be where all the politics operates. This ought to be where basically every creative profession operates. This ought to be where a huge amount of the economy operates.
They were always on to such a big idea. And then with everything, it's not a question of what kind of product you can build around that and how big you can get people to pay for it. But I've always viewed that the economic opportunity around that core innovation that they had is just much, much larger than anybody has seen so far.
Dwarkesh Patel
But how specifically do you monetize that graph?
Marc Andreessen
There's a gazillion ways. Tons and tons of ways. Elon has talked about this publicly, so it's not spoiling anything. Twitter is a promotional vehicle for a lot of people who then provide you stuff on other platforms. He's talked about video. People create video, they market it on Twitter and then they monetize it on YouTube. Why is that not happening on Twitter? Musicians will have followings of five, 10 million people on Twitter. They aren't selling concert tickets. They ought to sell out concerts.
I first noticed this with Conan O'Brien. When he got famously fired from The Tonight Show, he did this tour. I was a fan of his, so I was following him. He did his first live comedy music tour. He sold out the tour across 40 cities in like two hours. How did he do it? He just put up on his Twitter account. He said, I'm going on the road here of the day. Click here to buy tickets. Boom, they all sold out. Now, click here to buy tickets was not click here to buy tickets on Twitter. It was click here to buy tickets somewhere else. But why isn't every concert in the world, why isn't every live event getting booked on Twitter? As Elon's fond of saying, it's not rocket science.
Dwarkesh Patel
It's funny that a few revolutions in the Middle East were organized in the same way that Conan O'Brien organized this tour, just by posting it on Twitter.
Marc Andreessen
This is the thing that got me so convinced on social media relatively early, even before the Arab Spring. I remember that there was this overwhelming critique of social media between inception and like 2011, 2012. There was a decade where there was just this overwhelming critique from all the smart people that was basically, this thing is useless. This is narcissism. This is just pointless self-ego stroking narcissism. Nobody cares. The cliche always was Twitter is where you go to learn what somebody's cat had for breakfast. Who cares what your cat had for breakfast? Nothing will ever come from any of this.
And then I remember Erdogan, when Erdogan was consolidating control of Turkey, he came out and he said, "I think Twitter is the primary challenge to the survival of any political regime in the modern world." I was like, okay, all the smart analysts all think this is worthless, and then a guy who's actually trying to keep control of a country is like, this is my number one threat. The spread of what that meant, of what the outcomes were, I was just like, oh my god. My conclusion was Erdogan is right, and all the smart Westerners are wrong. Quite honestly, I think it's still early on. I think we're still pretty early in the long arc of social media.
The high-level thing here is that we're new. The world in which 5 billion people are on the internet is still only a decade or so old. That's still really early. The world in which 5 billion people are on social networks is like five years old. It's still super early. If you look at the history of these transitions in the past, like the printing press as a precedent example, it took 200 years to fully play out the consequences. We're still in the very early stages with these things.
Dwarkesh Patel
I was like 10 in 2011. So I don't know if I would have personally caught on if I was older, but maybe not. It's hard to know. But it is kind of interesting. You are basically personally invested in, I think, every single major social media company. So it's interesting to get your thoughts on where that sector might go. Do you think the next 10 years will look like the last 10 years when it comes to big tech? Does it just keep becoming a bigger and bigger fraction of GDP? Will that ever stop?
Marc Andreessen
As a fraction of GDP, it's only going to go up. It is the process of tech infusing itself into every sector. I think that's just an overwhelming trend because there are better ways to do things. There are things that are possible today that were not possible 10 years ago. There are things that will be possible five years from now that are not possible today. So from a sector standpoint, the sector will certainly rise as a percent. I'm putting my money where my mouth is in the following statement: entrepreneurial capitalism will deliver most of that.
A lot of that gain will be companies that were funded in the venture capital, Silicon Valley kind of model, for the basic reason we discussed. You need to have that throwback bourgeois capitalist model to do new things. Incumbents are generally still very poor at changing themselves in response to new technology, for the reasons we've discussed. So I think that process will continue to play out.
Another thing I would highlight is that the opportunity set for tech is changing over time in another interesting way. I think we've been good at going after the very dynamic but small slices of GDP in the last 50 years. And I think more and more now, we're going to be going after the less dynamic but much larger sectors of GDP. So education, health care, real estate, finance, law, government—they are really starting to come up for grabs.
They're very complicated markets and they're hard to function in for startups. It's harder to build the companies, but the payoff is potentially much bigger because those are such huge slices of GDP. So the shape of the industry will change a bit over time. But look, it's very basic. What is technology? Technology is a better way of doing things. At some point, the better way of doing things is the way that people do things. At some point, that does shift market share from people doing things the old way to people doing things the new way.
Dwarkesh Patel
But if you build a better education system somehow, the government is still going to be dumping trillions of dollars into the old education system or the old healthcare system. Do you just accept this as a lost cause? That basically 50% of GDP will just be wasted, but we'll make the other 50% really good? Or when you build the alternatives, can you just accept the loss of the existing system?
Marc Andreessen
Yeah, so let's take education as a great example. I think the incumbent education system is trying to destroy itself. I think the people running it and the people funding it are trying to kill it, and they're doing that every possible way they can.
For K through 12, they're prioritizing the teachers over the students, which is the opposite of what any properly run company would do. At the university level, the problems in the modern university have been well covered by other people. They have become a cartel. The student loan situation—was it Stanford now has more administrators than students? Again, you wouldn't run a company that way.
Dwarkesh Patel
There's a positive vision where you could turn that into one administrator per student.
Marc Andreessen
Yes, that's correct. You could, and they're not. That's exactly right.
And then look, you see the federal student loan crazy thing. And by the way, the universities are voluntarily shutting down use of admissions testing. They're shutting down the SAT, ACT, GRE. They're very deliberately eliminating the intelligence signal, which is a big part of the signal that employers piggyback on top of. They become intensely politicized.
We now know, by the way, the replication crisis—most of the research that happens at these universities is fake. Most of it isn't generating real research results. We know that because it won't replicate. You've got these increasingly disconnected mentalities.
There's some set of people who are going to keep going to these schools. But a degree from a mainstream university that costs, in 10 years, a half million to a million dollars, that has no intelligent signal attached to it anymore, where most of the classes are fake, most of the degrees are fake, most of the research is fake, where they're wrapped up in these political obsessions—that's probably not the future of how employers are going to staff. That's probably not where people are actually going to learn valuable, marketable skills.
The last thing they want is to actually teach somebody a marketable skill. Teaching somebody a marketable skill is so far down the list of priorities at a university, it's not even on the top 20.
A lot of it is just their cartel. They operate as a cartel. It's a literal cartel. The cartel is administered through the quasi-governmental bodies that determine who gets access to federal student loan funding, and those bodies are staffed by the current university administrators. So it's a self-governing cartel. It does exactly what cartels do. It's stagnating and going crazy in spectacular ways. So there is clearly going to be an education revolution.
Dwarkesh Patel
Does that happen today, or five years, or ten years? Does it happen in the form of new in-person institutions versus internet-based? Is it driven by you guys, or is it driven by employers who just get fed up?
Marc Andreessen
I don't know. There are lots and lots of questions about what's going to happen from here. But the system is breaking in really fundamental and obvious ways.
And then healthcare, same thing. Healthcare is just very broadly about outcomes. It's extraordinarily difficult to find positive outcomes in healthcare. In other words, there's lots of activity in healthcare, but it's very hard to find anything that causes people to live longer or to be healthier longer. Every once in a while there's a successful form of cancer treatment or something. But there are all these analyses that show massive investment in public support for health insurance and all these things, and then health outcomes basically don't move.
To the extent that people care at all about the reality of their health, they're going to have to find new ways of doing things. And tech is going to be the wedge of the market for people who have those new ideas.
Dwarkesh Patel
Hopefully these revolutions in education and healthcare are not like healthcare itself, where we're always 20 years away from a cure to cancer, and we're always 20 years away from making education technological.
You've talked about how big tech is two to four times overstaffed in the best case. I'm curious how overstaffed you think venture capital is. How many partners and associates could we let go and there really wouldn't be a difference in the performance of venture capital?
Marc Andreessen
My friend Andy Rachleff, who is a founder of Benchmark and teaches venture capital at Stanford, has a description of this that I think is correct. He says venture capital is always overstaffed and overfunded. His estimate is that it's overstaffed and overfunded by at least 80%. It's overfunded by like a factor of five. It should probably be 20% of the size that it is. There should be 20% of the number of people, 20% of the number of funds, 20% of the amount of money.
His conclusion after watching this for a long time and analyzing it was that it's basically a permanent five times overfunding and overstaffing. It goes to what I referenced earlier: the world we live in just has this massive imbalance of too much money chasing too few opportunities to invest the money productively. There's just too much money that needs long-run returns that looks to venture as part of their asset allocation in the way that modern investors do asset allocation.
The full version of this he describes is that there have only ever been two models of institutional investment. There's the old model, which is 60-40 stocks and bonds, which dominated the 20th century up until the 1970s. Then there's what's called the Swensen model, named after Dave Swensen, who created the Yale Endowment in its modern form. That's the model that all the endowments and foundations have today, and increasingly the sovereign wealth funds, where they invest in alternative assets—hedge funds, venture capital, real estate, things that aren't stocks and bonds.
Anybody following the Swensen model has an allocation to venture capital. On average, maybe that's 4% of their assets. But 4% of the entire global asset base is just a gigantic number.
Somebody once said it's like having a sixth marriage: hope triumphing over experience. The thing you'll hear from LPs is that every LP says they only invest in the top 10 venture capital funds, and then every LP has a different list of who that is. So it's this thing where everybody kind of knows that the whole sector is overfunded, but they all kind of know that they suffer from a real lack of where else the money is going to go.
And look, it's always possible that you'll have some great new fund that's going to do spectacularly well, that's got some great new sector that'll open up. A huge advantage that venture capital has is that the long-dated part of it means you don't suffer the consequences of a bad venture capital investment up front. You get a 10-year lease on life when you make a venture capital investment. You're not going to get judged for a long time. I think that causes people probably to invest more in the sector than they should.
Dwarkesh Patel
Is winner's curse also a big component here, where the guy who bids the most is the one who sets the price?
Marc Andreessen
That can happen. At the early stages, the best companies tend to raise at less than the optimal price because the signal of who invests is more important than the absolute price. Almost every investment that we fund at the Series A stage could raise money at two to four times the price they raised from us, but they value the signal. I think that's also true of the seed landscape, and I think it's still true in a lot of cases at the Series B level. Series C and beyond, it becomes much more of an efficient market.
Again, it's not a full auction. It's not just money. It's not a straight-up liquid financial market. These are whaling journeys.
And by the way, there's a much blunter answer to this question. People who raise seed money and Series A money from the high bidder often end up really regretting it because they end up raising from people who don't actually understand the nature of a whaling journey or a tech startup. Then they panic at the wrong times and they freak out. The wrong investors can really screw up a company.
At least historically, there's a self-correcting equilibrium that comes out of that, where the best entrepreneurs understand that they want people on their team who really know what they're doing, and they don't want to take chances that somebody is going to freak out and try to shut the company down the first time something goes wrong.
Dwarkesh Patel
Got it. But we'll see.
Closing remarks
Hey, everybody. I hope you enjoyed that episode. Just wanted to let you know that in order to help pay for the bills associated with this podcast, I'm turning on paid subscriptions on my Substack at dwarkeshpatel.com. No important content on this podcast will ever be paywalled. So please don't donate if you have to think twice before buying a cup of coffee. But if you have the means and you've enjoyed this podcast or gotten some kind of value out of it, I would really appreciate your support.
As always, the most helpful thing you can do is to share the podcast. Send it to people you think might enjoy it. Put it on Twitter, in your group chats, etc. It just splits the world. I appreciate you listening. I'll see you next time. Cheers.