Bill Gurley & Michael Mauboussin - Putting Theory into Practice

I know firsthand how complex the tech stack is for asset managers, and seemingly every new tool and data source makes the problem even worse, adding more complexity, more headcount, and more risk. Ridgeline offers a better way forward, one unified platform that automates away all that complexity across portfolio accounting, reconciliation, reporting, trading, compliance, and more, all at scale. Ridgeline is revolutionizing investment management, helping ambitious firms scale faster, operate smarter, and stay ahead of the curve. See what Ridgeline can unlock for your firm. Schedule a demo at RidgelineApps.com. Hello and welcome, everyone. I'm Patrick O'Shaughnessy, and this is Invest Like the Best.

This show is an open-ended exploration of markets, ideas, stories, and strategies that will help you better invest both your time and your money. Invest Like the Best is part of the Colossus family of podcasts, and you can access all our podcasts, including edited transcripts, show notes, and other resources to keep learning at joincolossus.com.

To learn more, visit psum.vc.

My guests today are Bill Gurley and Michael Mauboussin. Bill is a general partner at Benchmark Capital, and Michael is the head of Consilient Research for Counterpoint Global. While they are longtime friends with one another, I'd never heard them appear somewhere together, so it was a real treat to be able to do this with the two of them. They are two of the leading minds in their fields, and we combined their decades of experience into one wide-ranging conversation. We discussed the different kinds of increasing returns to scale, the issue of regulatory capture, AI, and hardware, plus a lot more. Please enjoy this great conversation with Bill Gurley and Michael Mauboussin. Well, guys, today's conversation is going to be an excuse to do something that I personally wanted to hear for a long time, which is just the two of you and your longstanding friendship and mutual interest in markets, business, investing, just to hear you talk together.

I don't think I've ever heard you, at least on the record, discuss a bunch of interesting ideas. And I'm going to view my role as cueing some things up and then letting you guys go. I thought I would start with maybe the most obvious place just because it's on everyone's mind and you both thought a lot about it. Maybe, Michael, I'll kick it to you to start. What are your thoughts on artificial intelligence? We're recording this on April 15th in 2024, just to give people context since it's all changing so fast. This seems like the most important thing maybe that's happened in a long time, but also quite hard to parse as an investor. What are your thoughts? What are you thinking about?

I want to kick it over to Bill because I'm sure he knows a lot more about all this topic than I do. A few of my kids are very involved in this from a professional point of view, so I try to glean some from them as well as I watch them do their thing. First of all, this is obviously very profound. It's very important. I don't understand all the details of how the technology works, but I think the most important question as an investor stepping back is going back using Clay Christensen's language of disruptive innovation versus sustaining innovation. So sustaining innovation using Christensen's language is something that is really different. It could be very material, it could be radical, but it's within the same, what he called value network.

It basically helps the incumbents become better at what they do. And then a disruptive innovation is something with a new business model that allows new companies to sort of break in and actually disrupt the more traditional incumbents. So as I look at this, the question is, does this just help the strong get stronger? And by the way, things like—you see some of the big AI players also are doing things like cloud computing. So they're big in multiple facets of this, versus does this allow new businesses to start up and emerge and disrupt the traditional guys? So, at least now, it looks like some of the big guys have the pole position: Microsoft and Amazon and Google and Meta and so forth.

What do you say, Bill?

Well, I think there's a number of observations you can make from a very high level that are important if you were thinking about a lens or a framework for investing in AI. One of them you just mentioned, which is this appears highly choreographed. The windows that open up huge doors for innovators and startups are usually not choreographed. You don't have the whole world saying, 'Look, this is where we're going.' And yet here you do. And I think one of the reasons why people are discussing what you just said, which is could the incumbents have some kind of advantage, is it was so highly choreographed. And give credit to OpenAI and some of the platform players who created APIs and made it very easy and leaned in and went after these people.

So that's one thing that has happened here. There's no one that's kind of asleep at the wheel like people accused Microsoft of being on mobile, which is where you see a lot of openings. So that may take away some of the opportunities. The second thing I would say is you have to separate LLMs from AI. AI is this trend that's been going on for 20 years. There's some amazing stuff happening. The example of Tesla switching to a full AI approach for their self-driving is really like a compelling thing, but it doesn't involve LLMs. And a lot of what we've gotten excited about are LLMs. They have some remarkable things that they can do specifically around writing code, but they also have some limitations because they're mostly about text.

And there's a lot of people I find on podcasts and other things where they're just breathing this expectation into LLMs and AI that I think are going to quickly outpace what's reasonable and what's likely in the short term. When people say stuff like, 'Maybe my computer, when it's idle, will just find a cure for cancer'—that's come out of the mouth recently of a pretty high-level person at one of these foundational model companies—and that's not going to happen anytime soon. And so keeping these things in context, I think, is really, really important. If you are playing around with LLMs, understanding where they're super powerful around language, a lot of the early big wins have to do with analysis and/or writing of lots and lots of text.

And they're remarkable, unbelievable at that.

Bill, how would you think about an investment in a foundational model company? Like if you think about this wave of new entrants that seem to have captured lots of value just in terms of like their valuations, the names we recognize, Anthropic, OpenAI, Mistral, et cetera, Llama—as of it, Facebook's a little bit different. It's these foundational model companies, the market caps or the valuations, I should say, are huge in some cases. How would you approach, like if you saw a marginal one tomorrow, how would you think about it?

It's already evolved to a point that's very similar to where Uber, Lyft, and DoorDash ended up, where there's just so much money moving around that I really don't even think about it as a startup market anymore. Or we should have been having this conversation four or five years ago if you wanted to be accurate, non-consensus in a way around these things. Part of it ties into the big guys being interested in what they're doing with their own balance sheet to reinforce this, which I consider to be remarkably dangerous and unhealthy. Having high market caps—I like to say market caps represent discounted future expectations—so in some ways, you set yourself up. There's an amount of money that's being handed to the experts that are hired by these companies, and all of that's reinforced.

So now the expectation of large secondary payments—this is pro baseball-level money at this point. All fair to them, I'm not complaining. But it's beyond what you would consider to be a startup world. One quick thing: I do think that the initial versions of these models are structurally flawed around providing personal memory and allowing someone to become dependent on one of these things. There are approaches like RAG and things; they don't really seem to get there, from my point of view. But if someone did this correctly, I think that could really change the game. So if someone had a foundational model where they'd figured out how to scale the ability to retune for each individual, or a small local model that people are talking about—Apple may have context for your state of your phone.

Any of that stuff could create a wildly new direction than what we have today.

Maybe a dumb and obvious question, but can you just click more on why the sorts of things that that would unlock that get you so excited?

I always like to reference the movie Her. This thing that walks around and has all of your previous knowledge, knows every email you wrote, knows every conversation I've ever had with Patrick beforehand. So walking in, I could say, remind me the last five times I met with Patrick, and it just starts spitting stuff back at me. And the LLMs today are really more like an advanced Wikipedia that can also code, but that has sucked in the world's information. But it hasn't really sucked in your information, and it doesn't suck in your marginal information. And it would be cost prohibitive today, but I know all of the foundational guys see this opportunity. And I do think it will be the game changer if someone can get it right.

Michael, we've talked a lot in the past about the way that new technology waves get installed, embedded, and then kind of normalized in an overall ecosystem. Everyone's in this sort of Everything Everywhere All at Once mode right now with AI, just expecting it's going to do everything and immediately, which obviously it won't. Can you lend that historical perspective and kind of how these things tend to play out in the abstract in a way that might apply to AI too? Yeah, I like the fact you say it in the abstract, since I don't know that much about what's going on.

But what history would tell us is that as new technologies come in, companies typically have some sort of a workflow and the new technology will come in and help productivity and efficiency for that company, but they don't completely reorient their workflows initially. So the first generation is it amplifies, but kind of in different spots. And then as time goes on, you reorient your business around that core technology that allows you to really unlock the different productivity potentials. Go back to electric engines back in the day, but the most recent example would be the internet where first wave was just sort of adding on to something that already happened. And then you had a whole wave of companies that were internet-first and completely reoriented.

We're probably early in this phase where companies are thinking—every company is talking about this. Obviously, you see this, for example, on conference call mentions and so forth. But very few companies have probably fully integrated this into what they're doing. And it's going to take some period of time to have that happen. I don't know, Bill, if you'd agree with that or if you see that in your own businesses you work on.

I do think if you listen in at the margin, there are people that are like, 'Oh, yeah, this thing's amazing. It can code.' And then people that have been around it, you talk to engineering leaders who've been around it a lot longer: 'The debugging is taking longer than it did before. It's writing code I don't understand. So now I've got to go fully understand the code.' I think people extrapolate at a level that has put us now in this middle land where, of course, there's amazing things happening. Of course, it's unbelievable what's possible, but the expectations have gotten to a place. Another one I hear frequently, especially from the most, we'll call them, AI optimists, is they start talking about universal basic income and productivity gains forever.

It's going to be so awesome. We'll just sit around and drink Mai Tais all day long or something. And I just think that's so impossible, so wrong. Just based on the stuff you and I have talked about over the years, Michael, most notably that competition is ever-present. Just because you get a gain doesn't mean you get to hold on to it forever.

Bill, why do you think that existing big companies have done such an arguably awesome job of facing this thing down this time around? It just seems like universally, all the big technology companies are so all over this that the leadership is doing what it's supposed to. And it makes me wonder, what does a disruptive innovation even mean? If these guys had been running these companies when mobile came, would that have been disruptive or sustaining?

There's a backdrop to that question that is probably a separate question that Michael's been thinking about that is new, which is in 2023, the Magnificent Seven or whatever, we have the largest companies on the American Stock Exchange at least outperforming most of the smaller companies from a growth, arguably, value creation and execution standpoint. And that's unusual in and of itself, whether AI is here or not. With AI, we talked about it a little bit, but because everything's now cloud-based and API-based, it's very easy for people to plug this stuff in. I think the companies that owned the creative productivity apps very quickly realized that this was going to be a required feature. So Microsoft and all their products, Adobe, those kind of things.

But they jumped fast, to their credit. And maybe it's because people like you, Patrick, talk about innovation and learning, you know, for so long that these people have gotten their stuff together. But it's very clear in this case that that happened.

What do you think, Michael, about that interesting phenomenon Bill pointed out, that maybe this is an excuse to have a big long talk on increasing returns to scale or something? But the dominant players, it's not like AT&T from when it was the huge percentage of market cap and it was slow and stodgy and was definitely nothing like Microsoft is today.

In my venture career, when I first joined, the number one question someone would ask an entrepreneur is, 'What are you going to do when Microsoft enters your business?' And it went from that to a full-on IBM equivalence where no one was afraid of Microsoft. They're kind of a really interesting example on this question because they're now back in full-on threat mode.

A few ideas come to mind on this. One is just to echo something Bill mentioned a moment ago, looking at the Magnificent Seven or whatever you want to call it, some sort of leading group of companies. Our numbers show that the Magnificent Seven specifically generated about 45% of the economic profit for the US stock market. So economic profit, just to be clear what that is, it's return on capital, less cost of capital spread. So you're earning above your cost of capital times the invested capital in your business. So it's a dollar sum, which is your return spread times the amount of capital. So the market caps, I don't know what they got, 25 or 30%, and that was actually below their economic profit contribution.

So maybe they don't earn that economic profit in the future, but just to make a point that this is not completely crazy. The second thing we've seen is that—and I'm not sure—what we've seen is the returns on capital for smaller companies used to be bigger or equivalent to larger companies. And starting around the year 2000, that flipped. So now larger companies are not only growing at pretty decent clips, but have very high returns on capital. There's a very interesting book by Jim Besson called The New Goliaths, where he lays out some of the theories as to why this is. I'm sure Bill will talk later about things like regulation and so forth contributing to this. But basically the Besson thesis is that these companies had enormous resources that allowed them to invest substantially in proprietary software.

which allowed them to enjoy the benefits of scale, but also do a lot with customization. And that those technologies did not diffuse throughout the economy. Unlike in past generations, technologies tended to diffuse. So this has given rise to these so-called superstar firms, which is super interesting. The third thing I'll mention is these big guys are spending a lot of money. And Bill mentioned sort of like what people talked about with Microsoft back in the day. And I used to use Microsoft as an example in my class of a company that was spending a lot on intangible assets and not that much on tangibles. So their R&D was a lot bigger than their CapEx back in the day. Those relationships have all changed.

As a little fun fact, if you just take the top five energy companies in the world and the top five technology companies, the technology companies are spending 2x the CapEx as the energy companies. And if you told me 25 years ago or 30 years ago that technology companies would be spending twice as much on CapEx as core energy companies, I think I would have questioned that potential. The amount of money these guys are spending is staggering and very difficult for new companies to come along and keep up with.

Bill, I'm curious when you put your investor hat on again, one of the original posts that you wrote on Above the Crowd, I think back in 2003 or something, was "In Search of the Perfect Business Model." And it talked a lot about increasing marginal utility to customers. And so I want to talk first about the virtuous version of this as a company scales and it actually is better for customers. And then later we'll talk about regulatory capture, which is like the other side of a similar idea or a similar coin. But why did you write that back then? And how does that relate to these massive 7, 8, 9, 10 technology businesses today?

I'll take a short history detour just because we didn't set this up and talk about how we met, but I was remarkably fortunate to land at CS First Boston when I did, primarily because Michael was there. And so I got to know Michael and we became thought partners and would exchange books and ideas. And one of the books we read back then was Complexity, about the rise of the Santa Fe Institute, where we've spent time. But it also outlined Brian Arthur's original thesis on increasing returns and network effects. And probably no other kind of theory affected my venture career more than that and trying to find companies where this is true. And the perfect example is Uber, obviously, where David Sacks had this incredible napkin picture which just showed the loop, but it's pretty simple and it's the answer to your question, which is:

If you have more drivers, pickup times fall. And if pickup times fall, the consumer value proposition gets better, and so more people want to use it. And if more people want to use it in more places, the availability spreads because drivers start popping up to meet the need. And so you end up with more coverage, quicker pickup times, more money for the drivers themselves. It's literally a win for all sides in the system, and those are rare. I think those are rare. When I push entrepreneurs to think about this, and it gets to your increasing of marginal utility, but I just say, as you penetrate a supplier base, one side of a marketplace, whatever, is the value proposition to the other side going up?

And ideally, it would be going up exponentially, which is super hard. It's arguable there are levels of increasing returns, like you could come up with some kind of scale or index because some of them are more linear and some of them can go exponential. But I think they're rare. They're so rare, in fact, that some people think they don't exist, but I'm pretty certain they do.

I think a framework that really helps also flesh all these concepts out is called value-based strategy. It's a famous framework by Brandenberger and Stuart. But the way to think about it is almost like a linear stick. So at the very top would be willingness to pay of the consumer, and then the price of the good or service, and then the cost to the company, and then willingness to sell. Your economic profit as a company is difference between price and cost. But again, these elements of willingness to pay and willingness to supply are very important. Willingness to sell, very important. And just to recast what Bill said, because when he explained Uber to me when I was first learning about it, this very much is the way it came into me, right?

So what happens when network effects—network effects by definition mean that the value of the good or service increases more people use that good or service, right? So willingness to pay goes up. So then you say to yourself, 'Okay, if willingness to pay is going up and I don't raise the price of the good or service, that's more consumer surplus.' More consumer surplus means happier customers and they stick around, and it's very powerful. On the other hand, if I do need to raise prices because willingness to pay has gone up, I have a little bit of headroom to do that. In fact, I used to think about pricing power as this sort of key issue. I no longer believe that's the way to think about it. I think the way to think about it is, am I increasing, as a company, am I increasing willingness to pay every single day?

Again, either adding to consumer surplus or giving myself future potential pricing power if I need it. And it also works on the other side, a little bit on the willingness to sell and the cost as well. So when you start to think about like Charlie Munger's idea of a Lollapalooza effect, you're creating value on the demand side, which is increasing willingness to pay. And because you get old-school economies of scale, your costs per unit are coming down. So more value for the customer, lower cost per unit. And you put those two things together and that builds sort of exponential. Huge value creation happens as a consequence of that. I agree with Bill 100%. I think these things are very rare.

I think people tend to think of them as much more prevalent than they actually are. There was actually an interesting academic study where they looked at companies seeking to become platform companies. And by their reckoning, it was less than, it was like something like one in six or one in seven that were able to achieve that. But when it happens, it can be very powerful from a value creation point of view.

People think about this a lot related to consumer companies, but I stumbled upon something recently that I thought was pretty cool, which is CrowdStrike, which is a huge, wildly successful computer security company. But they posted a memo four years ago titled "The CrowdStrike Security Cloud Network Effect." And the way that it worked for them and the way that the marginal customer ends up with more utility is if they are shared across a network, then if you belong to the biggest network, then you get the shared learning of everyone in that network, which lowers your threat exposure. And you wouldn't want to—this is always how I think about it—look at the value function to the customer and does it go up as the customer count goes up?

And if you were customer N plus 1,000 instead of N, would you have a better value proposition? And in that case, you do. You would never want to be in the second or third place security network. You'd want to be in the largest security network. So it can apply in other places, but I do think it's rare.

What would it take for you, Bill, to make an investment where you felt like you really couldn't see a path to one of these stories of increasing returns to scale across your career or today? It's rare. Not every company can have it. There's still big companies that don't have it.

It happens. There are other dimensions of success in venture than just increasing returns or network effects. And so it could be an early go-to-market advantage that a company has. It could be an incredible founder where you have confidence that those things might happen over time. But I look for them in most cases. I'm probably overly optimistic about being able to create them. I think it's such an unfair competitive advantage. I think it's just such a great way to win. And I find when I provoke entrepreneurs, they can think in ways they didn't in the past, just with some of the dialogue we've talked about. But not everyone can get there. It's hard. There's so many dimensions to corporate success.

You can't do everything. But I look for it all the time, all the time.

Michael, what role do intangibles play in all of these increasing returns to scale concepts and literature that you've reviewed?

I think they're really big. I mean, just to take a step back, I think that the investing world has changed radically in the last 40 or 50 years where, in the 1970s, tangible investments were bigger than intangible investments. And today, that relationship is almost completely flipped where intangibles are substantially larger. But I think the key question is, can we make some links between intangible investments or intangible assets and strategy, the stuff that we're talking about? There are sort of pros and cons on the strategy side. What's great about an intangible asset is it scales like crazy. You write software code, you record a great song, the cost of incremental distribution asymptotes towards zero.

So if it's successful, you can make enormous amounts of money, which is great. The problem is those same products are very much at risk for obsolescence. If you are no longer the most popular song or your code is not the newest code, it's worth very little, arguably zero. Intangibles are easy to steal. So if you come up with a great product, there are mechanisms through patents and so forth to try to protect it. But for the most part, people can replicate what you're doing. And so that makes it very difficult for you to sustain an excess return if other people can mimic what you're doing. And then the last thing that can be very bullish, I think, is something we might talk more about, which is intangible investment assets are subject to pretty quick manipulation and recombination to come up with new solutions.

So really, if you boil innovation down to its core, it's almost always the recombination of existing building blocks. And a lot of those building blocks today are intangible building blocks. Let's call them digital building blocks that we can manipulate in a faster and more effective way than we could before. So there's sort of the intangible investment: first, documenting that it's much bigger than it used to be; and then second, thinking about what are the specific strategic implications? There are positives in the column, and then there are negatives in the column. And then where do we go from there?

Michael, maybe talk about the research you've done on this notion of recombination of ideas, which is one of a subset of different kinds of increasing returns to scale. I'll just list them off just because it might provoke discussion amongst the three of us. First is economies of scale. The second is international trade. The third is learning by doing. The fourth is network effects, which we kind of talked about there. And then this idea you just mentioned, recombination of ideas. Maybe say a bit more—I love these five categories of increasing returns—sometimes within a company, sometimes across an ecosystem. Say one more click about recombination of ideas. And I'm curious how Bill has seen that play out in practice as well.

Just to state the obvious, this is an idea that's been around for a long time. Paul Romer won the Nobel Prize for his work in this area. When you think about economics, the classic way of thinking about this is some sort of production function. So you have input of some sort, usually labor and capital, and then you have some sort of output. And what economists observed for a very, very long time was that output was greater than the input would suggest. And so Robert Solow wrote about this many years ago. It's called Solow's residual. People knew that there was this thing called 'technology,' in quotation marks, that was contributing to being able to manipulate the inputs in some way that allowed for greater output.

So productivity, effectively. The early frameworks, there was an acknowledgment that was the case, but those were almost always exogenous to the model. The model was still about labor and capital. Okay. And so what Romer came along and figured out was we can make this endogenous. We can build it into the framework. This idea of recombination of ideas is, over time, we can figure out how to manipulate the inputs in such a way that allows us to have greater outputs. So that's the core idea of endogenous growth theory. So again, a long intellectual tradition. I learned about this actually from a very different point of view. I learned about it originally from the Santa Fe Institute and John Holland.

John Holland was a professor at the University of Michigan. He's purported to be the first person in the United States to get a PhD in computer science, interestingly, and was a professor of computer science, psychology, and engineering at Michigan. So this guy was extraordinary, and clearly one of the intellectual founders of the Santa Fe Institute. And he developed an idea called genetic algorithms. And the idea is pretty simple. You have a problem you're trying to solve, and you have some sort of fitness function to figure out how close you are. You throw out a bunch of computer programs, for instance, code, and you just test to see how fit they are in solving the problem. And then what you do is you breed...

literally breed the top programs, let them have a series of progeny of the same original length and see what their features are. And then you breed them again and breed them again. So it's just like breeding an animal or something like that. You breed it toward improving your fitness function and solving your problem. I know Bill will probably talk about this, but Matt Ridley's got this much more memorable phrase, like "ideas having sex." This idea of recombining ideas in such a way to get to some sort of objective goal, solve a specific problem. That's really, I think, at the core of a lot of what this stuff is about. Now, going back to Romer—I mean, what Romer emphasizes, these will be slightly technical terms, but this idea of rival versus non-rival goods.

So a rival good is a good that one person can use at a time. So the pen in your hand or whatever, the shirt on your back. And a non-rival good is a good that many people can use simultaneously. And then the second question is, can we keep other people from using those goods? So rival goods are easy to exclude from other people from using them, typically property rights. But non-rival goods are very difficult to have excludability. If it's out there, a recipe or an instruction, it's hard to exclude it. And what Romer came along and said is that some non-rival goods can be partially excludable, which means you can have the benefits of an intangible asset, but you can reap some of the economic benefits from that.

And I think that's sort of the big insight. So there are things that impede this recombination of ideas, but if you want to be optimistic about the world, this is really a foundation to do it, right? Because it's saying like, whatever problems face us, we have more tools in the toolbox, more building blocks to solve it. And because we have digital technologies, it allows us to do things, search the space much faster than we could before. And hence we can come up with solutions faster than we could before. If you want to be an optimist, that's certainly one way you could argue for that.

What does this look like in the wild, Bill? When you're engaged with people that are the recombiners of the ideas, trying to build something novel and new, how do you talk about this with them? Do they even think about it as such? Or are they just saying, "Look, now there's mobile and there's GPS, so I'm going to build Uber"? How's your experience of this in the real world?

Some of it might fall under the phrase 'best practice.' And so there are discussions, there are more discussions than ever because of podcasts and whatnot. And somebody is talking about their viral growth loop and how that works. And they come up with a framework and a model and they share that. They put it in a PowerPoint, and then other people are using that kind of thing. In a more concrete form, this is open source writ large. And I think open source is one of the most compelling innovations for human prosperity that we've ever had. Because in Michael's framework, there's nothing that prohibits someone else from using it once you've used it. And it has zero marginal cost. And so you can do amazing things.

Maybe you could argue one of the reasons why the big companies are succeeding is because, with open source, there's no patent protection for the smaller company. And so they can all take advantage of this stuff more quickly. And I do believe that most large companies have shifted over the past 10 years towards an open-source-first technology stack where they used to be an IBM shop or a DEC shop or whatever. And I think that's great for everyone involved. I mean, open source is kind of—there's this phrase used in economics, what is it, Michael? Perfect competition or pure competition. Like if you want pure competition, you have to get rid of patents and whatnot. It's just a wide-open free-for-all, which should be great for consumers.

Yeah. So anyway, I think open source is probably the most demonstrable example of this—ideas that can just be borrowed—and really starts with the farmer that learns how to plow a field and tell someone where to get seeds. And that example works pretty easily in people's brains. But when you take it forward to where we are today and those things are still going on, it's really cool. And I agree with Michael. It's a reason to be optimistic. I often wonder, could we use open source in nuclear manufacturing? Are there some areas where we're stuck from an innovation standpoint where we might be better off?

Now, how about autonomous vehicles? Just a little bit of history on this. This goes back to the work—I learned this from Jim Bessen's book on open source. I did not know this, but about a century ago, the automobile manufacturers together created an open-source consortium to work on various technologies. And so one of the interesting stories he tells in the book is in 1940, General Motors came up with the automatic transmission. Now, they did have some patents and stuff like that, but for the most part, it was very easy for others to essentially reverse-engineer what they were doing and to roll out the automatic transmission. And so within a decade, every manufacturer had offered an automatic transmission.

And this is 10 years with a war in the middle, right? So it was extraordinary. So the technology diffused. And I think the argument that Bessen has made is a lot of big companies today are investing huge sums of money on proprietary software. When I say huge sums of money, $200 billion, $250 billion per year on proprietary software they're not sharing with anybody else, which is really interesting. So in a sense, to capture economic rent, sometimes open source is not helpful for you. So it's a very interesting tension. And I'm with Bill. I think more of this would be better than less of it for overall innovation. But it's interesting, this tension between companies wanting to do it and not wanting to.

There's a really interesting trend that started probably with Android, but where companies that find themselves behind the eight ball from an offensive standpoint launch open source as a defensive strategy. Google did it with Android against the iPhone. And it's hard to go back to that moment in time when Apple launched only on AT&T, but all the handset manufacturers, all the carriers were very, very concerned about Apple. And they were very open to this idea of an open-source solution to combat that. They did it again in the cloud space where Amazon was running away with things with AWS, and Google created this technology called Kubernetes that allowed you to have portability between clouds, open-sourced it, got the Linux Foundation involved, got IBM and others to join along.

And so you're playing defense rather than offense, but you're fighting pseudo-monopolies, if you will. So it's just super interesting. There's a new map initiative that Meta and Amazon, and I think one other are working on to combat Google, which is an open-source data play.

The thing on Android that I think is fascinating in the context of what Bill just said. So, I mean, I was arguing for this idea of trying to increase willingness to pay. And one of the ways to do that is if you have a complementary good or service, so a complement, the idea in economics is if the cost of your complement goes down, the value of your good or service goes up. You sell hot dogs and the store gives away the buns for free, right? The value of the hot dogs go up. So in a sense, I always think about what Google did with Android is thinking like, what was their good or service they're trying to sell? What was their complement? And right. So for them, essentially, they're like, 'We're going to give away the complement.'

So we're gonna make our compliment basically free, which allows the willingness to pay to go off for what we care about, which is advertising. So in a sense, on mobile. In a sense there, it was a very interesting strategy in the context of that whole willingness to pay, increase the value of your good or service, get your compliment, drive it to zero. In this case, they just bought something, gave it away to everybody.

So it's very cool how this works out. So on consumer surplus and value creation, it seems clear that if we shared as a species more, probably we would get more and move faster. Companies and individuals are more selfish than that. They're trying to create and capture value. And there's lots of ways of doing that. Some do it very virtuously with increasing marginal returns, the Costco model, sharing back scale economies, things like this. Others do it more nefariously. And, Bill, you've written and talked a lot about regulatory capture recently, which I'd love to explore as a group here. But I'm also just curious for how you feel your way to the right line. Yeah, you know, everyone will say a drug patent makes sense because without that incentive, the R&D wouldn't happen.

And there needs to be some period that you can earn great returns if you're the one that discovers a novel new drug or something like that. Whereas like Epic Systems or something like, you'll just find literally everyone will bitch about it, except for the people that probably own Epic Systems, which I think has been pretty good to them. How do you explore this soup of incentives and what's best for companies and drives R&D and innovation versus what's good for humanity? Like, it's a big question, but I know you've thought about it, so I feel comfortable asking.

I'd probably push back on the drug comment. I don't know that 17 years makes any sense whatsoever and why that industry gets protected down to a very minute detail of innovation, whereas it's very hard to have a software patent. Like if we wrote software patents the way we write drug patents, the software industry would slow to a halt. And the other thing is, I mean, just take something like Linux, which has been around for 20 years now, but the number of companies that invested billions of dollars in R&D in Linux, there's several. And so just the argument that no one will invest unless you can put a hard patent on it has been proven false. Companies do invest, or for the defensive reasons that I talked about.

So I'm not in that market, but I'd like to see some kind of reform actually in the drug market. I think 17 is too long and they're notorious, based on what I've read, for making small changes and then reapplying the patents. So, like, intentionally drawing out these windows.

Say everything you've learned about regulatory capture and the crusade you've been on to try to get people more aware of it. Well, a widely watched one.

Yeah, I mean, I mentioned this in the speech, but George Stigler won a Nobel Prize at the University of Chicago. He was the one that really should be mentioned as the one that figured this out. But by the way, I think this problem gets worse as the country ages, if you look at it from a very macro standpoint. The more people hang around Washington for a very long period of time, they just learn how to influence it. And that's true on both sides, both the lobbyist side and the senator and congressman side. And then some people flip back and forth, a particularly scary situation. But you look at major pieces of legislation like Dodd-Frank and others, and you see competition goes down, not up after they happen.

And the phrase I use is, "Regulation is the friend of the incumbent." But there's just way too much proximity between the industries that are being regulated and the people that are regulating them. And there's not an offsetting balance to prevent that from happening. There's almost nothing. Someone told me after my speech, I should have included a slide that showed Boeing's headquarters move over time because it was in Seattle and then it was in Chicago for a while. And now it's in Washington, D.C. There's a lot of talk about Boeing these days and a lack of competition and coziness with regulators. And it's a really big problem in the industries that have been around the longest. I think it's a massive problem in healthcare.

Just a massive problem. There's other problems, but there's a massive problem there. I think it's a pretty big problem in finance. I think it's a problem in telecom. The industries that have been regulated the longest are captured. They're trapped.

Patrick, I'll just tack on to that because Stigler, as Bill mentioned, won the Nobel Prize for this. The way to think about this, there is good regulation, and he called it public interest regulation, right? So you're really trying to make sure that things are good for the broad public. And then there's regulatory capture, right? Sort of the bad regulation. People throw these all in one big bucket, but they're actually two very distinct things. And I'm very sympathetic to Bill's point that we've sort of gone away from or keep taking our eye off the ball of public interest regulation, which I think most of us would agree is a good thing, and gotten into this regulatory capture environment. And man, would it be great to not have money, not in Washington, D.C.?

The mix between companies and politics and money seems to have really distorted a lot of outcomes. And so I don't know how you reverse that because obviously it's a whole allocation of power and so forth, but that would seem to be a huge, huge step in trying to get this cleaned up, this whole issue.

It's likely a massive advantage for China, where a more autocratic government can just decide, okay, we're going to start doing things a new way. And where we have just painted on layer and layer and layer of regulation, and we can't reverse those or turn those off as we start to do something. There's been some super interesting articles in the past month or so about how Texas is crushing California from a renewable standpoint. And I don't think anyone thinks that's because the people of Texas care more about renewables. People would argue till they're blue in the face that the California populace is way more liberal, way smarter, cares more for the planet. And I think we all know what it is. It's just bureaucratic bullshit red tape that exists in California.

They can't get out of their own way.

Bill, you've got some great story about, was it some bridge in Pennsylvania that was rebuilt? There was a bridge across from Harvard, Harvard University, you know, over the River Charles. And it took them, it was supposed to be like a year-long project and it cost X. And it took them like 15 years and like 5X because of all the regulatory stuff.

I think that was the Boston Tunnels project. But interestingly, more recently, the I-95 bridge went down and they got it back up in 12 days. You know, this is, to me, the exact same as San Francisco when Xi visits and all of a sudden they clean up. This town that's been dirty for 25 years gets cleaned up in two days. And the governor, I think, in the bridge case, people are saying maybe a presidential candidate. Now, when we start celebrating people moving bureaucracy out of the way, in some ways, we're admitting that we've painted ourselves into a corner. And rather than view that as the exception, how do you make that the new reality? This is where I think China has a massive advantage. They could just clear the decks for good for every project going forward, whereas we're now making exceptions.

Maybe with the TSMC plant in Arizona, they're begging for exceptions to regulation, and we grant it as part of the winning process of getting the bid. We've come so far, we're so far over the line that we celebrate backing up as an exception, but it makes it really hard to think about how would you back up permanently. Sometimes when companies do layoffs, they do zero-based budgeting, which is a great term, but where you say, just imagine we are starting fresh. How would we ever create zero-based regulation? Start over, because we've layered it on so much, it's so problematic.

What did you see happen in the wake of the Twitter layoff decimation, Bill, amongst the community of founders that saw him lay off whatever crazy percentage and Twitter didn't break? What impact did that have?

The story's not over yet, because I think we have to wait until it re-emerges as a successful company. It certainly didn't break, which is an interesting data point in and of itself—stayed up. And I think there are numerous entrepreneurs and venture capitalists who believe that that is a proof point that many of these systems are overinvested and that there is a level of efficiency that you can reach if you're willing to put your mind to it. You could look at the layoffs that Zuck did over the past two or three years and apply the same lens because they're thriving in the absence of that. Yeah. Now, the perspectives on Twitter and Elon and all this have so much built into them that many people, I think, have a hard time looking at that example as a standalone learning point.

The two of you are uniquely qualified for this next question. Michael, you've studied it empirically. Bill, you've been one of the key participants in the ecosystem. We know the benefits of venture capital, that it's been behind the world's biggest companies in many cases. It's an incredible fuel for innovation. If you had to take the other side of it and talk about today's venture capital ecosystem and identify things that you don't like as much or that you're worried about, what would those things be from the inside and from the outside?

I mean, the first thing is just taking a step back and just talking about public equities versus private equity. Private equity having two flavors, venture and buyout. So we'll just focus on venture. First thing to say is, I mean, I'm going to get these numbers wrong, but roughly speaking, the market cap of public equity is probably at year-end, 50 trillion-ish. The assets under management for a US venture, a little over a trillion, maybe a trillion and a half, something like that. And when you zoom in on venture and you talk about what Bill and Benchmark did over the years, the capacity at early-stage just doesn't seem that big to me. I don't know, Bill, you probably know these numbers, but your last few funds were probably $450 or $500 million with an M—million.

Versus these other funds, companies going out and raising multiple billions of dollars. So I think what happened was in a low-rate environment, a lot of investors understandably had to pursue returns. And so they moved out on the risk spectrum and they said, 'Where do we go get returns? And that's going to be private.' So in buyouts, you get higher returns because you're levering businesses, and venture you're getting because you're buying younger companies with higher failure rates. It's become sort of this automatic: 'To get excess returns or higher returns, I need to go out there.' I think there's been much less sensitivity to the actual capacity and the actual experiences people get into these areas.

The other interesting thing about venture, of course, which has been extremely well documented is it's one of the few asset classes with high persistence, which is to say the past winners tend to be future winners and the past losers tend to be, well, often the losers can't keep going, but the winners continue to be winners. you have to get access to these guys. And that's also extremely difficult to do. Just take a step back. Who knows what normal is, but certainly if the US 10 years at four and a half, four and three quarters, we're getting back to more in quotes, more normal return environments in terms of equities and other parts of the credit world. And so perhaps that just turns down people's

Burning desire to get into less liquid and perhaps riskier areas. So those would be the ones I would say. And then Bill's talked a lot about this, but venture is fascinating. Venture in particular is fascinating due to the cyclicality of it. And so a lot of it depends on when you get in and when you get out. And so being sort of attuned to that is tricky, but also obviously seems super important.

Patrick, I'll start with, I've always felt it's funny because you read all these books and you study all these industries when I go out and make investments and you don't use them to reflect on your own. And when I did apply Michael Porter's Five Forces to the venture industry, I'd say it sucks. There's high competition. But the big problem, and this is structural, is there's low barriers to entry and high barriers to exit. And so it's very easy to raise a fund once you do. These days, you're on the field for 15 years, I think. You think it's gone from 10 to 15? And so that money stays in the system. You go into a period like we had three years ago with ZIRP, and the amount of money that's raised is so gargantuan, and then it has to work itself through the system, decimating returns along the way, just from a supply-demand standpoint.

I don't know how to fix that. You'd have to rewrite the standard GP/LP agreement, I think. I'll spend more time thinking about that later, but it is structurally flawed, I think, from a cyclicality standpoint. The other thing that's happened that ties into all the stuff we've been talking about, about the big companies doing better these days, but the number of public companies has shrunk dramatically. And Michael's written about this, but we've gone, I think, almost half. Is that right, Michael?

Yeah, down 46%.

You have to ask why that is. I personally don't think it's healthy because the minute that happens and everyone realizes that there's less companies going public and companies staying private longer, then the SEC, I think in a well-intentioned way, goes, "Oh my God, the average investor is missing out on this asset class. We have to fix this." But what they want to do to fix it is then some kind of institutionalized investing in private companies, which I think will fail massively. Or we're going to tell people they can put their Social Security check into the next Andreessen fund. I don't know how you would even organize that. And then you start doing things like you start trying to create an alternative public market.

You start putting rules in place in the private markets that, in essence, just recreate the public markets. And this may tie into what we've talked about with regulatory capture and bureaucracy. I think Wall Street's become a big company game. The big banks have no interest in a small IPO. Years ago, there were these banks called the Four Horsemen, which were banks that just specialized in small IPOs. And Silicon Valley leveraged them to the hilt. And you just had more companies getting public sooner. But I would rather the SEC try and understand why that's not happening. Why is it? And people like Gurley are saying you got to have a billion in revenue. If all that's true... this game's a very different game than it used to be.

And it has less of the American spirit. Anyone can do it. Anyone can start a company that can go public. If you have to get to a billion in revenue, it's just a totally different game.

Hey, Patrick, can I just take a moment to say why I love Bill so much? It is a little moment of appreciation here. So first of all, obviously, he's extraordinarily smart and curious and is a lifelong learner. So those are all incredibly great qualities. Obviously, I knew him when he came out of school, business school, and was a sell-side analyst and was a great sell-side analyst. Not a good one, but a great one. In part because he embraced and I think utilized really effectively some of the core tools to think about the value of businesses. So things like returns on invested capital and Porter's Five Forces to sort of be rigorous and think about things strategically. Understanding the basic unit of analysis of how companies make money.

And so there are a few analysts that do that, not that many in public markets. But when you get into venture, it seems to be a lot looser around the edges in terms of how people think about business. And one of the things I've always loved about Bill is you can always talk to him and he can always bring things back to some of these core principles, and reminds people, reminds his entrepreneurs, reminds his investors that these really core principles are really important. When you talk about universality of investing, what are we here to do? Buy something for less than what it's worth. I just really appreciate the fact that he's consistently and thoughtfully and obviously very successfully applied a lot of these ideas over multiple decades, multiple cycles to the benefit.

And he shared a lot of his thinking along the way with the rest of us. So just a little moment of appreciation.

I don't know if Bill remembers this, but the first time I ever had lunch with him and Sam Hinckley in the benchmark offices many years ago, I had all these questions kind of like the ones I'm asking today. I don't get to ask one. We spent the whole time talking about how one company that he was working with was allocating its capital, not just theoretical, but applied. I could second the notion. The idea of capital allocation is, I guess, it's the whole ballgame, right? Like it's what we're talking about. That's the topic of our conversation. And we've had this cool period where For, I don't know, 10 years, we had zero interest rates. That should affect capital allocation, how you raise capital, cost of capital, all these things.

And there's like a should-versus-did gap between what companies maybe should have done theoretically under those circumstances. Bill, you worked with lots of them. Michael, you studied lots of them. And what they did do, and I'm just curious—it seems like a real-world experiment that was run for a bad reason, the global financial crisis at its start, but probably one we can learn a lot from. So what did we learn in that period of like should-versus-did and capital allocation inside of businesses?

I'm happy to take a first look at this because we looked at this for public markets, and maybe Bill can talk about from his vantage point. We talked about this period of sort of easy money from 2009, Patrick, as you pointed out, right after the financial crisis through 2021—and of course, a little amplified in 2020 with COVID and central banks around the world coordinating their behaviors. So if you said, 'Well, interest rates are going to be a lot lower'—and they were, by the way; cost of capital was markedly lower—you'd probably expect three things to happen. One would, companies would invest more because your cost of capital, your hurdle rate, just went down. If you have a ranking of projects, more things get funded.

Second is you'd probably expect companies to hold less cash because cash is earning lower returns, not much benefit to that. And third, you might expect them to take on more debt because interest expenses are lower, you could keep the same coverage ratios and have more debt, and so forth. And so we examined what companies actually did in that 13-year period versus the prior 13-year period, which, by the way, included the dot-com boom and bust and so forth. And what we found was almost the exact opposite. Investments were actually down, the one exception being intangibles, but overall investments down. Cash balances actually up for companies, again, led by many of the large companies. And then finally...

leverage levels actually on average went down, didn't go up, which is interesting. So that's sort of the interesting questions, like why did that happen? The other thing is talking about capital allocation, there's just a wonderful research by John Graham at Duke University. He gave the 2022 presidential address for the American Finance Association, which is written up in a paper that came out in the summer of 2022, surveying CFOs over 33 decades to find out what they actually do And one of the big fascinating takeaways was companies actually don't even use the cost of capital. They know what it is. They know what their cost of capital is, but they use a hurdle rate that's roughly twice as high as their cost of capital.

And it doesn't really change. And it's roughly 15%, right? And so it doesn't matter if rates go up or down every day. They don't really care. They use a 15% discount rate or hurdle rate to accept projects, which is interesting. And then they're very, very sticky with things like their capital structure decisions and so forth. The last thing I'll mention is the one thing that actually did tick up in the ZIRP period was buybacks. And it's very interesting because often you hear people say, 'Oh, buybacks, companies will do them.' Like buybacks are done for—the best motivation would be, 'Our stock's undervalued and we're trying to signal that we think it's undervalued.' In reality, companies buy back their stock largely because it offsets dilution from stock-based comp.

And second, given those low interest rates and where multiples were, it was actually accretive—added to earnings per share. And by the way, buybacks are not guaranteed to be accretive to earnings because you're either foregoing interest income from your cash or you're assuming interest expense from debt that you raise. And it's the relationship between that interest expense and the inverse of your P/E multiple, essentially, the math of it. So what happened was we were in a period where buybacks were wildly and, across the board, pretty much positive to earnings. And we are now in a world where they are essentially neutral. The EPS gains from buybacks—that gig is largely over. Again, different companies in different sets of circumstances.

So anyway, I just sort of throw that one out there as this sort of funny takeaway from what you would expect companies to do and what they actually did, because their behaviors are actually markedly different than what the textbooks tell us they should be.

What did you see from the inside, Bill? Like, how good on average are founders—who are often, the successful ones, great at building a product—how good are they at allocating capital, which is the luxury they earn if they're successful at that first thing?

I mean, I think the thing that we lived through with the zero interest rate period—you go back to when we were reading Brian Arthur's work at the Santa Fe Institute, and we were reading so much other stuff, so many other great thinkers. And there's all these types of game theory dynamics and whatnot. And I don't remember when it was, but it was way before the ZIRP period actually happened. I said to Michael one time, 'What if there were multiple players in an increasing returns game, and they knew what the outcome was going to look like? How would it affect their behavior?' And it's actually an interesting conundrum. And I think for the early parts of, let's say, the internet era, most people didn't think that way.

And so maybe Bezos was the only one thinking that way in his case. So he's flooring the gas. Everybody else says, 'You're crazy.' Walmart said, 'You're crazy.' Everybody else says, 'You're crazy.' The other venture capitalists who are backing his competitors give up, and he wins. But we get to the Uber-Lyft wars and all this, and I think everyone knows the game. Everyone knows the game. And it's hard. How far should you invest if you know it's an increasing returns game and you're losing or in second place? What should your strategy be? And the dollar auction example that they use in these finance classes comes up as an equivalency of that. But maybe the reason the Magnificent Seven exists, maybe the reason why those price wars went to where they did, is just everyone got educated.

Everyone understands how these things work. And we're now living in a new world, a new reality with more educated participants in the game. It makes for a bloody mess is what it does. And maybe that's what we're seeing with the foundational models. There's no way, no way possible from my standpoint that the companies, the foundational models, what are they burning? $100 million a year? $200 million a year? There's no way that's high-quality capital allocation from my point of view. There's no chance. But maybe they don't have the alternative. Maybe they can't not do it, which is kind of the trap that we got stuck in with the ride-sharing price wars.

I have a feel question for you, Bill, which is, I can't remember which of your partners it was, Miles or Peter, someone said, we're 30 years into the internet. Surely the lowest hanging fruit models, whether that's a marketplace idea or whatever it is, have been tried. Maybe some things are being retried. AI may enable stuff. But what is like the opportunity set out there for innovative new businesses and companies feel like for you today relative to your long history of really having your fingers in this?

I don't think there's any way to have a point of view there that's interesting. Because the answer I want to go back to is there seems to always be innovation. And if it were so easy to predict exactly where it was, then we'd all just go be the best venture capitalists we could. But they pop up in weird places and new technologies. It's a combinatorial effect of what's happening. Clearly, the experimentation with AI is where so much of that's happening right now. So... Every time I've read, 'Oh, it's all over, no one's innovating,' I don't believe that. There's always something new coming around the corner.

One way to think about that maybe would be, what are the problems that need to be solved that haven't been solved yet? And those problems, that's a shifting set as the world changes. One huge one, which we could point to right away, is energy, energy usage. So Bill's talked a lot about this with nuclear and other things. So there's an example where you definitely could see innovation, how that ties in a lot of the themes we've already talked about. So part of the way I might think about that is: are there things that aren't good now? Are there problems that need to be solved? And do we now have...

...the pieces and tools to put it together in such a way to solve these problems? What about the physical world? I mean, energy makes me think about it. It's been one of these clichés that it's a bad idea to invest in anything that is hard and you can touch because it's just a much harder business to build. But obviously there's great examples of companies like that, Tesla and others, that are fantastic, huge companies now. Bill, what do you think about the physical world, whether it be energy or robotics or other new applications of technology that aren't just pure software all the time everywhere and start to touch the real world again?

I agree with you about the general principle that venture capitalists got away from these categories is simply they don't bend like software does. So the amount of exponential growth—it ties into what we're talking about, ideas having sex—the ability for software to replicate with zero marginal cost and to have low capital costs, it just makes sense that that would be tied to a higher return on your investment dollar. And the other problems that have existed—material science typically didn't follow the same innovation path that, say, Moore's Law did. And so this happened in solar. There were a lot of shots on goal in Silicon Valley. It didn't pay off. And there's been improvement. It's just been more linear.

And then regulatory plays a big role. So if it's energy, your ability to get the government to hand you money may matter way more than what you accomplish in the R&D lab. And are the founders set up to navigate those waters or not? And just because Elon did it, I'm not sure that means everyone can do it, which is another problem. It's pretty heroic what he's achieved, not just there, but SpaceX. The good news is, I guess, for the global populace, is there are plenty of venture capitalists who are standing up today and pounding the table and saying they do hard tech. And so we're going to see investment. I hope they don't get stuck in the regulatory morass, especially around energy and those kind of things.

It's fascinating that it's called Silicon Valley. Like, it started with something that was physical and then went so pure software for so long, and now we're coming back around.

Yeah, it's hard, though. I mean, we used to do a bunch of semiconductor investments, and it got to the point where it was $50 million per silicon, the first tape-out for your alpha chip or whatever. And that's a lot more like some of these biotech plays. In a TSMC world, if you can't get on their schedule, you're toast. And NVIDIA's got a lot of power in that, in keeping you. So it's just a harder world to play in with those dynamics. Whereas with software, no one can stop me from writing software tomorrow.

I want to go back to the increasing returns to scale and the notion of learning by doing. Michael, can you outline this idea, which sounds self-explanatory, but I think is really nuanced and an important point around increasing returns?

Ken Arrow won the Nobel Prize for his work on equilibrium markets and general theory of equilibrium. Ken Arrow was also, interestingly, a very early participant at seminars at the Santa Fe Institute. So, notwithstanding he was sort of a general equilibrium guy, he actually was one of the early guys sort of encouraging work in complex systems. And as a side note, I'll just mention that he was in his nineties at the time, but we were just sort of standing around having a coffee, and he was really proud of the general equilibrium theory work. He goes, 'But it kind of got our whole profession off the track. Understanding complex systems is really vital to markets and the economy.' So I thought that was kind of an interesting observation.

The reason I bring up Ken Arrow is that, notwithstanding his Nobel Prize, I think his most cited paper is actually about learning by doing, which is 1962. And the idea is pretty straightforward: the more you do something, the better you get at it, and you become more productive. And so why is this important for all of us now? Well, just think about different industries. Bill mentioned a moment ago, solar. Solar is a fascinating one. If you have more solar facilities, increase your output, your costs can go down. So the classic formulation is something called Wright's Law, named after this guy, T. P. Wright. And Wright's Law says that for every doubling of cumulative output, your cost per unit goes down by 20%.

He wrote that paper back in the 1930s, 90 years ago, roughly speaking. And it turns out, by the way, lithium batteries for automobiles: perfect Wright's Law. Cost per solar unit: almost perfect for Wright's Law. And in fact, a number of scientists at the Santa Fe Institute tested 60 different technologies to see which model best predicted the actual cost dynamics—the most famous one being, of course, Moore's Law. And they found that Wright's Law actually was the most effective model explaining all this. So the answer is that the more you do something, the better you get at it. Learning by doing is a pretty big deal. And you think about, again, electric vehicles, Tesla, and Tesla's cost-per-unit advantage.

Well, they're just so far out ahead of everybody in cumulative output that their cost per unit is lower. It's one of those why it's increasing returns because if you're ahead of the pack, you continue to expand that lead versus giving it up.

So that's the basic idea. Bill, I'd love to apply that to your time as an investor. If you had to consider your skill today, you've done a lot of investing versus your skill, let's say in 2003 or something or 2004, 20 years ago, where do you think you of today would most trounce you of 2004?

It's funny, man. My initial reaction is that there's not much in venture, precisely because the world changes and it's so dynamic, both from the new technologies that are coming, but also the competitive nature of venture is changing constantly. Different people try different forms and it's just so dynamic. And Benchmark always favored bringing on new young partners. And we found that there's a point at which they start to wildly outperform the old guys. So I have the opposite reaction. I'm not sure what there is. The stuff that you learn that you need to know, you can learn pretty quickly. And then you got to be on the field. And it's a crazy game. It's fun because it's so dynamic, but it's ever-changing.

All right. So Bill, can I ask you a follow-up question to that? Because we wrote a report late last year about the topic of pattern recognition. And a lot of investors say they like pattern recognition. And it turns out the group of investors who seem most enthusiastic about pattern recognition are venture capitalists. And the premise is this idea of it's intuitive expertise. You have this flash, like I've seen this before based on my expertise. And that expertise comes from being around. So to Patrick's question, if you're around 20 years more than your younger version of yourself, do you not have more patterns that you've seen in a sense and ability to anticipate what's going to happen? Or is it a sufficiently changing world, as you point out, that you can't rely on those patterns anymore?

I think the patterns have a half-life and they decay. Or I saw Tobi from Shopify was talking about the VCs passing on them. I fear we were one of those. And one of the rule sets that caused that kind of mindset was in the software world, Intuit was the only company that succeeded with small business and everyone else failed. And so you develop this anti-small business mindset, which keeps you out of HubSpot and Shopify and Twilio. You then miss massive amounts. And so you need a framework that allows you to benefit from the group dynamics of pattern recognition, but to not be holistically tied to any pattern or not let one pattern break. This is a new subject for this discussion, but we always talk in venture, like there's two error types.

Only one of them matters. Missing the big one is all of the game. Getting the negative right is of little value.

It's so asymmetric. So anyway...

There is some. I don't want to say there's none. When you put it at 30 years, I'm just not convinced that the guy that's been a venture capitalist for 30 years has these massive competitive advantages. I don't think there's much data that supports that. And another thing that happens is you just become cynical.

You become rich?

I think you might do, then. Maybe both. Maybe both. Maybe both. Maybe both.

What about this incredible stat that people have cited, but Michael, I saw you revisit recently, and it's obviously like the whole idea of venture is predicated on this power law idea? And it's like a very venture concept. But then when you zoom out and look at all of markets, you see this crazy trend, the idea that a tiny percent of companies not just represent all the market cap, but also like even outperform basic things like T-bills or something like this. And I'm just curious for you both to react to this strange feature of markets. This seems to always kind of be the case, and what its implications are for like how we should think about investing in general.

Well, Patrick, maybe I'll kick off and just give the basic stats on this. This is work done by Hendrik Bessembinder, Arizona State, which is really interesting. It's got some limitations. I don't think people should run too hard with it. But the basic setup is he looked at all public companies since 1926, and what he found was just under 60%. So, just got to let that number sit in your head, 60% fail to earn treasury bill rates. So, by his reckoning, they destroyed $9 trillion of wealth. The other little over 40% did create value. They earned above treasury bill rates, and they created an aggregate $64 trillion of wealth. These numbers are through 2022, by the way. And so 64 minus nine is $55 trillion of aggregate wealth creation in the US market in the last, call it, century or so.

But what's fascinating to your point, Patrick, and this is sort of the venture capital-esque statistic, is that 2% of those companies created 50 trillion of the 55 trillion. So 2% of the companies were 90% of value. You can even go down a bit. The handful of companies were actually a fairly substantial percentage, including Apple and so forth. So obviously venture, the average, Bill, I think, said it's expanding out, but let's say the average venture fund is 10 years, but we have data on the returns for 30,000 venture investments over the last quarter century or so. And yeah, 55 or 60% lose money. It's actually not dissimilar at all, and then you get some that make money, and then you got your right tails that pull up the whole portfolio.

So, as Bill mentioned, that you need to capture the extreme events to make the whole portfolio go. So one is looking at a century, another is looking at, call it, a decade, but you're getting sort of the same essential, almost like a fractal pattern. The other thing to say is interesting is that Bessembinder also collaborated with some folks, academics, and looked at the same markets outside the United States. So they looked at an aggregate 64,000 companies, including developing and developed markets around the world; same basic patterns held true. Really is interesting. So as an investor, obviously you say to yourself, if this is the pattern, I'm sure Bill and his colleagues think a lot about this, but if this is a pattern, how should we behave?

That becomes the interesting question. Should we try to identify those companies? Do we want to make sure that we own those? That the kind of errors we make, we have to own the—guy, the companies are going to create all the value. Even if we own some of the bad ones, it doesn't really make a difference if we own the good ones. It's going to make up for it all day and then some.

If we were to be a fly on the wall in a Monday, all-day Benchmark partner discussion with this topic in mind, what sorts of things does that then constantly bring up amongst you and your partners as you're considering an investment? Like, if we're going to heed this truth, and you just said you can't miss the huge ones, and if you go to Sequoia, they'll all talk about, like, 'Be the person that finds the next logo that goes on that wall.' This seems to be like the game that everyone's playing. What does that feel like tactically every Monday as you're talking about founders and customers? What are the sorts of questions that forces you to ask of young companies?

So, a couple of different things. One is just to embrace that attitude. I think Bruce came back from reading one of Ridley's books and used the phrase, 'What could go right?' at a partner meeting. And so, it's very easy, especially with a big group, with a group bigger than about five, it's very easy to fall into cynicism as a sport, to start taking shots at stuff. And so, having this 'what could go right' attitude—in other words, make the primary part of the discussion, 'How big could this be?' rather than trying to nitpick whether or not it might fail. And so, that's one. Two, I think it requires just exhaustive behavior. You can't stop looking. How would you know that you looked under every rock? There's no way to know that other than to

be exhaustive about it. And so, creating a culture where everyone feels that responsibility is important. And then, you know, never lose at the finish line. Never, ever, ever lose at the finish line. Once you've made a decision as a firm that it's something you want to be in, make it happen.

How would you do that?

Maybe give me, like, the saffron or, like, one ingredient in there. Yeah.

Well, I mean, you have to have built the right relationships. You have to call in the right favors from a reputational standpoint so that you're bringing the reference calls to the finish line. Be tiresome, don't lose on price.

There's the obvious idea that you want to be non-consensus and right. How often would that manifest in the deals that you did that turned out to be great ones where it wasn't that competitive when you invested? Or did the great ones feel really competitive, requiring the favors called in, et cetera, at the time of the round itself?

I'd say it's going to be somewhere right in the middle, Patrick, because a lot of the big returns either have some momentum, even at a very early stage, or they have an individual repeat-entrepreneur-type situation where they're highly competitive. So, it'd be very rare for there to be a 100% absence of competition and then that become big. It's probably happened before, but I think that's a rare event. So, it's somewhere in the middle. If it's a pure jump ball, that's what's happened with a lot of these late-stage rounds these days. I mean, if you're paying 90 billion post for Stripe or OpenAI, you're probably not looking at the type of returns Michael was talking about.

What in the world of technology or just, like, the landscape today has you the most interested or excited, Michael, maybe starting with you, that we haven't talked much yet about today? Yeah.

I would defer to Bill on this one. I guess probably two or three big things. One is, how is this AI thing going to unfold? I think Bill made a really important point, just to reiterate, distinguishing between AI, broadly speaking, and these large language models or generative AI, right? So, those are really two distinct things. We should keep those separated, at least a little bit. That's going to be an incredibly important area to understand. I do think that stuff that's related to the physical world is also going to continue to take on significance. We haven't really been completely overt about this, but when you think about competitive strategy, really what you're after is ability to generate good returns and barriers to entry.

You're getting good returns and it's difficult for someone to replicate what you're doing or to take away some of your economics. Those opportunities obviously do exist in the physical world as well. So that's probably another area just to straddle these two things, I think, are super important. The other area, again, not an area of expertise for me, but it feels like it's really ripe for a lot of change—and we would have said this 10 or 15 or 20 years ago—is healthcare. When you think about AI, you think about generative AI, you think about what could happen in the world of healthcare, how much money goes into it, how poorly managed it is as an industry, broadly speaking, the regulatory hurdles, that just feels like an area, too.

I don't know how easy or hard it is to make money, but that's another area that's got to be ripe for some change.

I would agree on all those fronts. As for the talk I gave on regulatory capture, I look at the energy situation where we know, and it's not just us, it's us leaning on the greater minds of our world, the Steve Pinkers of the world. Like, we know... that the most efficient way to create energy is nuclear. And we stepped off of the learning curve that Michael talked about. And how do you both get the red tape out of the way and get back on that learning curve in a fast way? I mean, it's awesome that for the first time, the Biden administration, I think, is going to help restart a plant. But where you want to go is so much further from that place. It's great to see us reach a guardrail and come back, but how do we accelerate in the other direction?

It probably would take a very active group of people in Washington that would want to see that happen, to rewrite regulation, to potentially incentivize the best and brightest minds to get us back on that curve. On healthcare, I totally agree with the need. I think if you didn't have the regulatory problem you have, entrepreneurs could do so much more. But take a simple construct like price transparency. Any marketplace company in the world would obviously want to expose all pricing so that people can make good decisions. We can't even get that. Congress demanded that the hospital systems disclose their prices, and several of them just sat on their hands and paid the fine. And these are pretty noteworthy institutions.

And the press doesn't really take them down for doing that. It's going to require such a different mindset than where we are today. And I just don't know how you get there. I don't know how you get out of the trap.

So, Patrick, can I mention one thing, too, like Bill mentioned before, zero-based thinking? There are two areas I found fascinating that if you would ask people in the 1950s, what will the future of these areas be? I think they would have said they're very bright. One is psychedelics, which in the 1950s were demonstrably helpful for certain people for things like depression and addiction and so forth. So the medical, careful medicinal benefits were quite clear, and then they just went wildly out of favor for a very long time and are now just getting... back into the mainstream understanding. A lot of people have suffered in between because of it. And the other is nuclear. If a Martian came down and said, "Oh, you guys need energy, energy is fundamental to everything, what do you guys got?"

You show them around, they'd be like, oh, this is the obvious thing you guys should be doing lots of. It's so straightforward and obvious if you took it from a zero-based point of view. But because of all this regulatory and emotional and political baggage, look at Germany now having to reverse decisions because they're recognizing... Anyway, so it's fascinating that you get these things. These are sort of like half-century trends that you have to bend it back around, where if you just said, if I had no prior knowledge or understanding, it would be obvious that these things are beneficial to the world and they're just not where they should be.

I love the zero-based idea applied to ideas. Nuclear and psychedelics just not mixed together. Two closing questions for you both. The first is about working with and being around what I don't have a better word for than just genius. And this is based on a few people that have passed recently and that you guys have both had some interactions with, one or the other. One is Danny Kahneman. Most people listening will have read his work in Thinking, Fast and Slow and heard the name a lot. The other is Cormac McCarthy, who, Michael, I know you worked with and Bill, too, at Santa Fe. And I think, Michael, he was really close with your wife. Such a cool relationship. And Murray Gell-Mann is the third, where these are three people that

are probably genius by any definition. I'm curious about those three, but also just what you both have learned about a certain category of person like that, what they're like and what it's like to work with them. Michael, maybe starting with you.

I think genius is a very fair term for all those folks, but they were, for me, very different experiences. I would probably have to say, if you had to say, pick the person you thought was the smartest person you've ever been around in your life, I would say Murray Gell-Mann. The guy was just extraordinary. And Bill's got a Murray Gell-Mann story, so we'll maybe let him tell his Murray story. But if you said you could only pick one person to recreate the knowledge of humanity, I would have picked Murray. The guy was extraordinary. And he obviously won the Nobel Prize for Physics in 1969, leading light in that area, but incredibly interested in lots of different things. Obviously, a founder of the Santa Fe Institute, so dedicated to understanding various disciplines.

So, just an extraordinary guy. Cormac McCarthy, of a different ilk. Obviously, he was a writer, an amazing writer. He didn't like—at least with me—didn't like to talk that much about his writing. He would talk a little bit about it, but not that much. But I think—and Bill, maybe you can back me on this—I mean, I think if you walked around the Institute, people would say he was like the smartest guy there, even though he wasn't a scientist. So he would sit in on all the meetings. And oh, even I used to have conferences in Santa Fe for business, you know, investors in the 1990s, and someone'd say, 'Cormac has asked if he can sit in.' And I'm like, 'Yeah, absolutely.' So he would come and sit in, like, business presentations.

So the guy was extraordinary, and he could talk about any topic and was an extraordinary storyteller. And then when you read his stuff, I mean, it's almost like—someone says, if, you know, I go over to Bill's house and he says, 'I've got a $3,000 bottle of wine,' because that's how he rolls, I'd be like, 'I'm not good enough. I don't have the quality. I don't have a palate that's refined enough to appreciate how good this wine is.' That's a little bit almost like I feel when I read Cormac. It's almost too good for me. I need to slow down or have someone walk me through it. Danny Kahneman's another guy, and I met him probably close to 20 years ago. First of all, just an extraordinary person in the sense that he made extraordinary contributions to understanding how people behave.

He was incredibly measured, thoughtful. It was like sitting down with your very wise grandfather. There was a recent tribute that was put out about him. And one of the common threads was the fact that he actually not only was willing to be proven wrong, he actually sought views that were different than his own and was almost happy to be proven wrong about something. And his take was, if there's a truth out there and I now no longer harbor a false belief, I'm stepping closer to the truth. And boy, what an inspirational way to lead your life. Every day say, 'If I believe something that's not what it should be, I'm going to step toward understanding the world in a better way.' And that requires an enormous amount of mental energy and an enormous amount of mental flexibility.

But he, above all, has embodied that. And that's leaving aside, Patrick, as you pointed out, reading Thinking, Fast and Slow—like all the lessons in there that people really should internalize. Investing, business, and it doesn't matter, like in your life, these are really, really core concepts.

Bill, what's your Murray Gell-Mann story?

So by the way, before I tell that story, since Michael was so kind to offer thoughts about myself, I would say one of the most amazing things about Michael is his ability to synthesize. And so Michael can go read a book that's too dense for most of us to make it through or go to a lecture from a professor that most of us couldn't stay awake through and come away with the three tidbits that we should all know. And I think part of what lands in Michael's books is this ability to synthesize from some of the smartest people on the planet. And that then gives us the opportunity to be proximate to them. We were visiting Santa Fe a long, long time ago. I'm going to guess late '90s.

Yeah, late '90s. That sounds right. '99, 2000.

So I'm hopefully 30 at the oldest. By that point, we had read the book and we knew who all these people were. And somehow I find myself at a breakfast table with only like four or five people, but one of them's Murray Gell-Mann. And it's not like an organized breakfast. It's like people grabbed a burrito and sat down next to each other. And someone asked Murray what he's spending time on, and he says quantum computing. And they say, how's it going? And he says, well, we're struggling because we can't measure the state of the system without messing up the system. And through what I will only describe as either egoism or ignorance, I blurted without thinking, so this was whatever that is, system one versus system two, the phrase Heisenberg strikes again.

To which Murray's face turned bright red, his head spun on his neck, and he looked right at me and said, 'What did you say?' like in the sternest, meanest tone possible. And I went... I felt like the smallest human on the planet at that moment in time. So it turns out that the thing he's describing is the observer effect, which I now well know. And even on the Wikipedia page for the Heisenberg Uncertainty Principle, these things are often confused and it's highlighted. But this was a pet peeve for him, people that confused the observer effect with the uncertainty principle. And he let me know it.

What about working with genius, Bill? Is genius, is that a useful concept? Are the people that you seek out geniuses? Does that matter? Is that a stupid word to you? What do you think?

I would probably say that the more interesting thing, which Michael talked about, are these people that are just kind of truth seekers that are just always open for discussion and always trying to get to the next place. They bring a very different mentality to the table. Michael and I have spent a lot of time in the past couple of years with Geoffrey West, and just your ability to sit down with them and say, 'OK, you just said this, but my mental model conflicts with it in this way. Can you tell me?' And like for that to be intriguing to them and to be them to be open-minded, to wanting to explain that and to try and figure that out together makes the relationship, at least from our side, so powerful, which is part of why we spend time there.

There are other people who just wouldn't have time for you and just say, 'Get out of here.' But I think if you have that plastic mind where you're always looking to learn, it's really powerful. You can walk away with something from a conversation the other party may not even know they conferred anything to you.

I'll also say Bill gave another brilliant talk, 'Running Down a Dream.' One of the things I took away from that—he features a number of people who are extremely distinguished in their fields: Bob Dylan, Bobby Knight, and Danny Meyer. But I think the point I took away, Bill—hopefully that was the message—is that these guys worked incredibly hard at mastering their craft and put in lots and lots of hours, lots of passion, lots of sweat. So what you see now is the output of all that work without perhaps fully recognizing how much they put into it and how where they are today was formed by all that effort. And by the way, lots of diverse influences. It wasn't just one influence, but lots of diverse influences, which they took and made into their own to some degree.

So I think that's another really powerful lesson We talk a lot at the Santa Fe Institute, like why being exposed to diverse ideas can really be helpful. And again, it's not because you want to take the best of those ideas and use the tools that are helpful for you as you sort out your day and solve your problems. That talk is obviously for old people and certainly for young people is really, really powerful. And it underscores, by the way, also the importance of hard work. And there's no substitute for hard work.

Michael was the second guest on this show way back when. And I think this is our sixth, maybe something like that. I think you two are actually the two most frequent guests. So it's so cool to finally do this with both of you together. We used to call it the 'Mauboussin bump' on my team because every time he would come on, the audience would like double and then stay that way. And so I owe a huge amount of my success. And then the same thing started happening with you. You both seem to have what seems like a felt obligation to do a lot of teaching—not just doing, but teaching, synthesizing, giving back, putting things in a way that people can consume them. And I wonder what you would say to others that have that capacity, that ability to teach and give back in that way, and the benefits

of it to you. Obviously, there's benefits to everyone that's listening to this, to everyone that reads your guys' stuff, that has learned from you along the way. Maybe just as a point of inspiration, say a bit about why it can also be a nice, selfish thing in some ways and can make your life return better because you've both done a lot of it.

It happened to me anecdotally because I became a sell-side analyst, and that's your job, to publish what your thesis is. And so it was a necessity. I didn't do it for the sake of it. Very quickly thereafter, I learned that the broader my distribution was, the more power came back to me, to your point. And then later, I think in venture capital, I learned that having a reputation as a thought leader was very helpful in closing the deal. Back to your earlier question about winning and being able to help. And so the reputational benefit, access—we talked about being able to sit down with geniuses and talk to them. Getting that access is partially tied to reputation. And so that's another benefit.

Along the way, somewhere I learned that I just think better when I write. You've heard this from others, but putting things down in a structured way and having to defend an argument, it ties back into Bezos's six-page letter kind of thing. It just causes you to think better. There's some consequences. You write everything down, you're going to get some stuff wrong, and then people 15 years later will say, 'See?' That comes with it. And then as I've gotten older, the giving back part is really, really rewarding. If someone comes up to you and says, 'Hey, that really helped change my life, my direction,' that's pretty powerful.

Yeah, I don't have much to add to that. I mean, I've been incredibly blessed that basically my career has evolved to the point where a lot of what I get to do is to learn and to share those ideas with others and hopefully to make them more effective at what they're doing. So that's amazing. I would echo what Bill said is that I find that teaching and writing—and they're both to me very related—really compel thinking. I often think I understand something until I really go to write it down. Patrick, that Increasing Returns is a good example where I was familiar with all those ideas and had kicked them around and talked about them for a long time, but really hadn't gone deeply into each of those things.

And just spending time to do that allowed me to understand the links between them, understand them in a deeper way. So writing them down, your thinking and understanding really benefit.

Yeah, well, it's very cool for me to do this with you guys. I'm really appreciative of the time. Michael had such inspiration on my time as a quant. Bill, your inspiration on my time as a private investor has been enormous. This is so cool to do with you both. Thank you for the time. Thank you. Thank you. If you enjoyed this episode, check out joincolossus.com. There you'll find every episode of this podcast complete with transcripts, show notes, and resources to keep learning. You can also sign up for our newsletter, Colossus Weekly, where we condense episodes to the big ideas, quotations, and more, as well as share the best content we find on the internet every week.