No summary for this episode yet.
Welcome to The Knowledge Project. I'm your host, Shane Parrish. I'm the author of the Farnham Street Blog, a website with over 70,000 readers that's dedicated to helping us learn by mastering the best of what other people have already figured out. In The Knowledge Project, I interview amazing people from around the world so that we can all learn from them, expand our minds, and challenge our thinking. On this episode, I have Chris Dixon. Chris is a partner at perhaps the most famous venture capital firm in the world, Andreessen Horowitz, or commonly known as A16Z. We talk about the history of venture capital, why companies fail, the future of artificial intelligence, and the idea maze.
I hope you like this interview as much as I did. I'd love to hear your feedback. I'm at Farnam Street on Twitter. That's at F-A-R-N-A-M-S-T-R-E-E-T on Twitter. Chris, thank you so much for coming on. Thanks for having me. I'm wondering, what's a typical day look like for you? For VC? Yeah, so it's a good question. It's an interesting job. I guess I would divide my calendar somewhat between, I guess, two broad categories of things. One would be meeting with entrepreneurs who are starting companies and raising money and coming to us to talk about potentially investing. And then the other half of the time, working with existing investments to try to help them, you know, everything from maybe they're raising more money or they're trying to recruit somebody or close a sale or some other kind of thing like that.
So roughly, I would say kind of half and half, kind of looking for new things. And what that means is basically a lot of meetings. So it's a big change for some people. Like I had a background in computer programming. There's a great Paul Graham blog post. It's called like maker schedule and manager schedule, I think, where he can trace kind of the schedule of somebody who builds things. So like a computer programmer or a carpenter or what have you versus sort of a manager. And VC is very much a manager. So it's sort of one hour scheduled meetings as opposed to kind of, you know, eight hours of getting into a flow state and thinking about a topic.
So the bad part is it's sort of a lot of state switching and kind of jumping around. The good thing is it's a fascinating job. You get to meet with incredibly smart and passionate people who are doing interesting things and they tell you all about what they're doing. And so, you know, if you're kind of like intellectually curious, it's a pretty amazing job because it's sort of, you know, one minute someone's telling you about some breakthrough in biotech. And the next minute you're talking about data center infrastructure. And the next minute you're talking about computer security. The next minute you're talking about, you know, I don't know what the transportation industry, you know, you name it.
So it's a fun job. Sounds amazing. Can we just back up for one second just so I don't make any assumptions here? So you work for like a venture capital firm, A16Z. I'm an outsider. Can you explain what does that mean? Yeah, so yeah, I worked for Idrisen Horowitz, which is the shorthand is A16Z. It was a firm founded in 2009 by Mark Idrisen and Ben Horowitz. Mark and Ben both prior to that had – Mark had co-founded Netscape, which was the first kind of commercial popular internet browser. And then they went on to do a bunch of other interesting startups. So, and maybe if you want, I can kind of back up and tell you what venture capital is more generally and a little bit of the history of it.
Would that be helpful? Yeah, that'd be awesome. Great. So, basically venture capital, I mean, as an industry started in the, I don't know what, the 40s and 50s or something. I mean, the kind of the practice goes back further. If you go back and look at kind of wealthy, you know, entrepreneurs, even in the, you know, whatever, 200 years ago, A lot of them, after they made money, would then invest in other entrepreneurs who maybe had a good idea and some good new technology, but not enough money to run it. So actually, I read a great book recently about how the world was won. It was about the laying of the first transatlantic cables in the 1840s, which they actually laid the tables onto the ground.
And, you know, if you read about all those things, they were always funded by, you know, in some cases governments, but a lot of times some crazy, you know, entrepreneur who made money and then was investing in some other new thing. Right. So there was always this practice going way back. But then what happened is in like the really, I guess, really in the 1960s, it became kind of formalized as an industry called venture capital. and there were firms kind of around, it all happened in Silicon Valley, around the rise of firms like Intel, Apple, those kinds of things, Microsoft. And there were firms like Sequoia and Kleiner Perkins and other firms that basically started off as people investing their own money in new technology companies.
And then at some point kind of became formalized and became the way it works today, which is today we don't really invest our own money. we invest other people's money. Specifically, we invest, you know, a lot of our money comes from places like universities. But you're not investing your own money? No, no. So, most VCs, I mean, we do invest our own money in the... So, the way it works is we have a fund and we raise money for a fund. Some of that money does actually come from us personally, but a lot of it doesn't. A lot of it comes from, like, for example, large universities. A lot of This is pioneered by places like Yale, for example, very famously started doing this in the 60s, where basically they have their endowment.
You might read about it. Yale has a large endowment. And they basically put some portion of that in bonds and some portion in stocks. And they want to put some portion into other asset classes, as they call them, that have long-term horizons. So basically what's nice about those pools of capital is that they kind of plan things out in a 10-year, even 30-year horizon, which matches kind of our time horizon. So basically that's what this industry is. It's actually – it gets a lot of attention in the press venture capital, but it's actually a very small industry. So there are a few dozen kind of firms that most people – that constitute the majority of the industry.
maybe a few thousand people work in the industry at the most. The amount of money invested is on the order of like 10 to 20 billion per year, which sounds like a lot, but it's actually smaller than the R&D budget for a lot of, you know, for Apple and Google, for example. So as much as it gets a lot of press, it's actually what most people would consider kind of a cottage industry. The tech world, I mean, it seems pretty clubby from the outside. To what extent is that true? Yeah, I guess it depends who you ask. The cynics would say it is clubby and insiderish or something. I would argue, my feeling is it's small and people know each other, but there's an ethos of inclusion.
anyone who's worked in the tech industry for more than a few years has seen people rise media like like you know um incredibly quickly um so you know i knew you know take people like mark zuckerberg or all these kinds of entrepreneurs like this like anyone who's worked in the industry for 10 years has met these people has met you know people like that who are now incredibly prominent back when they weren't and is very used to kind of new very successful people coming out of nowhere and as a result the industry is very i think it's very sort of inclusive and And people just sort of expect, you know, new things to pop up.
And people are very responsive, I think, to, you know, I don't know, new people kind of coming. It's everyone there in Silicon Valley. And by the way, Silicon Valley, people say Silicon Valley. But I think it's also that kind of spirit is now happening in places like New York and L.A. and Canada and Europe and Asia. Do you think that's a byproduct or do you think that's something conscious? Like people are trying to develop the same culture or do you think it's just happening naturally? Why is that? Why is it spreading? Yeah. I think part of it is people see the success of Silicon Valley and want to emulate it.
I think part of it is, you know, I see a lot of people who move to California to join the tech industry and then decide, you know, it's too much of an industry town and then they want to move to New York, for example. Right. To have more diversity. and to have... What's great about places like New York and LA, as an example, I spent time in both places, is you're surrounded with people that are in the arts and media and all sorts of other kinds of industries, and that kind of creates a different creative dynamic. And so I think it's just the natural kind of maturation of an industry, and as it spreads out, you have it kind of propagating to more places.
I mean, China, there's other specific things, like China is its own story probably, where they're really, you know, it's a country that sort of decided that tech is strategic and has invested heavily in it. So it's a multifaceted kind of story there, I think. So the firm that you're a partner at, A16Z, has a stellar reputation. I mean, how did that come about? What do you guys do differently? Yeah, well, so, you know, the kind of philosophy of the firm is a little different than, I think, than the traditional philosophy in the industry. So traditionally the industry, there were basically, there were very few venture capitalists.
And so what happened was if you were an entrepreneur, you had to go and you had to, you know, basically go to one of these 10 or so firms and pitch them your idea. And, you know, I think these firms, they kind of thought of themselves like the little bit of the way that maybe a hedge fund thinks of itself as their job is to come up with theories about where the, you know, the future is going and, you know, pick the best entrepreneurs. And then once they invest, they kind of hang back and kind of monitor their investment in the same way that a hedge fund or someone might. With our firm, we think we've kind of flipped the model where we think of ourselves primarily as a service firm.
So we think of ourselves a way of maybe a law firm or a talent agency or someone would where our first job is to provide services for the entrepreneur. And so we – and our secondary job is to sort of pick the right company. So the service we provide for entrepreneurs, we basically have, we're staffed very differently and structured very differently than most VC firms. We have over 100 employees who are not investors at our firm. Right. Whose sole job it is to help companies do things like recruit employees, you know, build their, you know, their customers base. Imagine if your phone made you feel like a wizard wherever you go.
With Xfinity Mobile, intelligent Wi-Fi and 5G work together like magic to keep you connected throughout the Emerald City. And now through December 16th, Xfinity Internet customers can get an unlimited line included for a year. Learn more at XfinityMobile.com. Xfinity. Imagine that. And see Wicked for Good only in theaters November 21st. Restrictions apply. Requires Xfinity 300 megabit internet or above. Taxes and fees extra. After promo, regular rates apply. Reduce speeds after 30 gigabytes of usage per line. Data thresholds vary. So now that people are seeing success with that, are your competitors copying that model?
I think to some extent, it's also a very different financial structure. So it's hard to copy because basically the traditional structure is basically that VC funds charge fees and then most of those fees go to the partner salaries. Instead, we put our fees towards these operating teams. And so for our competitors to copy us, they'd have to kind of dramatically change their own compensation and pay structure. Which isn't likely, right? Well, I mean, I think what it ends up, I mean, I think we are seeing probably, I think you will see more and more of our kind of style. I think it will probably come from newer firms.
And, you know, look, I mean, my broader view would be, I think it would be great if more firms did what we did. I mean, yes, it would be competition on the one hand. On the other hand, I think it would be good for entrepreneurs and good for the... Better for the system in a way. Yeah, it's better for the system just to have kind of... It's just more alignment between the investors and the entrepreneurs. And it's investors acting themselves more like entrepreneurs who are taking risk. Right. And our sort of compensation is aligned with the entrepreneurs. Like we basically, you know, most of the money, if we make money on this, will be because our company is successful, not because we collect fees.
Right. So, yeah, you know, so we've worked very hard to help entrepreneurs, which we think is primarily where hopefully, you know, our positive reputation comes from. It seems like from an outsider's perspective that companies are staying private longer than longer in the funding cycle than they ever have before. And the valuations for some of these companies, I mean, before they be the rumored valuations before they become public, like Uber at 50 billion or something. What do you see as the implications and second-order effects of this? This seems like an unprecedented kind of scale. That's a great question.
So, I mean, part of the answer is why are they doing it? Because they're able to do it. Because basically what you're basically seeing is that if you read the press, they kind of confuse this issue a lot. They say that VCs are investing in companies like Uber at later stages. Actually, I mean, we don't do those kinds of investments for the most part. Or it's actually what is happening is firms that historically have been public investors. So, for example, Fidelity, T. Rowe Price, Wellington, all of these kind of well-known public market investors have now moved to invest in private companies.
And so they're sort of the firms that are leading a lot of these late-stage investments. And basically, for a variety of reasons, I mean, so it's a complicated story. One reason that companies are staying private longer is the perception among the technology community that the public markets are somewhat short-term focused. So, you know, if you look at, if you just go read whatever the, you know, Barron's or the Wall Street Journal and things, there's an extreme focus on kind of what happens next quarter. Do they make their, you know, their numbers that quarter as opposed to are they investing for the next, you know, five to 10 years?
I just pushed back a little on that. Aren't Fidelity and the T. Rowe prices of the world who control hundreds of millions, billions of dollars in shares in a company, aren't they the ones that could be setting that in the public market to drive the expectations to be longer term? Yeah, that's a good point. And I think that would be a good counter argument. And I'm not saying this is a subtle question. I think this is sort of two sides of the debate. So one side of the debate would say that public market investors are short-sighted. The other side would say what you said exactly, which is there are these very long-term investors.
And to your point, look at Amazon as an example where it seems as though the investors have accepted the idea that they'll be investing for the long term and forego profits for a very long time. But there is on the flip side, Facebook and Google, most prominently, they have dual class stock, which means that they basically, before they went public, the founders structured it so that they could basically never get fired by Wall Street. What do you think of that? I think it's great. I'm a little opponent to that. If you just look at what's happening, I just don't think you can plan technology investment on anything shorter, like just the way technology products and life cycles work.
I think it works on a minimum of, let's say, a three to five-year cycle. And I just think it's very, very hard to have that kind of managed by a committee. So I'm not saying that those founders, you know, it's less that they have sort of superpowers and more just simply that you really need sort of a small group of people or one person who's managing for a very long-term horizon. and it seems to be I mean if you just look at I don't know I just look at what Facebook and Google are doing right now I mean I also have specific experience with some of these companies where it's just I don't want to name the specific companies but some of these public companies which where I think the CEOs felt like they were completely handcuffed or something just simply couldn't to make the kinds of investments they want to make.
So I don't know. I generally think things need to be, in technology, things need to be planned on a longer-term horizon. There are a variety of different ways you could accomplish that. One of them is dual-class stock. I mean, there are other proposals out there, for example, to increase short-term capital gains to disincentivize short-term trading. I think that's another good idea. There's a variety of kind of proposals out there. I do think, though, that in general, that, you know, long term planning, long term thinking is very good for us as a industry, country, you know, world. Yeah, I'm a proponent of long term thinking.
So and I think dual class stock is one of the one mechanism to get there. I don't know if it's the best mechanism, but I think it's one of them. So I have a question, but maybe you can explain funding right before that. But my question is, like, to what extent is the first round of funding really about preserving optionality for the future? so you can double and triple down on success? Or is it more about funding the idea fully? Or like, maybe you can walk me through some of the thinking. From the investor, from the venture capitalist perspective? From your perspective, yeah. I mean, I don't even have a, what are the stages and why would you invest at a particular stage?
And what are you looking for? Sure. Just briefly, yeah. Yeah, so basically there's, I mean, just some quick nomenclature. There's generally like what's called seed investing, which is, you know, one or two entrepreneurs and you know maybe a couple and an idea and that's also called angel investing sometimes and usually it's people writing you know individuals writing checks or small firms and maybe they'll raise something like a million or two million dollars and that kind of gives them enough money to build a small software team that can build a first product okay and that's that's something i used to do um at an you know at a16z it's we do we don't do much of that anymore we do some of that, but not as much because we have a bigger fund and we tend to focus on what's called series A and series B.
And so series A is after a company has, usually after a company has built like an initial version of a product and is now ready to, you know, build, build, build, kind of build out the product more and start selling it or taking it to market. Series B is usually a little later when they've, you know, got some initial results and now we're trying to accelerate those results. And those will typically be in series A, let's say, you know, $10 million might be an average investment size and in series B, $20 million or something like this. And so. And your series A deal usually includes the right first right of refusal, I would guess on further funding or.
Yeah. Yeah. It's, it usually lets us have what we call pro rata rights, which means we're allowed to invest a certain portion in the next round of funding and mathematically it's like enough that we can preserve our kind of ownership percentage so yeah yeah so basically it's like a yeah it just lets us kind of keep investing some not not a whole not the whole round um it's kind of technical but basically the idea is just that we have the right to kind of keep investing some amount yeah so that's that's usually it's the industry's changed a lot like in the past um maybe 10 to plus years ago vcs would actually take control of a company in many different respects, including the board of directors.
Like a hands-on kind of thing? Yeah. And well, it actually could like, there's all these horror stories of like them firing the founders and things. Oh, wow. That's not something that we do. We really just don't even take control for the most part. Most of us do now. We couldn't fire the founders if we wanted to. Not that we do want to, but we couldn't. And if you get a reputation for that, you'd probably stop seeing deals, right? You'd have a very short career, basically. So not that we want to, but even if we did, it's just not the norm to have those kinds of provisions. So basically, for the most part, it's a very simple transaction.
It's actually in all the areas of finance, it's probably the simplest, which is we give somebody, let's say, $10 million in an exchange. Let's say we buy 15% to 20% of the company, which means if the company sells for whatever, $10 million, we get 15% to 20% of that. For the most part, that's kind of what it is. There's a little bit more structure. There's things called preferences. which basically means that we get paid disproportionately more on certain downside cases and things like this. But it's relatively simple. To what extent in this, you say you do more Series A than angel investing.
And if I understood what you were saying correctly, you're more investing on people in the angel stage. I mean, to what extent in the Series A are you investing in people versus investing in the idea? Great question. I think it's Series A. It's definitely, certainly people is 90% of it. And the idea is also important with the proviso that the idea will, at that point, we know it will change. So it's kind of more like you're investing in the general direction of the idea because just the world changes. I'll just give you something like, you know, I remember when Dropbox, I'll just give you just to take an example, Razor Series A.
I think it was like 2008, you know, and at the time it was, you know, it was really pretty mobile. I mean, the iPhone had come out, but it was, you know, much less widespread than it is today. Right. And so, you know, if you look back at the original pitch deck for that company, or let's say for Facebook for that matter, or LinkedIn or all these companies today, Pinterest, none of them really had mobile as a big part of their business plan because mobile just wasn't, you know, it was still feature phones, right? It was still like those little Motorola phones where you type the, you know, you have the little keyboards and stuff.
And so the world, the computing world dramatically changed in the, you know, in the last seven years. And so all of those companies, so, you know, so if you invested in those companies early on, like a Facebook, you knew you were investing in a social network. You didn't know you're investing in a mobile apps company that eventually would buy a messenger and buy Instagram and all those other things. Right. So I would say you kind of directionally, you're investing in an idea and you're investing in people, but you also know the world will change dramatically in unexpected ways. And so, you know, what you really are kind of looking for, it's kind of like, you know, these kind of black swan anti-fragile ideas of you're really looking for kind of, you know, what some people call, you know, optionality, meaning things which, you know, you can't predict the future, but you can see that there are certain scenarios where, you know, these people… You're disproportionately rewarded.
Exactly. I mean, so like the – that's the thing you have to understand. It's hard to understand about the model of VC. It's very hard to internalize, I should say, which is that the best VC funds lose money at least half the time, which means half of our, if we're doing a good job, half of our investments will fail. And then some small portion will be huge hits and some other portion will be modest hits or something. I wonder if most people even get that in the stock market. Yeah, no, it's very skewed in that way. And in fact, it's interesting. I wrote a blog post about it. if anyone's interested, it's on my website, cdixon.
org, called the Babe Ruth Effect. And actually, we have data. I remember reading that, yeah. Yeah, there's a lot of data in the VC industry that actually, interestingly enough, the best firms actually have a higher loss rate, meaning they lose money more frequently than the less. But when they win, the magnitude is so much greater. Exactly. So it's like, I don't actually, I'm not a big sports guy, but in baseball, it's what you call slugging percentage, which is on runs you hit, even if you have more strikeouts. the two tend to be correlated so it takes it it's sort of an unnatural way to think in some ways because when you meet entrepreneurs you're not you're sort of thinking somewhat like will they succeed but you're also thinking probably more about if they succeed how big could it get right and so you have to kind of train yourself to think that way and and frankly train yourself to be accepting that a lot of what you do will will fail um and you know it's a little bit it's just It's one thing to realize that in the abstract and to write a blog post about it like I did.
It's very different to actually experience it because these entrepreneurs are your friends and you're rooting for them. And the reality of this job is you spend a lot of time kind of helping people in tough situations. So if you had to group the failures into kind of three buckets between leadership, execution and idea, what percentiles would you kind of put on those? I think that's a good question. I guess it depends on the stage. It's very different in different stages. But there's some reasonable percentage of the time where the entrepreneur kind of does everything right and just the market, you know, whatever.
You know, it gets, you know, bundled into, you know, Google releases the same product and gives it away for free or something. Right. Whatever it might be that, you know, just sort of like things happen that are beyond your control that just make it, you know, or regulators just decide it's, you know, you create a new kind of drug and the FDA decides it's, you know, not to approve it or something. You know, like there's certain things that are just external factors and that's probably some, you know, it's like, I don't know, I'm just making up a number 25% of the time. There's some external factor that is completely beyond your control.
You know, and then I think some portion of the time, sort of the hypothesis is wrong about the product and the market. And that's a pretty high percentage of the time. I think then the question becomes, you know, I think with really good entrepreneurs, they're able to kind of adapt then. And, you know, as some people call it, a pivot or something where you change what you're doing. And so, you know, that's always an interesting kind of scenario. But I would say that my overall learning, having done this for, I don't know, eight or nine years now, as a, I'm not, I've only been a VC for two and a half years, but I was investing personally before that for whatever, six and a half years.
Successfully too, am I done? Yeah, it was pretty good. I would say my biggest learning is it's probably more people than I ever, like I probably thought originally it was 70% people and now I think it's 98% people or something like that. Like it's a lot of people. That kind of begs the question, like what's the difference then between a bad founder and a good founder, so to speak, not to categorize them. Yeah, I think a lot of it is not necessarily that they're good or bad, but it's what we have a concept we call founder market fit. So the kind of fit between the founder and the market, meaning are they uniquely suited to do something in that market?
And so a lot of times in our business, that means they have a strong technical background. So maybe they have a PhD from Stanford and MIT in computer science. And that's probably, frankly, I don't know, a third, if not half of our investments are like that, or people just with very, very strong technical backgrounds who, you know, worked. I worked in a lab. It's very typical stories. I worked in a, I was at Berkeley and I worked on in their big data lab and I invented this new, you know, open source data analysis tool. And now I want to go make a business out of it. And that's literally a company we funded called Databricks, which is a technology called Spark.
Like, you know, that's probably a third to half of our company. So someone with very, very deep expertise. And then they have to learn, you know, they're obviously their background is in, let's say, computer science or some other technical field. They have to then go learn kind of how to run a business and how to hire people and how to, you know, get customers. But, you know, we kind of make the assumption that that's easier to learn than the opposite. It's easier to teach your computer science business and vice versa. You're never going to teach a business person computer science on the job.
You have to go to school for that generally or have some kind of work, like long work experience. So a lot of it is that. It's like technical expertise. Sometimes it's domain expertise. So, you know, someone will come out of a person comes out of, you know, the media industry or the fashion industry or you name it, right? Whatever industry it might be. And it says, you know, I've been working in this industry and I realized, you know, it's done and there's a whole bunch of things that are done in backwards ways and I have ideas on how to improve them. And, you know, and it comes from years of experience and deep expertise in that field.
That's another common one. You know, another one will be kind of like, you know, maybe like Airbnb where it's just for whatever reason, it seems like those founders, you know, kind of were part of a certain cultural movement that was, you know, around, you know, just sort of maybe it was a generational thing. I don't know what it was, but people, you know, they had been sort of sleeping on friends' couches and things. Right, yeah. Seeing that behavior emerge and sort of, you know, built out, you know, kind of rode that cultural wave. So that's typically a very important, you know, the sort of founder market fit.
I think also a lot of it is just, you know, tenacity. Almost all companies we're involved with run to extreme adversity. I've almost never been involved with a company that didn't have moments of almost failure. And so it's how resilient are the entrepreneurs. How do you go about determining that? How do you go about testing their grit, their tenacity? It's a good question. It's very hard to do. We do spend a lot of time with the entrepreneurs and try to get to know them. I think a lot of it will come in through their personal backgrounds. You know, it's one reason why you'll see a lot of VCs will invest, you know, in repeat entrepreneurs as an example.
Right. So, you know, like if you look at like, you know, Travis who founded Uber, you know, he had been involved. And I think he started two companies before and, you know, had a long track record and people, you know, and he had varied levels of success. But people who knew him spoke very highly of him as a tenacious and resourceful founder. There's lots of examples of that, of people with some kind of track record. If they don't have a track record, it's hard. And it's something you really don't know until the moment the adversity comes. What are the obvious things you're trying to avoid in founders?
Well, I think at the moment, startups are having a moment of kind of pop culture trendiness or something. There's a lot of news articles about startups and venture capitalists. It's become very sexy, right? Yeah, you know, and the social networking moving things. So I think what we're having now is a bunch of people entering the industry who maybe are, you know, coming in for the wrong reasons, who come to, you know, try to make sort of quick money or something and don't. And it's just really just not. They don't appreciate how hard it is. And how do you pick somebody like that out? Like when they come and they present to you, how do you determine that, you know, oh, I think they're in it for the money versus I think they're in it because they're passionate about the idea or some other, you know, narrative that we want to wrap around that?
Yeah, a lot of it's just depth of experience. How, you know, how long have they been working on the problem? And, you know, I'll just give you an example. Like I was a, I was an early investor in Kickstarter and, you know, and they didn't have, you know, Perry and Nancy and the founders of Kickstarter. They didn't have kind of the classic computer science background I described. Right. They had basically been working on the idea for, I'd say seven years at the time. Oh, wow. And, you know, had tried everything to kind of get funded and, you know, and had, and, you know, and you talk to them and, and it was really motivated.
The original idea for Kickstarter Perry was he was living in New Orleans and he was involved in the kind of music and art scene and had wanted to actually serve as a Kickstarter for himself because I think he had tried to organize a thing where like a band that he wanted to come play would come play and he had a bunch of fans who wanted to see them. And he just didn't have a way to kind of coordinate the two things to have the fans put up the money. He didn't have Kickstarter, right? Right. And he kept thinking about that and he kept thinking about, you know, the kind of going back in the history of the arts and the patronage model, you know, the going back to the Renaissance Italy and things and how the Internet could kind of let you reimagine that model.
and you know when you when you talk to him um so i think i invested i don't know when it was like 2008 or 9 um when he was first starting you know it was clear this was a person who was you know this was his kind of white whale he'd been pursuing for you know forever and time yeah and and you ask you could tell just you ask him questions and this was you know he had the depth of thinking he had thought of everything you know he had gone through we have this concept we call the idea maze and the idea maze is sort of the idea that startup ideas aren't really just kind of a static thing. It's kind of like you see the TV or the movies and they have the way they kind of have these, someone has this epiphany and I imagine that it'll be like an intermittent windshield wiper or something.
In reality, it's much more of kind of a maze, meaning like you sort of imagine how the product might work, but then you imagine if the world responds in a certain way or the technology changes in a certain way. Here's how I'll adapt. And you sort of imagine yourself traversing through a maze and at various points in the maze, there might be a dead end or there might be a trap or there might be a prize or something like this. And you don't really know how the maze is going to turn out when you first start, but really obsessed founders will have thought through all the possibilities. And so a lot of what I like to do, at least when I meet with entrepreneurs, is kind of try to traverse that maze with them and understand the depth of thinking that they have, you know, kind of gone through to get there.
And so in a case like Kickstarter, you know, I mean, it was just, it was, I mean, like it was, I'm not saying it was easy, it was an obvious investment or that it was, you know, obviously going to work. But I will say that like it was obvious that they had thought it through very, very deeply. It was a mission for them. It was not, you know, it was sort of a fun, you know, whatever, a new career choice or something or something done for some kind of more mercenary reason. So I don't know. But the answer, look, there's no great science to this. People have tried many, many times to use data science and other things to try to quantify these kind of questions you're asking.
And the results have been pretty poor. It's been very hard to predict these things. So I don't have... I was hoping you had like this secret recipe for us. I wish I did. And I've certainly tried. Many people have tried. A lot of it's like any kind creative endeavor. Um, um, you know, how do you pick a musician early on? How do you pick, uh, you know, a writer early on? Um, there's, there's certainly like having spent years practicing in the field is very helpful, but ultimately a lot of it comes down to kind of an art, I guess. Um, so, so many things do. So switching gears just a little bit here.
Um, what's one thing that you think the future holds that no one is talking about? uh oh good question um i don't know about no one because i think we're very few people then i mean that's like maybe maybe i could change the question just say that some of the things i'm excited about i mean i think some of the investments i've made have been things that are somewhat unpopular um so for example um i'm not unpopular but i would say uh i don't know what controversial or i don't know um so i'm an investor a company called coinbase which is the leading Bitcoin company. So I'm very excited. Bitcoin is an example, which I think is somewhat controversial.
So digital currencies. Digital currencies. I was an investor in Oculus, which is a virtual reality company that Facebook acquired. I'm very, very excited about virtual reality. How do you think that's going to change our lives, virtual reality? So I think it's the next day. I think it will be, like when we look back on the history of computing, it will be, the key milestones will be, I'm at the extreme end of excitement here, but there'll be the PC, the Macintosh or something like this, then the internet, because you have the next key moment, and then the mobile phone, like the iPhone, and then I think virtual reality will be the next wave.
What about artificial intelligence? That's another interesting one. That's sort of separate but related. I'm happy to talk about that too. I think virtual reality, though, will, I mean, I don't know how long it will take. Oculus is going to release their, it's announced they're going to release their consumer product at the beginning of next year. And I think initially for the first year or two, it'll be primarily used by people playing video games. But I think in the next few years after that, it will be used much more broadly. I think it will be the predominant way that people at some point interact with computers and other people at long distances.
I think in 10 years, we'd be probably having this conversation in virtual reality and we'd be, you know, looking at each other across the room and it would feel like we're in the same room. Um, and, uh, and, you know, I think, um, it's going to have, it'll be, the implications are far beyond gaming and it'll be all kinds of, you know, I think movies, um, there's lots of interesting health related applications, communication, social things. Do you think that'll bring the world closer together? I do. I think, I think it will. I think it will. Um, you know, there's some great videos on YouTube. I'd encourage people to go on there and check them out.
If you just search YouTube for Oculus or virtual reality, you'll see a lot of them where there's one the other day where it was a guy who was using virtual reality, a demo to experience. I think it was like the Apollo moon landing or something. And you watch these videos, people are literally crying at the end of the day. I mean, I've never seen a computing medium that has such a strong emotional impact because in that case, it was a guy who had dreamed his whole life about seeing this and he was crying and he was like, I would never be able to see this in any other way. And, you know, I think what it does is it lets people, you know, I think, for example, a big application will be virtual tourism.
So just going and simply, you know, visiting the Great Wall of China and the Taj Mahal, all those other kinds of things. And things which right now are very expensive. And, you know, one of my favorite demos, I say demo now because it's all very early and it's not like they're not full products, but it's a thing called Ocean Rift, which lets you just scuba dive around the ocean and observe different aquatic sharks and underwater. It's almost like having the experience but not going, right? Yeah, exactly. So I think it'll be all sorts of things like that. I think actually the gaming side is probably in some ways overblown because… Right, but we underestimate everything else, you think.
I think so. I think so. So that's VR, so I'm very excited about that. AI, when we talk about AI? Yeah, I'd love to hear more about that. The singularity, let's talk about this whole like. So, okay, there's a lot to say about AI. So I think that, well, I think there's two things. When people talk about AI, they often are really kind of talking about multiple things, right? So there's AI, like there's sort of like the how singularity, like when we have a talking computer. And then there's automation and like our, you know, will computers take jobs away and things like this. So I think a lot of people have what I would call kind of a world's fair view of technology.
So you remember like the World's Fair and I don't know if you see like the Captain America movie or a bunch of other movies where they show, you know, it was like Howard Hughes type guys and they're showing, you know, the Tesla coils and the Android robots and the flying cars. And so a lot of people, you know, sort of a simplified way to think of technology is like the, you know, the robots are coming and people are going to build robots that take away our jobs. If you actually look at how automation works, it's actually, I think, a lot more nuanced and less obvious. So I'll give you an example.
You just kind of take almost any technology company that's on our website that we've invested in. I'll just take an arbitrary example, a company called Zenefits. So Zenefits is a web product that lets you, if you're a small business, go and sign up new employees for health care and other benefits. Right. So it sounds like just like, you know, it's whatever, it's a benefit software, right? Actually, what it ends up doing is it ends up letting you hire fewer people at your company because you now no longer have to hire, you know, somebody to do that job, right? So what I would argue is things like a lot of automation doesn't really look like automation.
It looks like just regular software. Right. That's a lot of ways what AI really is, is sort of taking what smart people do and embedding it in software and giving that software out to lots of people. And so every new piece of software that you see in some ways is a, is sort of a, a piecemeal form of AI. And then when you sort of, when you combine it all together, what you get is kind of this broader kind of functioning super system, all the software interacting together. I think a lot of the real AI, the kind of the real automation ends up sneaking up on you. Now there's this other kind of AI, which is the kind of more spectacular stuff that you read about, which is, you know, it's the headlines.
Yeah. You know, so Siri is an example, like, you know, which is, which is, speech recognition. And then Google's doing a lot of interesting stuff with image recognition. I do think this stuff is at a... A lot of people in Silicon Valley believe that this kind of AI is at an inflection point now. Specifically around a technology called deep learning, which is basically a... I don't know if you remember neural networks. So neural networks are kind of the trendy... There were a lot of books written about them and things in the 90s, which were basically computer systems that were kind of designed to replicate the way that the human brain works.
Yeah. And it was sort of held out as its promise. And then it was sort of there was a letdown afterwards because it didn't kind of deliver the results people wanted. But basically what we've now discovered is it turns out that if you do neural networks and you use a lot more computing power, which we now have available because of Moore's Law, which is the idea that basically all computing gets faster and cheaper very quickly over time. Basically, if you take neural networks and you make them a lot more computing power, a lot more storage, a lot more memory, a lot more networking, a lot more computing resources, it works really well.
And that's what a couple of years ago, Google did a very famous experiment where they basically took, I think it was on the order of tens of thousands of computers, had them study YouTube videos, and at the end of it, those computers were able to correctly identify cats, like whether it contained a cat, with a very high degree of accuracy. That was one of the kind of results they released that really kind of shocked people as to how accurate it was. Because basically a lot of this stuff in AI, and maybe if you use Siri as an example, a lot of it is sort of relatively easy to get to like 80%, 90% accuracy.
It turns out if you just like a regular programmer downloads a bunch of open source software and spends a weekend, you can make like a decent replica of Siri in like a weekend. Oh, wow. Okay. But to get higher accuracy on that is like exponential. Yeah, it gets like 80% really quickly. And then it turns out all of the work is in the last 20%. So like self-driving cars are another good example where if somebody tells you, oh, I saw a self-driving car and it was able to drive on the highway during daylight, that's actually not very impressive. And that's actually something that almost anyone can build.
I mean, anyone with a programming ability can build. what's hard is all of the millions of edge cases. So by edge cases, I mean it's dark, it's raining, a dog jumps out, two dogs jump out, you know, a shadow that looks like a dog jumps out, like, you know, whatever, like you name it. There's a million little special cases where to learn all those different special cases takes lots and lots of additional effort. And so the kind of the big breakthrough with the cat video and Google was that they'd gotten kind of to this point of like 99% or something like that, which no one had ever gotten to before.
It might have been 97. I forgot about the exact number. But it was a very high number. And they've since gone on to do more experiments where they've done things that do like what they call image classification, which is basically take an image and describe what's in the scene. And the results are getting very, very good there. So, you know, you'll take an image and the computer will say, this is, you know, three children eating pizza. And it's right, you know. and like things like that. And so, um, so there've been a lot of really promising results and it's still very early. I mean, you know, look at your phone and, and the autocorrect.
And every day you'll see it makes like ridiculous mistakes. Right. I mean like, so, you know, we can't, we can't make an autocorrect today that, that, that, that seems to work, you know, even most of the time. Doesn't even get the swear words right. You know, so I still think we have a long way to go. I think the, the, the sort of the, I would call the laboratory results are very promising. Right. Those laboratory results require like 10,000 computers. Right. Yeah. It's unfeasible right now to have that. Yeah. That's right. That's right. So one of the questions will be just kind of how long does it take for that kind of computing power to, for the price to drop and become more ubiquitous.
And, you know, there's Moore's Law, there's all sorts of questions around Moore's Law. Some people think Moore's Law is slowing down. I think one of the big potential interesting things here is what's called quantum computing, which is this whole new kind of theoretical area of basically how to build computers that use quantum effects, so things from quantum mechanics. I would say the optimistic people, including some very well-respected computer science professors at Stanford, for example, believe that in five to ten years we'll have quantum computing in the mainstream. if that happens you could see a uh a dramatic ai will take off it could it could lead to a dramatic acceleration in the performance of ai yeah that there's a bunch of things one of the reasons it's very hard to predict these technology things is that you often have these things that have kind of feedback loops right which means like if we get quantum computing and if we you know you know and if we get you know that will let us compute things faster which will let us store more and then we'll be able to store more data.
And that'll have all sorts of second and third order effects on everything that's possible. Yes, exactly. It seems to be very complex kind of feedback loop systems. And so if I had to bet, I think we're still pretty far away from kind of the singularity. But there are certain scenarios Dupil can conceive of that it's much sooner. So with all these different companies, how do you filter information? And how do you do that personally? How do you know what's important and what's not important? And like, how do you determine signal from noise when you're, you know, surfing the internet? That's a great question.
I'm a huge fan of, you know, Twitter, for example. I use Twitter constantly. And for me, it's probably one of my most important work tools in that I've, you know, have a carefully curated list of people I follow who essentially I have whatever it is, 2,000 of the smartest people in the world finding information for me and telling me what to read. That's how I view Twitter. So that's obviously very important. And, you know, at the firm, we have a whole bunch of different things we do, including, you know, lots of people that we interact with and that we talk to regularly. We try to do things like, for example, we have a big academic conference coming up in a few weeks where we invite 50 plus of the top computer scientists in the world to come and kind of do like a mini, almost like TED Talks or something at our firm.
We, you know, we do lots of outreach with academia and things. We try to get involved also in like the open source communities, you know, go to lots of events, do lots of press outreach. You know, a lot of what we try and do is just kind of be in the flow of a lot of different, you know, interesting groups of people working on new things. Right. So what do you think people are focused on that's a waste of time? Like, what do you think misplaced attention? Where would that lie? Good question. So in the tech world specifically? or yeah or in general i mean like other than donald trump um i think that uh um i'll give you well i guess i'll give you one example of the food industry um right so i'm an investor in a company called soylent which you may have heard of which is kind of yeah i've tried their stuff okay so so so i think with soylent i mean the idea with with soylent is that we um we're trying to create kind of what we would consider scientifically perfect food um kind of the idea is you go and And the guys, you know, have a team of scientists who went and read every scientific paper about nutrition and then designed and built this perfect food.
I think when you look at the food industry, there's an interesting movement. Soylent's one example. There's a bunch of other Silicon Valley startups that are trying to do new things around food. I think that's an industry which is just a very backwards industry today. If you look at the U.S. right now, the diabetes and obesity are really at the epidemic level. And a lot of that's caused by excessive sugar and other kinds of ingredients like that. And if you just go, you know, just down just now this morning trying to find something healthy to eat at the local store and, you know, everything is filled with junk and sugar and all sorts of other things.
And it's really just an industry built around advertising and marketing and distribution and almost no money is put into actually researching healthier and better foods. So that's something I'm very passionate about. I think we like to say sugar is the new smoking so when you look back 20 years from now people will just be stunned by the kind of foods that we ate today I think the whole organic food movement is a great thing I think that's mostly only accessible to wealthier people so I think a lot of what I think is interesting are the people that are trying to think more broadly about how to reform the industry not just for people that can afford organic food A lot of my friends would call that the luxury of the rich, right?
Yeah, exactly. So just what can we do more broadly? So that's an interesting one. I'm very interested also in healthcare generally. We're doing a lot more, spending a lot more time making investments in sort of areas that intersect between healthcare and computer science. And just think there's a lot of things there that, you know, if you just look at the statistics of, you know, why are healthcare costs going up so dramatically? A lot of it has to do with the inefficiencies in the system. Everything from medical records are still kept on paper. The insurance system is very complex and in many ways backwards.
The incentives seem all over the place. The incentives are all over the place. It costs more and more now. People debate the exact reason, but basically it costs more and more to create new drugs. right so there's all there's all sorts of interesting things there i think they can be improved listen i'm totally conscious of your time here where i'm nearing the end i have three questions that i always ask everybody um so what's the one book you've read that had the greatest influence on your life oh man i think uh you know i think when i was in high school i read um uh girdle escher bach you know double soft setter's book um i don't know if you know that book it's no i'm gonna look it up now though so i was interested in computers since i was a kid and this book was sort of it tied together computers and philosophy and music and for me it was really important because it really broadened my horizons and I ended up majoring I went to when I was in college I majored in philosophy and I think that book kind of got me to do that that would be a huge one for me also anything by Daniel Dennett, do you read Daniel Dennett?
yeah yeah consciousness explaining there was this whole kind of a thread of, I think, you know, Oliver Sacks, who sadly just passed away and schooled. I used to just read all of those kind of popular. Why are you so smart? I don't know. Those are great. Those are great teachers. So when I was a kid, you know, whatever, high school, college, I read all those books and I, and those were all just sort of hugely influential on me. So what's on your nightstand right now? What are you reading right now that you're really into? I just read, what's called the three body problem just finished it you've read this book it's um yeah it's this uh it's this chinese author who just wrote it's a science fiction book um it just won the hugo award or something it's really it's really interesting book um what else did i read i read um i'm reading i was i just bought this uh was it called the martian which is i guess it's a popular book that's now made into a movie and uh i you know i books are one thing i like to i i i'm I'm not uh digital on books I buy only oh you're still physical why is that I just really like physical books I like having them on my shelf I like the feeling of reading a book um I don't know I just it's it's something where I'm the same way I live in this like weird world where you know I read physical books and then go out and buy a kindle copy just because I can't keep every physical book right like it's that's true that's true that's a problem I just like the feeling of them And it's also, I feel like I look at the screen too much as it is.
Yeah. And it's distracting too to have like the thing I, you know, I don't know, I read on the, on the iPhone and then I can like check Twitter and it's like, it's too distracting. Oh, I just read the Elon Musk book, which I thought was good. Sort of the biography of Elon Musk. Right. And, oh, I read a really good book called Sapiens. Have you heard of this? Oh, I heard of that book. Yeah. Somebody else recommended that to me. It's sort of like, it's kind of like Gun, Germs and Steel. Like one of these, I think they're calling the genre big history where it's kind of the panoramic view of history.
and it's just the history of homo sapiens. I've heard that was amazing. It was really awesome. I highly recommend it. And so you know what I'm trying to do with the Knowledge Project? Who would you like to see on the show? Out of anybody? Yeah, out of anybody in the world. Can you narrow it down a little? No, who would you like to hear me interview, I guess? That's a good question. I'm going to call them and tell them you recommend it. Okay, that's all right. It would be good. I know who'd be great is do you know Venkat from Ribbon Farm? I do, yes, yes, yes He's one of my favorite writers He's already agreed to come on Oh wow, okay, well there you go I have too much overlap That's awesome, no, that's a good choice You know who's incredibly brilliant is Ben Thompson He has his thing, it's called Stratechery Yeah, definitely I just started following him on Twitter Yeah, he's awesome Well thanks so much Chris, this has been great fun I really appreciate you taking the time.
Yeah, well, my pleasure. Thanks for having me. meat filled with all sorts of brain food? Go to farnhamstreetblog.com slash newsletter. This is all the good stuff I've found on the web that week that I've read and shared with close friends, books I'm reading, and so much more. Thank you for listening. you