Introduction

Patrick: [00:00:57] My guest this week is Chris Dixon, who has written some of my favorite essays on technology and venture investing. Chris is a prolific investor and thinker, having been an entrepreneur, angel investor, and now partner at the well-known venture capital firm Andreessen Horowitz. Our conversation focuses on major trends in technology including cryptocurrencies and the future of autonomous vehicles and drones. Chris has a rule of thumb for technology trends. "Find out what smart people are working on during the weekends and you'll know what others will be doing years down the road." After surveying his old essays, it's clear you could use Chris' writings as a similar litmus test. Please enjoy this wide ranging and great conversation with Chris Dixon on the future of tech.

Studying Philosophy

Patrick: [00:02:05] So Chris, it's not often that I come across a fellow philosophy major on this podcast, so I have to start there. We'll go all the way back to college and ask, why that was what you decided to study?

Chris: [00:02:16] Great question. I had been interested in computers my whole life, like as a kid growing up and programmed computers. And then, I read a series of books, Douglas Hofstadter, Daniel Dennett, these kind of... I don't know if... people that know will know and people that don't know may not know. But, sort of AI, cognitive science, philosophy of mind, logic, and I just thought it was like kind of... to me, it was this really interested intersection of all of these technical things I was interested in, but then these almost science fiction-y futures of artificial intelligence. And then, you naturally, as you follow that path you're like, "Well, what is artificial intelligence? What does it mean to have consciousness? What does it mean..." So, that's the rabbit hole I started down.

And then, I ended up focusing on philosophy of mind, logic, philosophy of language, kind of very analytic philosophy. Philosophy is kind of divided between continental, as you know probably, but continental and analytic philosophy, and so I was very much on that side. I had a wonderful experience. It was a great thing to study. It turned out that the job market in technology was probably a more attractive job market than the philosophy job market, so I ended up switching at some point. But, I still go back and read it. I actually recently wrote an article for The Atlantic, which was really--

Patrick: [00:03:22] Oh really?

Chris: [00:03:23] Yeah. Which was on the history of logic and computers. And so, I still read a lot of this stuff. I'm interested in it, but I realized at some point I'm never going to be anything more than a mediocre philosopher, so I'll try my hand at something else.

Patrick: [00:03:36] How much has your, or maybe where has your view changed the most since that early reading to now? Having done a lot in the real world and business and investing?

Chris: [00:03:45] In terms of philosophy interest?

Patrick: [00:03:46] Yeah.

Chris: [00:03:47] That's an interesting question. It's interesting because I go back and I try to catch up on what's going on in philosophy and it turns out it doesn't change that much. It's a relatively slow moving field. But, that's a great question. I think the stuff I did was... I don't know, I just think as you get older or something you get more interested in other questions. So, analytic philosophy is very... The stuff I did at least, was very narrowly focused on things like logic and philosophy of language. It was kind of technical and I think just as I've gone on, I've gotten more interested in questions of really going back to the Greeks. If you go back and read Plato and Socrates, the questions they asked, "What does it mean to lead a good life? What is it to be a good person? What is the nature of reality?" I don't know, these kinds of more almost like--

Patrick: [00:04:22] Big, big--

Chris: [00:04:22] --religious, metaphysical questions. To me just personally... I don't know, maybe just as time goes on you start thinking about these things or something. And so, if I were to do it again I'd probably have maybe a broader view and think about more fundamental questions of living and life or something, as opposed to something so technical. But, at the time that seemed really interesting, the technical stuff.

Patrick: [00:04:42] How about tying that into technology? This is something I think a lot about. Working with a lot of data on my side, I know that the things that you pick to optimize for determine a lot about what gets built. I'm curious, if you think philosophically about technology, what it is that all the things you write about, think about, invest in are optimizing for? Is it human growth, is it happiness, is it wellbeing, is it laziness? How do you think about what is effectively... Back to talking about the Greeks, like Aristotle's prime mover? What is pulling us and driving the technology that we create?

Chris: [00:05:17] I think it's... The actual motives are very different, right? I think there are a certain set of people who are sort of engineers, who have engineering mindset, who get a great kind of delight out of the creative act of building something. I used to be like this when I... I know that feeling. It's this sort of feeling of deep satisfaction that you've had an idea, went and worked on it, and then saw something that was actually built. I think if you talk to most engineers, like computer programmers, so much of what they do is about seeing something being used in the world and going from an idea to something that actually affects people, right?

So, I think there's the engineer's perspective, which I think a lot of people in Silicon Valley were once engineers and see that as... I see what I do as sort of an extension of that, which is we help to allocate capital to people on those projects and in some ways help decide what projects make sense and connect to the broader world. If you look at it from a societal point of view... I mean, this is very controversial these days and you just showed me a New Yorker cover picture which was talking about the coming automation and the various threats. There are all these discussions going on in the news today, in the news and Twitter and Facebook, et cetera about are autonomous cars going to take away jobs and AI is going to displace people, and things like this. Which, I think there's a real important conversation to be had there and I think sometimes it's clear that technology does have that impact.

I think if you... My own view is that if you look at the history of technology that it has caused some short term displacement, but over time has created more jobs and has improved the overall quality of life. And so, if you look at... There's a great website. I think it's The World of Numbers, I believe it's called, where they go through... You can look at any metrics, like GDP, infant mortality, just overall percentage of the world population and poverty. Just objectively, everything has gotten much, much better in the last 200 years. I just fundamentally believe that will be the case and will continue to be the case.

I forgot which economist said, "Human needs and wants are unbounded." The idea that this, as they call it, the lump of labor fallacy. That once you automate a few jobs that there won't be new jobs created. I don't believe that's true. I think there's a fundamental assymetry and it's very easy to imagine... With new technologies it's very easy to imagine the jobs that will go away. It's very hard to imagine the new jobs created because by definition they don't exist yet. So, 20 years ago you would've said, "Okay, Photoshop is going to ruin print graphic design," but you didn't imagine it would create web developers and mobile app designers, and all these other things. In fact, there's a great chart somewhere where they show that actually the number of total net jobs created went up in those design fields. And so, who would've imagined today an Uber driver and a mobile app designer and social media manager. All these kinds of jobs that actually do exist would've been very hard to predict. However, you could've predicted that people that print newspapers and all sorts of other things would go away with the internet.

To answer your question, I think there's a local motivation, which is to build things. The core human need to create things, which I believe is a core need and a core drive that brings people happiness. And then, I think there's a broader societal goal, which I think is to just increase well being and welfare. I do believe that technology does that in general.

Hardware vs. Software

Patrick: [00:08:32] One of the things, the common themes of all my conversations recently has been the power of compound knowledge over time. A lot of it driven by technology and a lot of the discovery that's been a result of networks, both open and closed. So, I'd love to spend a decent amount of time on network design, on features of networks, both open and proprietary. Maybe you could begin by your own history with exploring networks. I know it's become a key part of your thinking from an investing standpoint and maybe a technology standpoint too, so a clean way to do it would just be to describe how you got interested in networks, network design, and we'll talk about where that's all going.

Chris: [00:09:09] Big topic. I have this joke, it's not a very funny joke, but it's a joke that computer science was mis-titled and it should've been called algorithms and network design.

Because, if you really look at what people do in computer science, it's generally those two things. As opposed to, you can actually do on a computer... Most of what computer scientists do actually don't require computers. When you do algorithms and network design, it's really much more of an abstract mathematical thing and it's, as they say, substrate neutral.

You don't actually need to embody it on silicon to have it be this... And in that sense, there's a true, I think, theoretical academic subject or something. But, it's that important, right? It's like half of computer science, what they call distributed systems or network design. But, it's so important. And obviously, the internet is the base layer of the... I think of the internet as... I know we're going to talk about cryptocurrencies in a minute, but I think of the internet, what we have today, as kind of the first couple of layers of what the internet could be. I think it's only at its very early stage of its--

Patrick: [00:10:01] It's 25 years old.

Chris: [00:10:03] --of it's potential. Yeah, it's 25 years old. Well, and there are two ways to look at that. There's the Carlota Perez framework, if you're familiar with that, which is--

Patrick: [00:10:09] Deployment age.

Chris: [00:10:10] You have these technologies like... One way to look at it is, it's similar to the automobile. Which is, yeah the automobile gets invented in 1900-ish, you have a bubble, as she describes, in that case it was in Detroit, and you had whatever it was, thousands of car companies, and then there was a crash. And out of that Ford and GM, and a few others emerged. And then you had this long period of what she calls the deployment phase, right? Which is, you have cars and they get better, but they're still cars fundamentally. They don't change that much, but what does change is all the applications for the cars, right?

So, you build the highway system and trucking, and fast food restaurants, and big box retail in suburbs, and all these things that couldn't have existed before you had cars. And so, those are the apps, if you will, of the car era, right?

Patrick: [00:10:52] That's very interesting, yeah.

Chris: [00:10:54] But, that's kind of how it plays out. And then she goes through... It's a fascinating book and I highly recommend it if people haven't read it. It's the history of financial cycles. I forget exactly.

Patrick: [00:11:00] I spent a long time talking to Jerry Newman about this on the podcast.

Chris: [00:11:02] Yeah, he's written some really good blog posts about it. And so, her view is you have this... And actually, the Gartner Hype Cycle, if you're familiar with that, that actually should be called the Perez Cycle. It came from her work.

Patrick: [00:11:11] this stuff plotted on that--

Chris: [00:11:17] Yeah, so it's sort of this initial period where everyone gets really excited and probably maybe over excited, then there's a crash. It's very much like a dot com thing. And then there's this long rebuilding. Okay, now I think there's a really interesting question as to whether that cycle applies to software the same way it does to hardware.

With cars, it's fundamental. Once you invent the combustible engine... Yeah, you can do an electric car, you can do ride sharing and things like this. It's 100 years later and we're now working on these things for real, right? But, I think of software as fundamentally different than hardware in that software is literally the... Going back to philosophy a little bit here, but it's like the encoding of human ideas. If you can think of it, you can write software that embodies it. And so, I think of the internet as... I think this will touch on some important issues. I think of it as a much more plastic malleable form of media than television, radio kind of thing. So, I mean, fundamentally if you look at television it was a hardware-based network, right? I mean, it was like you had certain RF signals and things like that. There was only so much you could do. You could send pictures on it and that's it. And then, at that point it was just like different content and maybe upgrade from black and white to color, but that's about it.

Here, on the internet, you can literally through software... The way the internet is architected is you have these relatively simple base protocols and on top of that you can re-architect almost anything. It's completely malleable and can reinvent itself. I think this is very important because if you're familiar with Tim Wu, who has a book--

Patrick: [00:12:23] Yeah.

Chris: [00:12:24] It's a great book, very interesting thinker. I'm a big fan. But, it fundamentally says what's happened is, with the big four internet companies today, we're seeing the same kind of centralization that we saw with--

Patrick: [00:12:35] Telecom, yeah.

Chris: [00:12:36] --TV and radio, and everything else. And, I think his conclusion is the only real answer to that is regulation. If you believe that the internet and software is just like hardware, then that's the inevitable conclusion. I actually believe something else, which is I think software can change and is much more malleable, and that there are potentially technological and market forces that can correct the centralization.

Patrick: [00:12:57] What are the features that make that true? So, I guess maybe the one I would think of first is just physical limitations for hardware. There's a limit to how many transistors you can jam on a certain amount of space and software maybe can iterate or go through stages of evolution generations faster. Is that the primary driver of why you think that's true? What are the reasons behind that?

Chris: [00:13:16] Why software is different than hardware?

Patrick: [00:13:17] Yeah.

Chris: [00:13:18] Yeah, you know it's-

Patrick: [00:13:20] Or why those other models might... The Perez and-

Chris: [00:13:23] It's the degrees of freedom. We know software, to be turing-complete computer is to be a computer that can embody anything, any logical... It's mathematically provable, let's go back to philosophy and logic, that any algorithm that can ever be invented can be embodied on any turing-complete computer, right? So, it comes down to the degrees of freedom essentially, right?

With hardware, you're fundamentally limited by the physics and by everything else, whereas software we know mathematically the degrees of freedom are literally the limits of human thought. There are some debates. If you read books where they talk about Gödel's incompleteness theorem and whether that means. But, for the most part we know that the degrees of freedom are almost unlimited. Which his, by the way, why it effects the business cycle too. Which is why software companies are commoditized much more slowly than hardware companies. I think it was Larry Page who said he expected all hardware to eventually be the cost of the commodity inputs, the silicon and everything else. It eventually just goes down, because in the end once you've built... There are only so many ways to put together a semiconductor or something, right? Whereas, there's just like infinite ways to put together a set of cryptocurrencies. There's just simply no limit. It's literally like you can have a new idea and you can invent a new way to do it. It's an unbounded design space.

The Internet and Open Networks

Patrick: [00:14:28] You mentioned that the internet is 25 years old, maybe we're just in the early stages of seeing... Obviously, we've seen a big suite of apps being built on top of that and capture an enormous amount of value. Maybe you have this over concentration of power or centralization on top of a very decentralized open network. So, maybe talk about what you view as the best features of both proprietary, maybe proprietary centralized networks, if there are positive features, and open networks. That will be an interesting way to get into crypto.

Chris: [00:14:56] Talking about centralized networks, so by that... Or centralized platforms, let's say, which is the obvious big ones are Google, Apple, Facebook, Amazon. The benefits are immense, which is like I have this magical super computer in my pocket that has access to all human information and I pay for the phone and everything else is free. I mean, it's unbelievable, right? Someone Tweeted, "You have more information than the President of the United States did 15 years ago." I mean, it's unbelievable. Three billion people have smartphones and it's a miracle. By the way, you go back and you read science fiction. An interesting fact of science fiction, no one predicted the internet. They predicted flying cars and robots and all sorts of like Tesla coils and things that fire laser, and everything else, but no one predicted you'd have a super computer in your pocket with access to all the information. And, by the way, I can instantly... My partner, Marc Andreessen, always jokes, "Of course people are looking at their phone. You've got three billion people and this filter of the smartest, most interesting things in the world. Why would you look at the person next to you when you can look at like this..."

Patrick: [00:15:44] The best of the best.

Chris: [00:15:46] The best of the best, right. I mean, like it is amazing. He's sort of kidding. We actually do believe you should go outside and interact with humans, so it's said a little bit tongue-in-cheek, but it is an unbelievable thing. And look, Facebook is free, Google... I mean, Google is a miracle. It's a true miracle. I think if you described it in a science fiction novel 30 years ago it would've just sounded like the most preposterous thing in the world. That you can just literally type any word or anything you think of and auto complete and think for yourself, and you'll have this magical crowd sourced encyclopedia. I mean, it sounds ridiculous, but it's what we live every day and everyone just expects it and thinks it's normal.

So, I mean, look... I mean, yeah, and they put ads on there, but it's actually the ads are pretty good and they aren't that annoying, and they don't have big pop-ups and things. It's amazing, right? It's amazing.

Patrick: [00:16:29] An interesting distinction would be between Wikipedia and Google. Obviously, Google is far more ubiquitous and people use Google constantly. I probably use Wikipedia as much or more than I use Google, and that is a very decentralized open network-based thing. I'm trying to really understand. I've really struggled with this, honestly. Like what the litmus tests may be for whether or not something is better off... As we see this proliferation of token-based open networks, whether or not something is better centralized or decentralized. Do you have thoughts on the criteria?

Chris: [00:17:02] Sure. I mean, I'll just briefly touch on Wikipedia because Wikipedia is kind of a miracle. Thankfully, the creators of Wikipedia, Jimmy Wales, they decided to make it a nonprofit. It does have certain technical features that lend itself... They do have to every year, and you'll see I think the banner up today, they're asking for money every year. But, the hosting costs are relatively light. It's hosting cashable static webpages. It's very unlikely you could run the operating expenses of a search engine or something in that model. Some of it happens to be just the operating model is relatively inexpensive, it's a nonprofit, it's worked like that. But, it's also just frankly, I think, an accident. A wonderful accident in history, that the founders decided to make this thing a nonprofit. And so, that's great. I don't think...

If we look going forward, the reality is that the vast majority of investment today on the internet is made by the very large incumbent companies and they're trying to build products that reinforce their platform dominance. And so, you'll see it happening where now they're all kind of doing everything, right? They're all building Google/Alexa style, Google Home, like things you talk to in your home. They're building... They all have their--

Patrick: [00:18:04] Kind of testing each other's boundaries.

Chris: [00:18:04] Yeah. They all have their Apple TV, they're all doing kind of phones, they're all doing... Right? So, sort of everyone is doing everything. Everyone has got a voice assistant. Amazon is now moving in... They all have AWS, Google Cloud, Microsoft Azure. They're all now... Eventually, they're all going to have business apps. They all want to kind of do everything, as you would expect a for-profit business to do, right? They're profitable, they reinvest the profits in new things that expand their dominion. And so, you just asked me... The benefits of centralization. There are lots of technical benefits of centralization. Performance is a huge one. Just from a design point of view, if you want to build something like a search engine, there are all sorts of benefits to having all of that web traffic flow through some central point. Now, interestingly, behind the scenes at Google it's actually a very decentralized system and they've actually famously gotten very good at... Bill said if you look behind the scenes at like Google's data centers and things, they are commodity Linux boxes running these very distributed algorithms. So, it actually is very decentralized behind the scenes.

There are all sorts of things you can do performance-wise. When we talk more about cryptocurrencies we can talk about that. That's one of the big challenges in the decentralized world. Decentralized systems, and so the internet was one. And, by the way, a very interesting case study is the internet versus AOL. I think people today forget that AOL was a very powerful incumbent at the time and the internet was the scrappy underdog. The benefits of decentralization. Sometimes people talk about this when we talk about Bitcoin and Ethereum and systems like this, they talk about things like censorship resistance, which is one of them. Which is the ability to withstand a government attempt to shut them down.

I don't think that's a particularly interesting feature of decentralized systems. I think a very important feature of decentralized systems is the ability for users and developers, and people participating on that platform to feel like they can build on that platform and not have the rules of the game changed later on. To me, that is the key feature of decentralized systems. I think to understand that, you have to look at why did the internet beat AOL? The internet beat AOL... The simple way I think about it, is there are a few million, at the time, now probably 20 million great software developers in the world. As Bill Joy famously said, "No matter how many smart people you have working for you, the smartest people don't work for you." The smart people work for somebody else. And so, regardless of if your Facebook or Google, I don't know how many engineers they have, it's in the 10s of thousands, right? I mean, 10s of thousands out of 20 million, right?

Patrick: [00:20:17] Yeah.

Chris: [00:20:17] So, they have a tiny drop in the bucket of the great engineers. In my view of the world, the systems that will ultimately win and the reason the internet beat AOL was that they captured the hearts and the minds of that massive developer community. And why did they capture their hearts and minds? Because, you knew if you were Larry Page, when they started in 1996, I think, '97... You knew if you build a search engine on the web that the web wasn't going to go say, "Oh, thank you for doing that. I'm going to now change the rules and charge you money." You know, like Apple charges you 30% and Google charges you 30% on their phones. Or famously, like Facebook and Twitter had a bunch of built apps like Zynga and things on top of them and they changed the rules on them. They knew they could build that thing and if it worked, it worked. Why? Because, the internet was a set of decentralized protocols.

There were some committees and things that decided, but they nonprofits whose charter was to encourage innovation, not to take the profits away from companies.

And so, it was like the highway system. It was like a public infrastructure that you knew you could build on and if you build this awesome trucking company on the highway system, the highway system wasn't going to come along and take it from you, right? Would you want to make that kind of investment on a private highway system? You probably wouldn't right? And, in fact, the history of technology has shown that if you did it's a very dangerous game to play. Investors have learned that lesson and don't invest in things like that.

And so, ultimately... And then, Mark Zuckerberg started Facebook in a Harvard dorm room, again on the open web. I worry that today it's not clear to me that you could be a Sergey and Larry in a Menlo Park garage or a Mark Zuckerberg in a dorm room and start an internet company and actually have a credible shot at creating a massive new company, because of the degree to which these things have become centralized.

Just a simple example is everyone uses their phone. That's the vast majority of computer usage today. And, most people find apps by going to the app store, and those app stores are controlled by Google and Apple, right? And so--

Patrick: [00:22:06] A very small bottleneck.

Chris: [00:22:06] Yeah. And, in fact, we haven't seen... I think if you look at the top 10, even 20 apps, on the iOS app store at least, I don't think a single one was created post 2012?

Patrick: [00:22:15] Yep. Talking to Dave Tisch in New York about this and he asked the question. He was like, "When was the last time you downloaded an app that like you actually use with any frequency?" And it's been a long time.

Chris: [00:22:23] Probably games. I mean, that's the one thing, but it's like a short... You play for a couple weeks or whatever, but that's it right?

Patrick: [00:22:28] Yeah.

Chris: [00:22:28] That's right. The kind of home screen everyday stuff hasn't changed, and you can just look at the app store. And so, that worries me. One, I guess is a selfish reason. We invest in startups, so you could say I'm just talking to my book here. But, I think you could make a strong case that this is an issue for society because you want people to be able to start something in a garage that becomes a big deal. Because, that's how real innovation happens.

Patrick: [00:22:50] Let's move to some of these same issues through the lens of some of the major cryptocurrencies. One of the things that in my investigation I found so appealing about this whole space philosophically, is that they are these open networks. But, one of the things I'm struggling most with is the tendency for concentration in this case. So, with TCP/IP there's no value within the protocol layer that could be accrued to whomever. Whereas, with Bitcoin you see... If you look at like a Gini coefficient, the concentration of ownership, the concentration of Hash Power, hierarchies just tend to emerge in these things when there's value or wealth associated with it. So, is that a concern for you? I know, obviously, the potential upsides of having these be more open networks for exchanging value versus information, but what do you think about that idea? Just the fact that they do hold value, kind of the fat protocol idea could just lead to all the same typical decentralization problems?

Chris: [00:23:45] I mean look, as your listeners may know, there's been a big debate going on, on Bitcoin for a long time over the what's called the block size debate.

Patrick: [00:23:50] Could you just highlight that a little bit for us?

Chris: [00:23:53] Yeah. So, basically... Let me just maybe, if it's okay, step back a little.

Patrick: [00:23:58] Absolutely.

Cryptocurrency

Chris: [00:23:58] Cryptocurrency really started in 2009 with a publication of the Bitcoin paper by a pseudonymous author, Satoshi Nakamoto. Then that sort of kicked off. At first it was people kind of playing around with it and experimenting, but then the price started going up and miners came in. Miners are basically people that provide the service side. They're kind of the AWS, they keep the ledger. They're sort of the service providers of the network. And then some users came on as people started building wallets and some companies got started. And then, it kind of really rose in public consciousness in 2013 when the price started going up and then it kind of crashed.

There was this debate going on, which was among the core developers of the Bitcoin protocol over what's called the block size, which essentially is how much memory to allocate to each block, which basically reduces to how powerful of a computer do you have to have, to be a Bitcoin node? And so, one school of thought is this should be something that's highly decentralized and that anyone with a low powered laptop should be able to participate in the network and be a node, and be part of the governance of the network. The other school of thought is no, we need to increase this over time so that the system can scale and we can lower transaction costs. And there's a whole bunch of... It became this very contentious thing and it still is very contentious, but what happened in the meantime is a certain set of developers said they want to try a different path and they created other protocols that were sort of inspired by Bitcoin. They took a lot of the ideas of it, but changed it and expanded on it. The most notable of those is Ethereum.

This was Vitalik Buterin and a group of, I think there's like 20 core Ethereum developers who created something like Bitcoin. It's a decentralized system that uses a blockchain. A blockchain is the security mechanism. It's what they call a consensus mechanism. It's where a whole bunch of different network participants called miners come together every X minutes, depending on the system, to vote on the state of the network, the ledger-

Patrick: [00:25:46] Who owns what, basically.

Chris: [00:25:47] Yeah. And so, Ethereum is kind of a generalization of Bitcoin, which goes to say that instead of just keeping the state of a ledger of currency, we're going to add a whole computing platform on it. It has a language called Solidity, which is very similar to JavaScript. It's a very powerful general purpose computing language. And so, on Ethereum in addition to transferring money you can also write arbitrary computer code that they call smart contracts, that can do all sorts of anything you can think of. One of the examples they had on their home page when they launched, of like the 10 cool things you can do, was crowdfunding. Fast forward to today, you may hear about ICOs. That's literally that code almost, slightly changed, that they showed on their homepage...so, it's a crowdfunding thing, which turned into... People sort of said, "Hey, let's do this crowd. So, you send us money and we'll send you back this digital good called a token, and this will be a new way to raise money and develop protocols." Then that kind of kicked off a whole other wave, and now we have... It's really exciting. There are thousands now of these projects and there are, unfortunately, as with anything involving large amounts of money, there are a lot of bad projects, there are some outright scams, but there's a lot of good stuff too. I probably see a few interesting... Per week, I would say these days I see a few very, very interesting and talented teams working on new protocol ideas inspired by these ideas of Bitcoin and Ethereum.

Patrick: [00:27:08] Do you think that is the answer to that prior concern of stagnation of the big tech players all having secured their network effects? And potentially, this is a way both through fundraising, but also through if the internet gave birth to everything we see, that this is a new layer of protocols? You mentioned that we might just be at the beginning, maybe this is just step two of 10 of what the internet could become.

Chris: [00:27:31] I think it's very early and so... I want to caveat a couple things. One is, nothing is say here is--

Patrick: [00:27:38] No, of course not. We're disclaimed very well on this podcast.

Chris: [00:27:41] Okay. Nothing here is investment advice. I do not think anyone should go out and buy cryptocurrency unless you've spent a ton of time researching it and everything else. Every prediction I make will be over a five to 10 year horizon, not a one to two year horizon. Because, I've made predictions in the past and then on Twitter I'll get--

Patrick: [00:27:55] --It'll come back to bite you.

Chris: [00:27:56] --a stock went down today or you know. So anyways, so this is all a very long term horizon and I'm not making any... I think you should only put a tiny portion of it... If you do any investments in these things it should be a tiny portion of your savings. So, okay. I also think, by the way, that it's very likely that the cryptocurrency prices have gotten indeed ahead of the actual progress and there very well could be some kind of correction in the next 12...I wouldn't be surprised at all, because there are a lot of limitations. That said, I think it's an incredibly exciting and powerful technology. I think we are in what I would describe as the infrastructure phase. It's actually very similar to how the web was developed. In the 1990s, if you wanted to build a website you had to go and literally go and get a data center and put servers up, and everything else. Then things like AWS and all these kinds of things happened where two people could just toss some code on a server and you could compete with bigger companies and things. And so, we're still building infrastructure.

Ethereum supports like 10 transactions a second. You asked earlier about the disadvantage of decentralized platforms, that's a big disadvantage. The performance is not as good. Now, there's a million... There's not a million. There's five to 10 interesting paths by which that performance will get improved and there are very specific proposals and all sorts of projects being worked on, including by the core Ethereum developers and things like this, so I'm very optimistic it will be. But, that's just the state of it today, right? The state of it today is even if somebody came up with something that was like a social network using cryptocurrency, which people have come up with and I think will come up with, it just simply won't scale from a technical point of view to the scale you'd need to really challenge the Facebooks and Googles of the world.

Now, that said, there's a ton of interesting stuff happening. There's a lot of money flowing in, there are a lot of smart people. It's hard for me to keep up on the number of interesting new projects. There's just so much. The speed of development is so rapid, which is... The time it reminds me of the most, I was an entrepreneur then, was sort of 2003-2005, which was what people called the Web 2.0 era. Kind of 2003 to '09 or something, which is when you had the early social networks. You had Friendster, you had Flicker, and all these other really cool things. You just had this feeling that... Every day there was a new project and how could you keep up? I remember YouTube coming out and then I remember like at one point it was like... Whenever TechCrunch launched, I want to say like 2004, because I remember very vividly we were all talking about it and having meet-ups, and emailing each other, "Did you see this project? This is really cool. And like, oh this new idea of like tagging." It was all these--

Patrick: [00:30:12] Friends lists.

Chris: [00:30:13] --and following. I mean, that's... Twitter invented... The first time they did this with the asymmetric friend of following, or maybe Tumblr did that first. But, anyway, I remember the first time TechCrunch came out I was like, "Oh wow, there's finally a website that's going to aggregate all these projects." It was funny, I remember this too, is that you would go to work in the tech business and people would say, "Well that's not... That's a toy. It's social networks, let's do the real stuff," which is like enterprise software and selling boxes of security boxes. I think it was the dot-com crash and people were like, "This internet thing is an awesome thing for humanity. Let's be serious and make real businesses, not these toys." But then, all the people I knew who were entrepreneurs and engineers and things, they would do that during the day because that's where your paycheck came from and then at night you'd go and talk about all the cool stuff happening in the Web 2.0 world, right? I remember that very vividly because it was always sort of just... And, of course fast forward, the stuff you talked about on the nights and weekends turned out to be--

Patrick: [00:31:07] The real opportunity.

Chris: [00:31:08] --the real opportunity. And so, I think that's a historical pattern that does repeat itself over and over.

There's a reason why so many things in the history of computing have started in garages and in hobbyist kind of things. Because, the hobbyists on nights and weekends, those people are thinking on a five to 10 year horizon, by definition. You can think about the average really smart engineer.

You go to work and your boss is sort of thinking on a one to two year horizon and giving you your assignments, right? You have nights and weekends, and you're thinking about the really cool futuristic 10 year stuff. In some ways, you can think of it as somebody voting with their time and if really smart engineers are all voting with their time to work on VR and drones and cryptocurrency, biohacking, and all these other cool things, to me it's been a very interesting indicator of the future. Today, cryptocurrency is known. I meet engineers every day who are at companies like Google and Facebook, and all they want to talk about is cryptocurrency.

Patrick: [00:32:00] Fascinating.

Chris: [00:32:00] During the day they're working on ad targeting or something, right? I mean, you can imagine because that's what makes the money today, right?

Patrick: [00:32:04] Yep.

Chris: [00:32:04] He's like, "How do you make this ad convert slightly better?" or whatever. And then they go home and they're reading about Ethereum and stuff, right. Actually, more and more of them, I'm seeing really a change lately, are leaving those companies and starting cryptocurrency. I've seen a couple this week from Facebook and Google.

Patrick: [00:32:20] Do you think about individual, call it tokens, as analogous from your venture capital seat to making seat investments in companies? Meaning, obviously there are companies that will be built, traditional C-corps, LLCs, on top of crypto tokens. But, from purely your seat, thinking about obviously wanting to invest in something that has the potential to get very big, am I thinking about that right? That it's almost like maybe Coinbase is an exception, it's done really well, but that the better or more interesting thing from your seat is looking at Ethereum or Tezos, or something like that?

Chris: [00:32:54] I'll just tell you my own path. Bitcoin came out and we invested, for example, in Coinbase in 2013 and had looked at a bunch of other things. At that time, in that era of 2013 to '15, anyone... Not anyone, but a lot of people who were doing things around cryptocurrency would come to venture capitalists like us and raise money. And then, I remember Ethereum happening, I think it was Augur did the ICO and a few other ICO type things. They called them token sales back then. And then, I remember just like not getting any phone calls as a VC. I was like, "What's going on?" And it was because they didn't need VCs anymore. They were all raising money from the crowd, which is kind of awesome.

And so, at least my... I'll tell you my own person process. I was like, "Okay, well this is interesting." My job, you could look at a VCs job and on the one hand it's to invest in Delaware C-corps or something. On the other hand, I think our job really more broadly is to try to invest in great technologists and to do that regardless of what the actual mechanism by which you do it is. We then co-led the first investment in a crypto hedge fund called Polychain. That was about a year ago. And then, since then have started making direct investments in these protocols. What that means, is we actually buy... They have a certain... Each of these protocols will have a certain number of tokens issued at the time and maybe some scheduled for issue over time, and we will just buy some portion of them.

Markets and People

Patrick: [00:34:11] How do you think about the dichotomy between... One of the questions in venture capital that I find so interesting is betting on founders versus betting on market. You mentioned you bet on technologists, but maybe also technology. How do you think about that blend or that mix? Is there one that's more important than the other? This morning on the plane I read your writeup on the 11 exciting trends in technology. It wasn't the 11 exciting technologists, so I'm curious how you think about that split.

Chris: [00:34:38] How do you think about the trade off between markets and people?

Patrick: [00:34:40] Yeah.

Chris: [00:34:40] I think it's very stage dependent. I think the criteria by which you evaluate a venture capital investment depends a lot on the stage. Let me put it that way. And so, at the earliest stages... So, it's seed and series A, as we call it, which is generally really early. You basically have a team then and some... Yeah you have a market, but often it's like a general idea. You don't really know how... It's not a market in the way you'd think of it in a business school sense of there's existing demand and people are clamoring for... Everyone wants a lawnmower or something and they need one, and you're now satisfying that demand. Who knows if there's going to be a market for you know, when we invested in Coinbase in 2013, there's going to be a market for wallets and exchanges for cryptocurrency. So, you're speculating on the market.

It certainly matters a lot. Getting the trends right matter a lot. But, my experience has been the actual product and even in some sense the market tends to change over time, and so you just...

The only thing that really remains a constant is the people you invest in. And so, the people at the early stage investments have to be at the core of it.

I think the general direction matters a lot. And it's specifically like do you have headwinds or tailwinds of technology? When you have things like all of the innovation happening in cryptocurrency, when you have things like more... The classic thing in technology has been Moore's law, which is the fact that semiconductors get cheaper and more performant every year and basically every 18 months or whatever double. You have these kinds of tailwinds that make everything better. Social networks that have network effects. The more people that get on it, the more useful the network gets. Having those kinds of dynamics are very important as well.

At later stages there's venture capital and private investment investors who do later stage investments. A lot of these later stage investments, like if you look at what SoftBank is doing today or something, it looks a lot... The type of analysis you would expect them to be doing would be much more like what a hedge fund would be doing. You're looking at margins and growth rates, and it's very similar to what you'd see in a typical Wall Street analyst report, right? So, it varies wildly. And so, with these crypto protocols it depends on... I mean, in an ideal world we'd love to see the network launch and we'd love to see some kind of usage statistics and other kinds of things.

Patrick: [00:36:40] Right, and KPIs.

Chris: [00:36:42] Yeah. That's not the current investment climate today. Today, there are people basically with a team and they call a white paper, which is like a protocol specification, will generally come and raise money. Sometimes they have code written in things and, frankly, we have to decide yes or no if we want to invest in that. We can't just say, "Come back in 12 months when you've written it," because by that point they'll have launched and the token will be traded publicly, and it may be too late for us to invest. So, that's kind of how it works now. That may change if the market changes and the standards change.

Organizational Structure

Patrick: [00:37:13] How do you think about, again, this dichotomy of hierarchies versus more open networks or decentralized networks within firms themselves? So, within the Delaware C-corps. I'm curious, as you make early stage investments and then obviously are there to foster growth in those companies, and obviously you're resented to do so, what observations you've had as an investor over the last decade, or however long, about how organizations are organized? Because, you see interesting writeups about companies like Valve, for example, that are very flat where--

Chris: [00:37:44] The holacracy concept.

Patrick: [00:37:45] Exactly, yeah. I'm fascinated by this and I'm curious if you've learned lessons about whether or not you see things changing? Like we talked about earlier, how hierarchies just tend to emerge, but then in hierarchies very often the bottom of the pyramid is just there to serve the top and the top doesn't get real good information, and there's this stagnation.

Chris: [00:38:04] I'll tell you, if you go back and you look at predictions from 10 or 20 years ago, let's say, where people talked about how the internet is going to change organizations. There's this idea of Coase's theorem. Coase's theorem, famous economist who talked about why you have a firm and the idea of transaction costs and... I don't know, there's a whole idea. It's sort of an economic--

Patrick: [00:38:23] It's fascinating, yeah.

Chris: [00:38:23] --foundation for why there's a firm. Why are there firms at all? Why isn't everything just done in kind of a marketplace? I think you would've expected 20 years ago that the internet and having all these kinds of new networks that we have on top of the internet, social networks and marketplaces and other things, would have had a much bigger impact on organizational structure than it actually has. As an example, if you go read S1, those are the filings that tech companies make. Go read tech S1. Tech companies, when they go public they file these financial statements. A huge portion of tech company costs are very traditional sales and marketing groups. And so, a lot of these, for example enterprise software companies, it will be maybe half or more of the costs are giant sales teams. I think 20 years ago you might've thought in the year 2017 you won't need a person to get on a plane and fly across the country and convince somebody to buy some software. They'll go to the internet and they'll do some research, and they'll buy it. It hasn't really changed. The way that big corporations buy software really hasn't changed. You still need a massive commission-based sales force. That might've been a surprise 20 years ago. I think one thing you've seen though is you have seen these examples. WhatsApp was one where I believe they were 50 people when they were bought by Facebook for whatever it ended up being.

Patrick: [00:39:41] 26 billion.

Chris: [00:39:42] Yeah, 20... Depending on the Facebook price, 20 plus billion, right? 50 people, right? Because, what they were able to do is they were able to outsource so much of their technology to things like Amazon Web Services and other things, right? And just have a small team that's doing design and engineering on the app itself. I think Instagram might've been--

Patrick: [00:39:58] Something similar, right?

Chris: [00:39:59] Yeah, or even smaller. And so, now that said, you could counter argue and say, "Well, they hadn't really gotten to... They hadn't gone and built a...They were just a product company. But, I do think potentially what we're seeing in the cryptocurrency world is you're seeing... Like Ethereum, there's no company. It's a foundation. It's a nonprofit foundation that oversees it, that's like the development team. Bitcoin literally is nobody. I mean, it was this pseudonymous person that created it and now there are various groups--

Patrick: [00:40:27] Factions.

Chris: [00:40:28] Yeah. Fighting on GitHub for who has administrator access. It's like literally that's the... There's no--

Patrick: [00:40:33] It's amazing.

Chris: [00:40:34] You know, and it's this thing that's whatever the current market cap is. I haven't checked, but like--

Patrick: [00:40:38] 100 billion maybe.

Chris: [00:40:38] Yeah, something like this. And there's just no organization behind it, right?

I've always hoped that the internet would let you have more freedom to have these loose pieces. Small pieces loosely joined, as they say.

Have these small pods that are focusing on design and you're focusing on engineering and you're doing your podcast and someone else is doing this, and like we would all interact through computer networks and other kinds of market-based mechanisms or network-based mechanisms or something like this. I always thought that was a promise of the internet and it never... I feel like it under delivered on that promise and I'm kind of cautiously optimistic that maybe we're going to see that in the next phase of the internet.

Patrick: [00:41:18] What has you... It doesn't seem like you have a luddite bone in your body, but if there are things that have you concerned over the major trends that you see in technology, what are those if any?

Chris: [00:41:28] Well, we talked a little bit earlier, I think it's... The centralized four companies controlling the internet was not the internet I wanted and I don't think it's good, for a bunch of reasons. You see the tension in politics now with these debates over--

Patrick: [00:41:43] All of a sudden, like everywhere.

Chris: [00:41:44] --fake news and who controls these things. These have become such influential and important public resources that I think you're going to see more. My guess is those debates are going to become only more intense. But, I think also, just from my lens of innovation, I think it's just not good to have so much power concentrated in those companies.

Patrick: [00:42:02] Are there means other than... Obviously, we talked about the token model that's maybe a shortcut around this problem. Are there other strategies that you're either directly funding or seeing pursued by founders that are trying to change or disrupt this concentration of power? Are there things that are working or is it really difficult to break--

Chris: [00:42:19] I think a lot of it is just people aiming for different things, so you go and you verticalize. You take AI and you apply it to healthcare or something, where you just know the big four aren't going.

Patrick: [00:42:31] Like niche verticals. It seems like--

Chris: [00:42:33] Yeah. I mean, you could argue they're not niche. Like healthcare, education. So, I mean, whether niche or not, but yeah that sort of verticalized. Go and be more applied than they are. What else? There are certain areas, like selling applications to Fortune 500 companies. Microsoft does that, obviously, with Microsoft Office, but generally like... Google has Docs, Amazon has got a few things they're sort of trying, Facebook has Facebook Work, but for the most part that area is... You'll see a lot of venture capital investment in what they call enterprise SaaS. That's been relatively... Salesforce and some other big companies, but they, just frankly, aren't as powerful as the big four. And so, that area is... There are pockets like that where there is still white space and not... But, like in social, just this last week what was it, TBH, that Facebook bought? Sort of the first app to pop up in the top 10 list in a while and they just, reportedly

Patrick: [00:43:19] Snapped them up.

Chris: [00:43:19] --for 100 million, just bought them off. Which is great. I mean, it's great for the founders. I'm sure it was a good financial outcome, but this is not how we're going to get challengers to the incumbents, right?

Patrick: [00:43:29] Sure.

Chris: [00:43:30] If everyone gets snapped up. I mean, one possibility is that regulators come in. There's talk of this. I don't know if that will happen, but...

Major Trends

Patrick: [00:43:37] Using your framework of nights and weekends. Cryptocurrency is accepted because we've talked a lot about that. What are you spending your nights and weekends reading about and investigating right now? What major trend or two?

Chris: [00:43:49] There are all sorts of interesting things happening besides cryptocurrency. This is not news to anyone, but I think it's real, is artificial intelligence and stuff happening now. Artificial intelligence to the field that since, that was sort of pioneered in the 1940s really, and has had all of these various ups and downs, and they call them AI winters, and periods where people thought they were making great progress. In the end there were some great successes like Big Blue winning chess and other things, but for the most part a lot of the promises of revolutionizing the workplace and thinking machines, and things like this have kind of under delivered.

It may be different now. The stuff going on with deep neural networks, like the results. You can just look at it through these various charts. For example, Stanford has a very famous contest every year called Image Net, which is basically a contest where people submit their algorithms. The contest is to correctly classify what's in a set of images, like I think it's 10,000 or 100... I don't know the numbers. Some large collection of images, and say it's a dog, it's a cat, et cetera, and the error rates of how often it was wrong was essentially like 20-30% or something for decades. And then, since the advent of deep neural networks it has gone now down below human levels, which means it's better than humans, right? What it turned out... It's a very interesting story. I don't think it's... It's been written about a little bit, but I think not enough. It was a small set of researchers, mostly based in Canada. Geoffrey Hinton, Yoshua Bengio, Yann LeCun, who were these eccentrics in the field who still believed in neural networks in the '90s and the early 2000s. They were eccentrics because it just didn't work very well, but they believed in them because they believed that they were the artificial intelligence methods that were most similar to the human brain. And since the human brain obviously works, they thought these would eventually work. It turned out they were right, once you had enough computing power. Once Moore's law delivered processors that could compute enough and you had enough data and you had enough storage, and all the other kind of computing resources, it turned out to be shockingly right. And now Geoffrey Hinton is the head of AI Google and Yann LeCun is the head of AI Facebook, and Yoshua Bengio is the holdout who is still at University of Toronto. And it was funded by some random Canadian grants because everyone in the U.S. was so sure it wasn't going to work.

It's a great story. I'm looking forward to the definitive book on it some day. If you just graph out the results. There's a lot of hype around it, but the results are really, really promising. The big question is will it continue at this same pace? And, if it does, I think you'll see a lot of... It'll also have all these secondary effects. There's the core AI research and then there's all the things you can do with it. We're seeing all this great stuff happening in healthcare where you combine, for example, AI with the fact that you now have this explosion of data in healthcare and you can do gene sequencing and all sorts of other things with images and other things. You combine the two and you have these incredible diagnostic tools, incredible ways to do real engineering. Instead of random drug discovery of trial and error, you can now do real biological engineering. There are all sorts of really interesting implications. That's a very interesting area.

Patrick: [00:46:52] What about in the physical realm? Everything we've talked about is very ethereal. Anything--

Chris: [00:46:57] Yeah, I mean, so... Well, I mean... I was going to say... The third thing I was going to say is what I call new computing platforms. I don't know if that's physical, but in that category I put autonomous cars, drones, augmented reality, virtual reality. I think this is the other obvious mega trend. Which is just you're taking computers and you're putting them in all sorts of different places, right? So like what comes after a computer in every body? It's like 10 computers per person instead of one. They're all over the place, in different forms. I don't know, is that physical enough?

Patrick: [00:47:24] Yeah, sure.

Chris: [00:47:25] Okay.

Patrick: [00:47:25] Let's keep going down that path, because I've really never learned much about this.

Chris: [00:47:28] That's the other really interesting... By the way, all of these things I think are all intersect. AI intersects, cryptocurrency intersects. I think these are all the 3D kind of things.

In the same way that mobile amplified the internet, I think all of these three will amplify each other and I believe these are the three important things happening over the 10 years in technology. At least the kind of information technology that I work in. Specifically, it's kind of a new computing platform.

By the way, these things go through these same... We talked about it earlier, the Perez Hype Cycle. Facebook bought Oculus and VR. Now everyone is saying it's over and it will never happen. Of course it's going to happen and it's just a matter of time. If you just track out how good the headsets are, they've gotten I'd say 2-3X better year over year on a price to quality basis. They're just getting better and better and better. They're going to get smaller and cheaper, and higher quality. I just think it's going to be a no-brainer. You're a 15-year-old in 2030, you're going to play--

Patrick: [00:48:19] Playing like Player One.

Chris: [00:48:19] You're going to put on your VR headset and it's going to be this light, beautiful thing that costs 100 bucks, and you're going to go into the Metaverse and you're going to play whatever, League of Legends version 12 or whatever the things that are popular in 2030. And you're going to have this... Of course you are. Do you want to stand across staring at a rectangle or do you want to be like fully emersed in this virtual world? Of course it's going to happen, I think. Augmented reality is a version of virtual reality, which is you still see some of the real world and you overlay a virtual world on top of it. Which, I think again is going to be... I personally am a little bit more bullish on VR because I think people are just going to want to go all the way and be in there, and have the full emersion, but AR is also very exciting. You see it actually on phones now, on the new iPhone. People will experience this thing called AR Kit, which is you'll see these demos maybe where you can like put your phone up and see the virtual Ikea chair pop up and things.

This will be kind of a stepping stone drug into the full AR experience. Eventually you're going to say, "Oh, that's cool. Now I want to have it all the time." And you want a headset and you want a... But, like that AR probably will hit next year with... Androids have got their equivalent for the AR Core. That will probably happen next year and that will be exciting. I mean on the phones and then there will be headsets. AR, I've already seen demos and this is working. It's just too expensive and not quite there yet. But, I've seen demos where like you just had your drone fly over a building and now you're sitting at this table, like this conference room like we're in right now, and we all see the building projected onto the... It's just like Iron Man. Iron Man is awesome. You pick up the phone and you're moving it around. It's exactly that and it's coming. Of course you're going to do it and of course it's a better way to experience it. And so, yeah I think it's very exciting.

That's super exciting. I think the self-driving car is super exciting. There's been a lot of talk about it. I'm sure everyone has heard about it and I think it's very real. By the way, all these things, I think they're definitely all going to happen. There are a lot of questions around the timing and it could be anywhere from, in many of these cases, three years to 20 years. You go back and, VR as an example, people were thinking that might happen in the 1990s and of course it was way too early. If you go back in mobile phones, by the way, there was like a whole series of Silicon Valley startups trying to do basically the iPhone in the early '90s. Besides the Apple Newton there was General Magic, and I forget the other ones. Anyway, so there's always huge questions around timing, but this will happen. I've ridden in autonomous cars, like it's real.

Patrick: [00:50:37] How do you think about that from an investing standpoint? I always try to remind myself of Andy Rachleff's 2 x 2 matrix of you need to be right and non consensus. A lot of what you're talking about is the big topics everyone is talking about, maybe consensus that these things will happen. Can you still make money investing in these trends if everyone else agrees?

Chris: [00:50:56] Yeah, I agree. We invested in Oculus, for example, in 2013, when I think it was contrarian then, and then they got acquired by Facebook, and then I think VR became kind of non-contrarian and we've done--

Patrick: [00:51:07] We're done.

Chris: [00:51:09] Yeah, we've done a few small investments. No longer in the headsets. We instead invested in the app layer. But yeah, that did... You're right, that did move into a different quadrant. I think what happened in self driving cars, really the acquisition of Crews. GM bought Crews for I think north of... I mean, it was up to $1 billion or something. That kind of changed the venture dynamics in that space and suddenly all the valuations went up 5x, and everything else. But I think, for example, VR now people are down on it and everyone is talking about AR, so it's probably a good time to invest in VR probably, I don't know. Self driving cars, it's a little harder. The other weird thing with self driving cars is it's very different than the computing markets in that there's basically 100 large car OEMs or tier one suppliers who are significant companies. Just like yesterday, Delphi bought a company. And so, there's just a lot of money and other kinds of competition going on there. It's a great question though.

One of the beautiful things from an investment point of view about cryptocurrency is that there's just simply--

Patrick: [00:52:02] Nobody owns it.

Chris: [00:52:03] I mean, and none of the big companies have any, as far as I know, any projects on it. In terms of the big tech companies. Wall Street doesn't think it's... You hear like Jamie Dimon and these other people talk about it. They don't think it's interesting.

Patrick: [00:52:14] It's changing though.

Chris: [00:52:16] Maybe, maybe. Right now it's sort of--

Patrick: [00:52:17] The winds of change are-

Chris: [00:52:18] Right now at least it feels like a wonderful thing in that they all think it's ridiculous. Maybe that will change, but so far it still feels nice in that way. It's a great question though, because you're right, you have to be both contrarian and right. Very often we'll just end up sitting out of a market because everyone is excited and the cat is out of the bag.

Patrick: [00:52:37] Coming full circle to the funny New Yorker cartoon that we started the conversation with, which basically is a bunch of robots walking their dogs and throwing some change at the human panhandling on the ground. Self driving cars and all these big trends that are changing, I guess taking over repeatable tasks from us. How do you think about, if you're talking to a 10-year-old or your father or an uncle or something, and encouraging younger people to maybe approach the world differently, approach a career differently than they would in the past when repeatable tasks could be a great way of building a career? How do you think about that? You mentioned that these will be new jobs that we can't maybe envision, and maybe that's the only answer, but I'm curious if you push people more towards say creative work versus repeatable work and how you think about the future of work?

Chris: [00:53:24] One thing I think, for example, there's a whole bunch of jobs, which actually are growing areas, which are kind of more emotional kind of work. In healthcare, as an example, right? Where there's always going to be... I think there's never going to be... Maybe in education. I mean, maybe you'll have a mook, like a video lecture, but I think you're always going to want a TA to go and grade the papers and talk about it. At least for a long period, that's the model that seems to work, right? To have humans involved. With medical care and there's all sorts of things where you want to go and you want to understand something, you want to ask questions. I don't know, I think those jobs are relatively immune from automation.

I think if you look at the history of jobs, what's tended to happen is automation on the farm and farming, it meant farming was a different activity. You were driving around in a tractor and planting crop rotations and doing all sorts of other things, instead of physical labor. So, I tend to think these things will change as people move up the stack sort of and you do higher level jobs. And then, there are other kinds of jobs that require more of a human touch. But, I don't know. I think obviously things like having some kind of knowledge of how to do engineering and things like this, I don't know, I think that's probably always going to be highly valued. But, go back to the earlier...I mean, I don't have a great answer to this because... I think it's almost impossible... I think if you go back and you look at the history of these things, it was very difficult to predict. I mean, who 20 years ago would've said social media manager, Uber driver.

Patrick: [00:54:51] Absurd.

Chris: [00:54:51] It just would've sounded absurd, right? So, I don't know what it will look like, but there are going to be all sorts of things that I think you need to have happen.

Patrick: [00:54:58] With this UBI discussion, I always remind myself of the Milton Friedman idea that if you want more jobs, give the guys spoons instead of shovels.

Chris: [00:55:06] Yeah, yeah.

Patrick: [00:55:07] And that technology is effectively just a lever for doing more with less. We've always figured out new frontiers. Maybe that's some faith in that, but--

Chris: [00:55:17] I mean, the idea that there won't be jobs, it's like this mental model that there's a fixed set of demand and once you supply that... Like people want whatever, a TV and food and shelter and healthcare, and then once you supply those things that's it, they won't buy other things. Which just has not been the case.

Patrick: [00:55:36] It's what we started with. Like demands are boundless, right?

Chris: [00:55:39] Yeah, and for example like maybe there's going be an explosion of more creative activities. Maybe people have more free time and they watch more movies and they read more, and there's just more people writing, and there's more... I would think you have now three billion people with smartphones, you have the internet, I think it's just a matter of time before we figure out better business models for how, for example, writers get paid. I think this is something I think about a lot. I think there's a version of the future I would like to see where you have creative people rewarded, paid much better, and reach much broader audiences through the internet. I'm optimistic, but I don't know the answer. It hasn't happened. There's been no... If you go back in history, you see all these cases where people say, "Oh this is the--

Patrick: [00:56:20] It's like the Malthus argument.

Chris: [00:56:21] Yeah, yeah. And it just never happened, and I just don't see why now it's different. So, I'm optimistic about it, but... Look, also there's a competitive aspect. Is America going to stop doing AI research and then is the rest of the world going to stop? Are we going to have some like agreement with China? China is doing as much AI stuff as we are. I mean, so are we going to let them capture all the productivity benefits of self driving cars? There's another... It's sort of this coordination problem too, even if you were going to hold back on these things.

Patrick: [00:56:46] Last question that I ask every guest, because it's a fun place to close, is for the kindest thing that anyone has ever done for you.

Chris: [00:56:54] I think there would be a series of things that I would say, which is the people that have helped teach me and mentor me, whom I'm very grateful for. My parents, who of course were always... They were English professors and always reading books to me and making me read books and do things like that. And then, all the great teachers I had and people I worked with. I remember I had, just in the Philosophy Department at my university, these teachers that would spend hours after class talking about any random questions--

Patrick: [00:57:23] --rabbit hole logic stuff.

Chris: [00:57:25] Yeah, rabbit holes and sort of things like that. So, I don't know. I feel like I've been very fortunate to have all these sorts of really interesting and helpful people along the way.

Patrick: [00:57:32] Fantastic.

Chris: [00:57:34] Thanks.

Patrick: [00:57:34] This has been... Man, we have covered a lot of ground. This has been a lot of fun. I really appreciate all your time. We'll have to do a round two and check in on all of these trends in three or four years.

Created by @babyxcar