The Pomp Podcast - David Pakman, Partner at Venrock: The Age of Automation

Episode Date: January 27, 2020

David Pakman is a Partner at Venrock, where he focuses on early-stage venture investing in consuming and enterprise tech companies with a recent focus on robotics, crypto, and consumer products. Formerly he was a Product Manager at Apple and was the co-creator of Apple’s Music Group. In this conversation, David and Anthony discuss his views on audio and voice, artificial intelligence, robotics, competition on the internet, how crypto can disrupt venture capital, and why he’s so interested in the idea of crypto collectibles.  BLOCKFI-----BlockFi allows you to keep your crypto, put it up as collateral, and receive a USD loan funded directly to your bank account. They do loans ranging from $2,000 to $10,000,000, and they're perfect for helping you reach your financial goals of all sizes. Visit BlockFi.com/Pomp to learn more about putting your crypto to work without having to sell it. UNSTOPPABLE DOMAINS-----Make your crypto currency payments simple and build censorship resistant websites. Visit unstoppabledomains.com and purchase your blockchain domains today! ETORO-----This episode of Off the Chain is sponsored by eToro, the smartest crypto trading platform, and one of the largest in the world. Join 11 million other traders and create an account at etoro.com and build your crypto portfolio the smart way.

What's up, everyone? This is Anthony Pompliano. Most of you know me as Pomp. You're listening to Off The Chain, simply the best podcast in crypto. Let's kick this thing off. David Pakman is currently a partner at Venrock, where he focuses on early-stage venture investing in consumer and enterprise tech companies with a recent focus on robotics, crypto, and consumer products. He also was the co-creator of Apple's Music Group and a product manager at Apple, along with starting and selling a number of technology companies.

In this conversation, we discussed David's views on audio and voice, artificial intelligence, robotics, competition on the internet, why crypto can disrupt venture capital, and what is so interesting to David about the idea of crypto collectibles. I really enjoyed this conversation. But before we get into the episode, I want to talk about the three sponsors that made it possible. The first, BlockFi. If Bitcoin and other cryptocurrencies have any chance of ever becoming the next global reserve currency, we're going to need a lot of infrastructure and wealth management services built. This is exactly what BlockFi is doing. As many of you already know, I'm such a big fan of the company that I invested, alongside being a big user.

Today, they offer three products. The first is giving US dollar loans against your crypto as collateral. The second is an interest-bearing account for your crypto deposits. And the third is allowing you to buy or trade cryptocurrencies. currencies. These three products are important in helping people do more with their Bitcoin, Ethereum, GUSD, etc. Don't want to sell your crypto but need US dollar liquidity? BlockFi can give you a loan. Or maybe you're a long-term holder like me but want to earn some interest on your assets. Obviously, rates vary, but right now BlockFi is paying 6.2% APY on Bitcoin and 8.6%

APY on GUSD deposits. Those are unheard of interest rates in the legacy fiat world. As if the products weren't enough, BlockFi also recently told their users that they're going to be launching a credit card in 2020 that pays rewards in Bitcoin rather than cashback or loyalty points. A Bitcoin rewards credit card. Not only sounds dope, but should help more people earn Bitcoin for their everyday purchases too. BlockFi has been a longtime supporter of Off the Chain Podcast. I'm a huge fan and investor. So go to blockfi.com slash pomp and check them out. Again, go to blockfi.com slash pomp. There may even be some discounts and surprises when you head on over to blockfi.com slash pump. Now, the second sponsor is Unstoppable Domains.

They've got a really cool name, obviously, but they also do great work. Many of you have probably heard about YouTube taking down crypto-related content or Metamask getting removed from the Google Play Store. Well, the decentralized web is going to make that kind of censorship impossible. There's a lot of companies working on this, but Unstoppable Domains is one that I'm super excited about. They've created a way for anyone, even those without technical knowledge, to launch their own decentralized website. You can literally go to unstoppable domains.com and use their dot crypto domains to build a decentralized website. That website's then controlled by you and you alone. It is done the same way you control your Bitcoin with your private key. I got a website

set up on decentralized storage, and now I'm the only one who could take it down. It allows me to host this podcast, a blog, or whatever else I want on the domain without worry of censorship. If that wasn't cool enough, I can even get paid Bitcoin directly to that domain as well. That's right. I've got Pomp.Crypto, and I can use it to get paid. It's a huge improvement from sending Bitcoin or anything else to wallets using hex addresses. So right now, you can already go to Trust Wallet, Atomic Wallet, and Coinomi to type a .crypto domain into the send field and pay somebody. In order to onboard the next billion people, we can't be talking about hex addresses anymore.

Bitcoin's just too hard with those addresses. Everyone uses domain names on the internet. It's a system that anyone understands already. So head on over to unstoppabledomains.com, get your domain, and stop sending out long hex addresses every time you want someone to send you money. Again, that's unstoppabledomains.com, unstoppabledomains.com. Now, the last sponsor is eToro. eToro has been a longtime sponsor as well.

They originally started outside the United States offering stocks, commodities, traditional currencies, and cryptocurrencies to users. Then last year, they came to the US and they started off by offering cryptocurrency trading. It's not just a crypto exchange, though. They pioneered two concepts, the first being social trading, the second being copy trading. Social trading is basically the layering of social networks with trading. So you can go to any asset, let's say Bitcoin's page, and you can see the conversation around the asset. You can see people sharing information, graphs, charts, opinions, etc.

You can also trade right from that same page. So social trading is kind of a friendlier way to trade and access to more information. But let's say you also find somebody there who's been trading the asset. You can see their performance. You really like them. You think that they're smart and you wish that you could trade like they can. Well, you can now click on a copy trader button that will allow eToro to mirror your portfolio to that trader's portfolio.

They buy an asset, you buy it. They sell it, you sell it. So it's literally just a mirroring your portfolio to theirs. So head on over to eToro.com and go check out copy trader and social trading. I think you'll really like it. Again, that's eToro.com. When you head over there, you can buy, sell, and trade cryptocurrencies. You can also use their social trading and copy trading features.

Again, hit up eToro.com. Now, lastly, Off The Chain is not only a podcast. I really enjoy recording all the episodes, but I also write a daily letter to our investors. Most people would keep these letters confidential and private, but I allow anyone to subscribe to read the letters every morning. You could simply go to OffTheChain.Substack.com and sign up today. Again, that's offthechain.substack.com and sign up today. All right, now let's get into this episode with David.

It was a ton of fun to record. I think you guys will really enjoy this one. Anthony Pompliano is a partner at Morgan Creek Digital. All opinions expressed by Pomp or his guests on this podcast are solely their opinions and do not reflect the opinions of Morgan Creek Digital or Morgan Creek Capital Management. You should not treat any opinions expressed by Pomp as a specific inducement to make a particular investment or follow a particular strategy but only as an expression of his opinion this podcast is for informational purposes only all right guys bang bang i've got uh david here

i am uh i'm really excited about this because uh i think that you have probably the best understanding of other frontier technologies of anyone we brought in here uh so we're gonna talk a lot about crypto but uh we got other things to talk about as well um so thanks so much for coming to do this. Thanks for having me. It's like to be here. Um, let's go through your background. Uh, I feel like you as a venture capitalist, a lot of people know, especially younger people, but, uh, you had a whole life before, uh, investing. I did. I still feel like I'm a new investor. Um, so let's start from the beginning. You actually were a trained engineer. Yeah. Computer science engineer from Penn wrote a lot of code. And then, um, how did you go from

there to uh all of the music work that you did well i was always a musician growing up okay uh played drums still do and um was fortunate enough to be born at a time when the internet in the 90s was just becoming commercially adopted and it was somewhat obvious that print and music you know audio were going to be the first two media types to be affected by you know digital distribution right so i uh i started working for apple during school and then afterwards i moved to gupertino and worked for apple and i was in the system software product group i was a product manager and worked on a couple different operating system versions and then um saw this collision of of sort of music and internet happening and um you know there was no need for if you can send a song

digitally over the internet you don't need a cd anymore um and so it was somewhat obvious i think that was going to happen. And I proposed to Apple that given, I guess my view was tech companies would be able to function as pieces of the music industry value chain, maybe even replacing record labels who I viewed largely as, you know, unnecessary middlemen that took too much of the value, still true today. Artists deserve more and we should try to get people, as many people out of the middle as possible and connect artists with fans. It felt like that's what the internet could do. So I proposed to Apple that they start a group focused on exploiting new digital distribution technologies and trying to build a new music business. The reason I thought Apple

could do it was not only do we have the technology, but we were also, Macs had been adopted by most creative professionals in the world, still true today in almost every creative discipline, but especially in music. Every recording studio had a Mac back then, they still do today. Every artist had a Mac. So we had a certain level of credibility, I think, with artists that could talk to them about, hey, there's this new world coming. And Apple said, that's a pretty good idea. Why don't you lead that? So with a couple other people, I started the music group at Apple and we started to investigate different ways of ushering in a digital future. Yeah. What was it like when you first joined Apple? So pre the music group, you know, I think a lot of people, especially in

technology just there were you know fanboys and fangirls of this company and the founder etc but what was it like when you first got there well i was a fanboy too um you know i saw the mac when it you know first came out and said this is amazing uh in fact i used a mac at pen in an engineering school doing all my coding i was the only engineering student who was using a mac everyone else was using pcs and i had to find alternate versions of the tools that everyone was using um so it's pretty rare then it's not anymore um and i love the products i just believe that you could create complex technology products that were easy to use that was apple's mantra make great products so i wanted to work for them in the worst way um someone at school uh i met was

a an apple student rep apple hires i think they still do today like a person on every college campus and sort of deputizes them to be like a junior representative. They paid you back then in gear. So I was like, that's good enough for me. So I applied and I got that job and I got a laser printer. I was like the only kid on campus with a laser printer. That's what I got paid in. So I'd wanted to work for them. I started working for them in the summers, like had summer internships and then went to Cupertino for a summer. And, you know, I was in Kidney Candy store. What was amazing, and then I eventually got a full-time job there. As soon as I graduated, I moved out to Cupertino. What was amazing was just the quality of the people. I mean, it was

kind of warped me in many ways because I was like, wow, everyone is much smarter than I am and knows way more than I do and is, you know, determined. They're passionate. And I thought every place in the world would be like that. Every place is not like that, but it is one of the things that sets Apple apart. And it was an incredible experience. Learned from so many people yeah it's interesting um i worked at facebook for two years and uh similarly i think that's the part that people don't understand unless you go spend a lot of time there or work there is the quality of the people it's almost like that's what they're focused on hiring right just the higher high quality people and like they'll surprise you what they could actually

produce um and it sounds like you ended up actually being one of those people with i don't know if they had plans ever going to the music industry right but hey we hired smart people they come to us with this idea. Yeah, go do it. And then you go execute on it. Do you think that companies can be successful without hiring kind of the best people? Right. I debate this all the time with founders who say, you know, we need to we need to fill that role. And how long do you wait to find the right person versus you need somebody to execute on it today? Yeah, I think it's it's tied to one other concept. You can't just try to hire the best people. I think you have to do two things. You got to hire the best people and you need to empower them. And I think that there's

sort of a set of managerial tactics that I would quantify as Silicon Valley best practices. It's how tech companies are built. And you working at Facebook understood this pretty clearly, Apple and Netflix and Amazon and Google. The concept is you push decision-making down, right? You try to get teams small so that innovation is maximized and you empower them to experiment, iterate, propose things. You try not to centralize all decision-making, try to decentralize it. You try to put all the innovation bottoms up. And what happens is you've got great people who are motivated and you sort of set them on a direction. You get great ideas as opposed to limiting the number of people who are allowed to make ideas, which is sort of the legacy way that companies

were built. And it's proven to be massively disruptive to the way industries evolve. There are courses taught. There are consultants that try to teach half the companies we're staring out the window here looking at in New York City, how to behave differently this way. And so I think that's, you got to hire great people, but you also have to empower them to go be creative and experiment. And I think sort of the core root of Silicon Valley's advantage is this, great people who are future oriented, who are motivated and incentivized to disrupt the past and make things better are allowed to experiment until they do yeah it's um it's really interesting to hear you say that because when you hear it it's like yes that

it's absolutely what's going on there um is that organizational structure or is that more cultural or is there kind of both yeah yeah i mean you have to build the org that way you have to tell people that's what we want you to do you have to demonstrate over time that that's the way products evolve um and uh and then you build an institutional culture around it but it's very common in tech companies i'm using the term silicon valley broadly right we do that at new york companies we do that in austin but tech companies have been are being built this way and have been for 25 years and it's i think proving to be just a leaner more nimble decentralized diffuse get the best out of people focus yeah for sure um when you started the apple music group uh

this is pre or post uh like Napster and kind of all of the decentralized file sharing yeah so um around that time there was this thing called mp3s right and uh this is like there's some young kid listening right now what's an mp3 yeah yeah well uh so this is like the mid 90s like 95 and I think by 96 or 97 the number one search term on the internet was mp3 really it was higher than porn Wow. And that's because that was the only way to get music digitally was to get an MP3. And Napster came about around that exact time as a way to make it even more convenient to get MP3s. MP3 was just a file format for compressing music in high fidelity. It was a format that was not controlled by the traditional music business. It came out of a German standard setting body and you didn't

need a license to use it as a consumer. You needed a license to put it in a product. A lot of people licensed it put it in their products and it became easy to make mp3s from cds and it became easy to play them back and so that was what was happening in the in the mid 90s so it was just as napster was emerging that i was saying this is this is what's going to happen and this is uh any idea of an ipod coming down or previous many years several years before right okay um and so as you started to get into this was it hey we know exactly what we want to go do or was it just there's apple kind of our ecosystem our creativity innovation and there's music let's go figure it out like how much was it a did not know what we wanted to do but we i would say there were two

main thrusts one was we had a set of technologies called quick time which was really a suite of audio and video compression and transport standards so that you could compress audio and video send them over wires like networks like the internet and you could view them on a computer and we wanted to promulgate those into a lot of different use cases and entertainment is an obvious one for that so how do we get quick time more adopted well let's go to musicians who make songs and music videos and start getting them to use quick time to distribute music and second i think it was clear that mp3s were looking like kind of a mess to manage a whole you start getting a whole lot of them and you need tools to manage them

and Apple's pretty good at end user software experiences. So maybe we need software to help people manage their libraries of MP3s. And then you eventually saw iTunes emerge, a tool for ripping CDs and storing your libraries. And this is all pre-streaming. So you had these very large collections of MP3s. Yeah. And the creatives response to this, right? I think today it's become much more popular to whether you're an athlete or a celebrity or a musician to be a tech investor and kind of be entrepreneurial etc back then maybe like the hip-hop industry etc was entrepreneurial in terms of they want to start their own labels and things like that but what was the reception when you started talking about more like hardcore technology right like quick

time and mp3s etc was it a good reception or was there kind of a big learning curve and education work on your guys end yeah you'd be surprised that um the the forward-thinking people were the artists who almost always if you think about it are the ones who are adopting technology very early in the cycle so in the 80s you had all the artists adopt synthesizers long before you know traditional music did and what did that do it created a whole boom of you know sort of particular style of uh style of music or genres drum machines same thing um now thinking back you know things like auto-tune and different algorithms used in music creation so artists are very forward thinking and love to try the new thing, helps them get a different sound. The laggards were

the record companies who at this time and still today are largely run by lawyers since music is really an intellectual property. And so lawyers are by their nature risk averse and I'm very control oriented. And so the record labels, the major labels were complete laggards about it. don't want to hear it don't want to meet we're not going to license any music in digital formats we like the control that we have through cds we want to prolong this as long as possible it's like a fear of disruption or just like a control it's just control we want to control it all um and but the independent labels the indie labels because by also their nature need to experiment they were willing uh to really experiment and the first labels that were

licensing music for use in digital download we're all indies yeah it's it's kind of the like the innovation comes out of necessity right totally totally and and yeah willingness to experiment for sure um so you guys obviously start apple music group it's done amazing things that many people don't even realize kind of what you guys are responsible for um you eventually move on and uh and i think you went to like a private equity firm and started to buy businesses grow them etc two things before that is uh and this is a very typical path when you go work at apple or Facebook or Amazon, you might meet a whole bunch of amazing people and learn best practices, software development, organizational development, how do you hire people.

But many times you might choose to take that to a startup. So most of my mentors, the folks who I admired and learned from who were several years older than I was, that was their path. They were at Apple for a while, then they left and went to startups. So I was like, oh, I'm going to do that. And so I left Apple and I went to a startup. And this was in the mid-90s. I went to a company called N2K, which was one of the first digital music companies. And it was trying to sell digital downloads online. Really? Yeah. And we had a CD store online, like we buy CDs, but then we were really pressing to try to sell digital downloads online very early. But the company went public and was pretty successful. It merged with its rival CD Now, and we were an online music retailer.

But we were also a record label. And the idea was, boy, if we could sell enough music online, we could sign artists and then the artists could sell music directly to their fans and that was you know probably 30 years too early but the concept was a good one it is uh funny that that was the idea because there's some people trying to do that now right exactly yeah um okay so you do that and then how do you go from n2k to um kind of the creation of this digital locker well so the thing that was somewhat obvious was people have a lot of mp3s and uh and they're and they're they're storing them on their local hard drives but the internet connects everything so what happens when you leave your house and you get to work and you want to listen to some music

all your music's back home on your hard drive well it doesn't make any sense right so let's move the music off your local hard drives to what to the cloud um and uh we started a service to do exactly that it was called my play you could think of it as drop box for music and one of the cool things was once you uploaded your songs you didn't have to download them again you could stream them Now, it was super choppy bandwidth at this point. So it wasn't the most reliable thing. And a lot of people had modems at home which weren't high enough bandwidth. So it wasn't a great consumer experience. But you could see it being the beginning of where streaming music would become. Certainly more convenient than carrying a hard drive around. And that company was called MyPlay. And we got to about 8 million users. And we sold that one to Bertelsmann, who owns BMG Music. yeah it's funny because uh i remember being a young kid and everyone was running around with uh you'd get in someone's car and basically the way you could tell if somebody was cool or not was what cds they had in their cd case right and that was kind of the rather than carry around the

hard drive you literally carry around a cd case and um you know god forbid you forgot it at home or you left it in your friend's car or whatever you had no music right yeah and so a similar thing there um got it and so as you're building that i think there's a lot of comparisons it sounds like to where kind of crypto is today um and we'll get into crypto stuff but what was the thought process as you guys are building a service that you know it's not perfect right and kind of you're at the beginning of hey this is going to improve over time there'll be better streaming the technology that the consumer has will improve etc but somebody has to be those first steps down that path, right? Are you kind of aware of it at the time? Is there frustration of, man, we wish

that we could technically solve this? Or do you understand, hey, we're early to this and it's just going to take some time to improve? It's a really interesting question, Anthony, because I was young then. And I feel like today that those years, I was very certain of where the world was going to go i was absolutely convinced that music would become streaming music that all music would go digital that you know cds were doomed that artists could take back some control um have direct relationships with their customers at least know who their fans are no idea who they were otherwise and um and bandwidth would improve like it was just so obvious to me that um digital connection was so much better than waiting for a newspaper to arrive every morning right or or having to go

put a cd in a truck like you know we called it atoms to bits right um that was so obvious to me so i don't think we we're probably more impatient right like why come on these damn phone companies are so slow at rolling out higher bandwidth connections because it will enable such great experiences for consumers so we had to compromise like we had to make websites that could load on slow connections. And we had to do some kind of cross sampling or down sampling for high fidelity tracks so that they could stream over lower bandwidth connections. And you don't have to do that as much anymore. So I think we just tried to accommodate. We were certainly early. I think your point is this is 1997, 1998, 1999. No one has broadband at their home yet. We certainly

didn't have wireless high, um, high bandwidth connections. So you couldn't do this to your phone yet, but, um, we were right. We were early. That's all. Yeah. It's like the, uh, was, um, uh, the ringtones was like a big deal. I think that's like some of the first digital music I remember interacting with. Um, but, but it's interesting because, uh, a good friend of mine constantly says that, you know, the best entrepreneurs are short-term urgent, long-term patient. Yeah. Right. And I think that it's kind of a perfect example of, um, I mean, still today, right. That there's still streaming improvements going on etc and we're talking you know 20 plus years later so also i think it's it's relevant to the crypto conversation we'll get to i guess soon because

i think a lot of people in crypto are are certainly certain that the world's going to be more decentralized and that trust in institutions will continue to erode and so we need to digitize trust um but but none of us know exactly when that goes mainstream in the same way that you know digital music was certain to me we just didn't know when it would tip and become mainstream. Got it. Um, when did you want to become an investor? Like, like what was the jump from building companies? You know, you obviously had success kind of being innovative and entrepreneurial inside of Apple, then a number of companies yourself. Like what was the, yeah, going back to Apple, a lot of my mentors, you know, stated Apple five or 10 years left there,

did a startup or two or three and then became VCs. Like that was the path. And I was like, Well, that seems pretty cool. I like the idea of helping entrepreneurs once you've been a good one. And I like the idea of learning about more than one area. So the real moment for me was after spending 15 years as a digital music entrepreneur, I did N2K, CD Now, I did MyPlay, and I teamed up with some guys and bought eMusic and ran that for five years. It was like, you know, there's got to be more to life than digital music, right? And mobile was starting to emerge and social was starting to emerge. And I said, I want to get smarter about these other markets. And by the way, they feel like they're much bigger. Like there's way more opportunity there. Um, I want

to play in the, where the big fish are, uh, swim with the, whatever. I want to fish where the big fish are. Um, and, uh, and that was really the motivation for maybe trying to become a VC. And I was fortunate enough to find a firm that was looking for somebody like me. Got it. Um, did you know what else you wanted to look at or was it just outside of digital music? there's a whole bunch of other areas I want to go help entrepreneurs in those areas. I'll figure it out when I get there. I've sort of been long internet since 1995, right? Like it was like, this is going to, this just changes absolutely everything. Like it literally takes the value proposition or the, sorry, the value chain of every industry and throws it up in the air and

it's all going to resettle totally differently. And we're still, I don't know what percentage we are into that, but think about the conversations you have with FinTech entrepreneurs who are like reorganizing the insurance sector, right? Or the way homes are bought and sold or the way appraisals are done. Like we're so far into just reorganizing everything thanks to the internet and software that I just knew that we have an infinite amount of opportunity. And I didn't have a specific agenda, but for me, the big one was mobile was happening at the time. It was like, this is, smartphone's going to change everything. Once again, we're going to reorient a lot of pieces everything has to be rebuilt for mobile yeah and it's crazy too uh because it's still

happening right i mean i see companies all the time getting started that i'm like wow that's a really good idea i can't believe that the internet hasn't touched that corner of the world yet um and you know good bet for you to be long internet since 95 but uh you know 25 years later still still happening yeah and i think what you're as an entrepreneur you usually want to ride a wave right because rising tide lifts off boats and and um you get credit for something that's pushing you along. I remember like, I mean, one of the ways that YouTube got big was they did an embed code that's sitting on top of MySpace. MySpace was big and that helped make YouTube go. And so there's countless examples of like companies, you know, Zynga that rode the Facebook wave. You're always

looking for a wave and mobile was a wave, social was a wave. There's a question, it's not, I don't we know enough yet, you know, is AI a wave? And is crypto a wave, right? And when you have waves that are really big, you might argue cloud was a wave too, right? Reorganize the entire sort of infrastructure stack and corporate software, enterprise software. If crypto is going to be a wave, then a whole lot of things are going to get reoriented as well. And that's, I think, what many of us, why many of us are poking around there. Yeah. So some of the things outside of crypto you're focused on right at venrock is um kind of audio and voice um ai especially like around software development robotics etc maybe let's go through each one of those before we get

to the crypto stuff um what's the uh kind of belief or or your view today of like audio and voice right why is that so important and kind of why are you spending time there yeah well i'm spending a little bit less time there okay yeah like i think that we when when you started to see these devices proliferate um so it's like the home devices yeah let's call it siri as a software device right and um you know the echoes and alexa as a hardware device the question many of us asked right away was so is this a new platform will you see audio apps like you saw apps on mobile and will we have a you know cambrian explosion of lots of new um products that are built for voice and uh you know we didn't know the answer at the time but over time it started to become clear that

actually, this feels a little bit more like an and not an or, like you're adding voice to apps like a feature, but it doesn't replace the last mechanism for interaction. So I think the place that voice has been most impactful is in podcasting, where we have very low cost to produce content. And look, in many cases, podcasting is better than radio, right? And it's way more variety, lots of convenient targeting. Here's a cool way to think about it. I read a Pew research report on why Netflix is popular. So I surveyed tens of thousands of people and said, why do you like Netflix? Number one, two, and three answers are one, it doesn't have commercials. Two, I can watch whatever one I want to watch when I want to watch. And three, I can binge watch it. Oh,

interesting. Number four was I like the shows. So the first three reasons why Netflix is popular is because it's better than TV. Yeah, the format. Yeah, exactly. The format corrects the annoying things of television, of linear television. So first thing you learn about that is if you are trying to compete with Netflix,

like your legacy guys, don't do things like we're dropping an episode once a week. That's not why people are going to Netflix. One of the reasons they're going to Netflix is because they can binge. So don't try to exercise control like you did in the legacy world. but by the way not everyone's doing that yeah uh so then you in certain but basically what they're

doing there it sounds like is the legacy guys realize hey there's disruption happening and rather than fully commit burn the boats instead what you do is you almost okay we're gonna go do this but we're still gonna do it in a way where we're fighting the headwind of the consumer trend or the consumer behavior and you know what it doesn't work because the yeah what i think most legacy media companies still don't fully get is the consumers in control. The consumer has infinite choice and just votes with their eyeballs and they go somewhere else. Right. So if you make it high friction or you are substandard experience compared to the competitors, you just lose their attention. Right. So Netflix wins on the attention. Right. Yeah. The thing that is really interesting

to me about the I'll call kind of the video platform wars, if you will, or whatever you want to call it is it used to be a thing where uh the studio created content that content was pretty much they were programming for you they called it programming uh the platforms either they actually own that platform or they had a very significant say because they were the ones providing the content of a lot of competition now what you're seeing is there was these aggregation platforms like prime or netflix they've then used that data and they realize it's all about the quality content well let's just cut out these studios and we'll go sink billions of dollars of our own capital into actually creating originals and i gotta say like it's pretty

high quality stuff right of course look it's um they have the data they know what the customer wants why don't we inform ourselves what customers want and then go pay people to make that you know And one thing I've said before is that guys who deliver entertainment over the Internet can buy programming just like people who deliver entertainment over coaxial cable and cable television. Right. I mean, it's they're all just buying content from creators. The creators are the show writers or showrunners and the actors, the directors, the editors. those are the creators. And the studios had basically a monopoly on buying from them because no one else was buying from them. But now all the internet companies are buying from them. They can do the same thing. Their money's just as good, but they can at least inform themselves

with data what's going to work. So to get back to audio, I think if you think about podcasts, they're kind of better than radio. Why? On demand, no commercials, or at least more tolerant amount of of interruptive ads or no interruptive ads at all maybe an expensive sponsor read um it's a little bit more like public radio that way um but uh i can listen when i want and i can binge watch right and so i feel like um podcasting one of the reasons for its success is it's catering to the consumer just a better product than radio and that's one of the reasons for it it's funny because like content we know higher quality content normally wins right if the consumer likes it they'll figure out they'll jump through hurdles to go get that quality can i push i'm not so sure

that's always true okay why depends what you need mean by quality a lot of times convenience can trump quality but i think we're using quality differently i'll just give you one example music listened to through earbuds yes in a mp3 format is lower fidelity than like something coming off a cd but i don't think that's what you meant by quality right if i like if i like song or yeah if i like uh one band over another yeah i will go and find that band even if it's harder for me to listen etc right um so so agreed the uh the the content in terms of i like it versus i don't yeah uh but that also brings up quality of the format right and so if you think of um one of the things i mean from a personal experience uh we do an ad read at the beginning

of uh each episode after the introduction uh advertisers want you to do them during the episode and there's people who do it i personally hate when i listen to something and it gets broken gonna buy ads so i just refuse to do it and um i always joke that like you actually probably can make more money by it by doing it but who are you optimizing for right who's kind of the customer um and i think that a lot of these uh platforms have figured that out yeah so so i think in voice my view is the the the biggest beneficiary other than the platforms themselves like google and and amazon um is is audio programming right and because it's it's effectively going after radio yeah um ai so you uh are pretty bullish it sounds like um and uh and have some very specific views

there maybe just generally like how do you think of artificial intelligence and where we are and then maybe we can get into something like the software development so yeah well there's just no doubt that it is um a significant change to the way we build software and the capabilities that software can achieve made possible by processing very large sets of data and having computers make some decisions that are, and when their computers are better at making decisions than humans are. So I think we've been looking at two main applications of that that are interesting from a venture perspective. One is, well, businesses, enterprises make a lot of decisions every day. Should I hire this person? Should I promote this person? Where should my sales

team spend their time today? Who should they call? What should I put on my website? What should I put in my email? And how should I merchandise my products? Where should I sell them? How should I price them? On and on. How much inventory should I make? How many different styles should I make? These are questions that businesses have been asking and answering for scores of years. But it turns out that humans are pretty bad at making almost all of those decisions. Given enough data and tuned algorithms, computers can make better decisions than the humans can. So you're really going after human decision-making in enterprises and trying to improve it using data. And that thrust is an investment area of ours. Some people call it

smart software or AI for the enterprise, or just software that improves decision-making in enterprises on a lot of different dimensions now to me that is um it's less revolutionary it's more of an evolution because you know i don't know 20 30 years ago there were people who just made decisions you know i always think like the madman uh you know what i put on the website let me lick my finger stick it in the air and you know i think this is the logo right or this is the motto uh over time the people who have been the most disruptive or innovative said well let's look at the data right who clicked on the most on that button right that type of stuff but it's still a human analyzing the data what you're really talking about is why don't we give the data

to somebody who's even better than yeah just put the computer in there and take the human out yep um so that's happening i agree with you you can you can draw it the way you just drew it and it looks evolutionary and it kind of is like it's just um putting more faith in the algorithms if you will but but what is different is the process by which you gather data label it train the computers and then tune the algorithms is kind of a break from um a more traditional way that software was built and um it requires a bunch of let's call it um artisanal expertise today to do really well and there's you know tens of thousands of people maybe a hundred thousand people in the world who know how to do it really well and probably 80 of them work for apple facebook

google and amazon some at netflix and you know a couple more tech companies right uh microsoft LinkedIn, Microsoft. But very quickly, you realize that over the last seven or eight years, the tech giants have tried to monopolize the talent on this, which just means that they're both advancing the state of the art and applying it in more ways and better than a lot of other legacy companies can. So it's created a bit of an unfair advantage. And that, I think, is what opens the opportunity to sell AI-enabled products to enterprises who can't hire the same talent and and use their own data to improve for sure how much of what you guys are seeing in this space is um what i'll call like the value being placed on the decision making itself versus more

like testing frameworks um in that uh data gathering right the example i always think back is uh when i joined facebook um i was uh running a team that was focused on growth for facebook pages and one of the things that just blew my mind was they had built custom software internally where you could basically feed a bunch of variables into it it would show all the different variable combinations to traffic and kick out you know a couple days later hey this is the best combination that would have taken you know 10 humans in a room five days to just come up with all the different combinations let alone run it and so to me that was less like uh the computer was making the decision if you will but it was more of using that testing framework and being able to actually

in a pretty scientific and methodical way, get an answer. So less AI, but more like testing frameworks. Yeah. So I think maybe the question you're asking is that are there investment opportunities around tools and the stack used to build AI applications? And I think there are. The place that we've been skeptic,

so yes, you need to gather large sets of data, you need to clean the data, you need to label the data, you need to build models, you need to tune the models. Sometimes the models should maybe be tuned like every day, so you need auto-tuning models, and you need to make the models run on low power and sit at the edge in many cases. This is another specialty, have small footprint, optimize models. So all of these are things that a bunch of humans are good at and a bunch of tools are being built to scale, and there are some investment opportunities there. the one place we've been sort of skeptical is ai as a service okay explain okay so like hi i'm a website uh with a bunch of really smart ai people who work here and you are we're staring out at

metlife an insurance company and you want to upload we have a service where you can upload all of your data let's say you can upload all of your claims and uh and all of your customer information and we'll build a model and it'll run and it'll come back and tell you which one of your customers are likely to file claims in the next year. Got it. Outsourcing the model. As a service. We're a little skeptical that you can genericize AI that way, that it turns out that most AI problems require a lot of customization on their solution to make it perform really well. And so it's hard to be a generic platform. Is that a, and this is a little bit in the weeds, but is that a problem where the contributor of the data, you've got a lot of worries about, did they collect it

correctly clean it label it etc and so kind of the inputs to uh the algorithm could be bad therefore you get a bad output or is it just really hard to customize actual algorithm to the problem i think it's more the latter yeah it's uh i mean it's not rocket science to figure out how to clean and label data um although not everyone's good at giving a lot of people credit yeah sure yeah but um but i think the bigger issue is do you have the sort of artisanal knowledge to to know which combination of models and algorithms and machine learning approaches will produce the best outcome. Got it. And so maybe talk a little bit about AI and kind of software development. We were talking before we started recording, and I think that was a really interesting view of how software

development might materially change as AI becomes more mature. Yeah. So right now, I think from a consumer standpoint, AI is being used as a feature in a lot of software. So I mean, obviously your voice assistant is using a lot of AI to understand what words you're saying and then understand the meaning of those words. Google Photos and Apple Photos are using a bunch of computer models, AI models to do facial recognition and find where your dog is in every picture. There's no human doing all that. And those are features that we're getting very used to today. So you could think of those really as being traditional software applications with an AI model sitting in the middle. But if taken to an extreme, software developers, when they make software, make a lot

of choices. Like what, we were talking earlier, what items should I put in the menu? And what order should those items be in? Well, a long time ago, we figured out, well, shouldn't the data tell us what order the items should be in? I remember an example way back, like 1999 or something. hp announced we looked at our website logs and the number one reason people were coming to the hp website was to download the drivers for their printer so we put that button on the home page and everyone's like oh my gosh that's so smart now a consumer can get right to it let me go look at my web blog exactly let me look at my logs so then a bunch of tools started coming out to analyze your logs but it still was like tools analyze logs human reads logs human programmer changes

website. Then we went into this era of personalization, right, in the 90s, where it's like, well, wait a second, why don't we just write a little algorithm that says, look through logs, find five most popular things for this person, and build the menu that way. So that's what we have today. Well, we have more than that today, but most websites are personalizing the experience somehow, putting merchandising products on the homepage that are right for you. Yeah, simple things are, you're in the United States versus somewhere else, show an English website versus something else, right? More complex is, hey, we have your purchase history yeah we know you like sushi so let's show you the food delivery for sushi yeah um but

suppose you took uh everyone's data who comes to your website and every version of the website you've tried and you threw that all into a model and you and you gave the computer a whole lot of other information like the time of day and the weather in the area where the customer is coming from and the computer looked for correlations that uh when you uh when it's cold out and you live in Duluth, Minnesota, and you have this purchase history and you come here, I should show you this parka. And while that's personalization, it may be building the rule itself rather than a human deciding on what the rules should be. And that's machine learning. So we're at that era now where in software development, it may be best to remove the actual

algorithms that a human wrote and instead let the computer build the model, let the computer build the algorithm to decide what should be shown or how to build, what features should be shown to the customer. And if that happens, if that really is true, then software development changes because your software looks like really just like one AI model talking to another AI model talking to another AI model. And there's not a lot of tools to handle that. A couple of companies that are doing that. So I think maybe software development itself could be implicated on a broad scale because of this change. Yeah. And as you think through that, it's almost to an extent your time the computer, hey, we don't really know exactly all the questions you should ask, right? So ask

what you think you should ask to get to these answers. And so it brings two questions. One is, now you're really reliant on the human telling the machine what to optimize for, right? So are you optimizing for purchases, right? You know, one of my favorite examples is, take a media company, they want to optimize for as much time on site as possible. That means you're reading a lot of articles and you know, looking at their ads, they make more money. Well, Google actually optimizes for the exact opposite thing right is how do i get you on the search page and you click the first result and get off the page um and so humans have to understand what they're optimizing for the second thing it brings into question is like the downside of ai right or kind of the fear i'll

call it not even the downside uh where what if the machines you know go off a cliff and they start writing all these algorithms and um you know the facebook example i think is the best one where everyone's like oh they were talking to each other maybe right um kind of how do you think about you know humans telling the machines what to optimize for and making sure you get that right and then also how do you kind of prevent the the doomsday scenario that is probably overhyped a little bit i like that you asked the question because facebook and twitter already are using ai created algorithms to decide which tweet or which post to show you and it's really good because you ever catch yourself just sitting there mindlessly scrolling and that's exactly what it's built to do

That's the the KPI. That's the objective function is to keep you on as long as possible. Right. Engagement and time spent. And it's working. And they're very good at it. They're really good at it. And I think as a result of that, we're starting to wake up to that. A lot of people are saying, wait a second, they're manipulating me. Right. Or I don't want to be mindlessly scrolling and I'm having a hard time stopping. And that's because like they know the psychology of the dopamine hit you get from seeing yourself mentioned or someone like you. So it's all built, it's all, it is manipulating us. It's just, you know, it's going to our base behaviors and keeping us on their website longer. And then some people are starting to

wake up and say, I don't think, I don't think that's good. So what are we going to do about that? You're going even further. So that's today, like that is happening today. You're going even farther is like, can, can the software start to do bad things to us that its masters didn't ask it to do? We start to get into a philosophical conversation here, but the, the distance from where we are now to like computers deciding to behave badly versus them doing exactly what they were asked to do. But it's, you know, some unintentional consequences. I'm, you know, we're just nowhere near general intelligence, artificial general intelligence. And I'm of the mind, I'm a computer scientist, but I'm not a data scientist. I didn't study AI in school.

But I'm still of the mind that, you know, the smartest people in the room tell me that uh, we're just nowhere near that. Right. And when you see how long it takes to get some of these algorithms tuned to just do very basic things, you realize that they're not smart in the, you know, they're not intelligent in the human sense. I always, uh, go back to this idea of, it seems like a lot and you're much more educated on this than I am. Um, but from an elementary view, it's like, look, the machines are just doing what the humans would do if we could, right. Which is basically taking a bunch of information, processing, make a decision. Yeah. What they're not doing is overlaying that with the one thing that humans are better at the machines which is

the human intuition and kind of all of the um the general intelligence etc what that leads me to is um i've had this idea for a long time that let's just you know kind of admit the machines are continuing to get better and better and better and eventually humans uh will trust those machines and software over other humans right but if you flip that back to what we're talking about the audio side it's almost as if all of a sudden the creativity that general intelligence will become more valuable in the future because the repetitive uh type tasks and decision making etc the machines will kind of have a monopoly on that what can the machines not do right and so whether that's content creation um i've recently met a guy who he's rolling up uh ad agencies um and his whole

idea is look there's everyone focused on how do you get more people to click on the facebook ad but there's still this niche of well you need somebody to come up with the ad campaign right and so his whole idea is that creativity will become more valuable over time etc do you buy to that or do you think that the machines can eventually evolve from just a repetitive task to they literally will create content and things like that that are just optimized for what we want to see today i think when we look at the application of machine learning towards creative things like visual storytelling or writing text right or even music creation because there have been a lot of applications people trying that stuff they um most of the stuff is interesting

but derivative because no surprise you feed it a lot of mozart it you know spits out something it sounds strikingly like Mozart. That's amazing in and of itself. But when you think of like, will artistry be subsumed by computers? I think two things, two reasons why it probably won't be for a very long time. One is, I think when you look back at the great artists, and I use that term super broadly in all forms, in all media, they were initially doing something that was a significant break from the past and may even be heresy right like this is this is what is this like i don't understand this this is terrible it doesn't sound anything like what other people are doing doesn't look anything like what other people are doing um and yet it's that um that that break

that is the demonstration of their genius it's their it's a it's original thought right it's something new and that's what can be the indicator that something great is happening and it takes time for the rest of us to catch up and say wait a second that is really great uh and even though it's different that's one of the reasons it makes it so great um so number one is i don't i don't know that we're going to be able to just tweak the algorithms and say yeah i fed you all the mozart but now do something completely original yeah the second is i think we as humans like to understand the humanity behind the creator what story led them to do this and we can make up a story about a computer that did something or pretend it wasn't a computer but boy I think

we like to attach what is the Billie Eilish backstory about why you know and boy and now I really appreciate her music even more right so and we're fascinated by that and that's never gone away so I guess I'm skeptical that like all the great moments in arts broadly will now be done by computers however there's a lot of fields where let's just call it derivative um art won't be humans won't need to do anymore so like production music background music what music do you put on the bed of a commercial or boy you know that's that can that will be able to be created by computer it already is um animation like i drew all the main characters and the cell and the main cells now fill it all in computer really good at that stuff um so i think there

it does impact human creativity, but not in the dramatic way like, okay, we're never going to have another great author again. There'll be no more Hemingways. Yeah, absolutely. So we've talked a lot about kind of automation in the software world. What about in the hardware world around like robotics and stuff? I know you guys have spent a lot of time there. What's kind of going on there? What are you guys excited about? What's interesting is the macro. So we've basically at full employment in this country and we have rising labor costs, you know, hopefully getting to $15 an hour everywhere, but then it doesn't stop there. It keeps going up. And so we have people who employ humans who need a bunch of different jobs done that at $15 an hour or more start to become economically not possible to do.

Right. You hear about farm labor, the labor costs going up such that they can't a farmer can't afford to produce food for us profitably. Um, you hear a lot about, um, um, uh, like a warehouse type jobs, you know, highly repetitive, low skill labor, where if the cost, if labor costs keep going up, we won't be able to profitably produce, um, different goods. And that's why they've, those goods have been moving overseas to be produced. So, um, one answer to that is, uh, automation. And the thing I want to stress in we've looked pretty closely at these many of these markets usually it's not it's almost never displacing human labor the humans are going doing something else that's slightly less repetitive something okay interesting right that because the higher wage jobs are for jobs that humans are good

at and that um that computers and robots aren't that good at um we have a company called simbi robotics simby yes simby s-i-m-b-e um the problem they solve is like literally half a trillion dollars every year is lost to you going into a store and the item not being on the shelf okay 500 billion dollar problem wow the items often in the store it's just not on the shelf it's in the back somewhere where it's coming in a few hours but it's not there um and another half a trillion dollars is lost to an item being mispriced or not being displayed properly. Can we chalk all that up to human error? Yeah. It's humans either not being good at that or just being like perennially bored at it. Like

who wants to walk up and down store shelves every couple of hours and see what's missing from the shelves, right? Or what's not displayed correctly. Humans hate that job. And that job is often unfilled in many groceries and retail stores because of that. No one wants to do the job. High turnover so this robot does that drives up and down store aisles finds all the things that are out does it a couple times a day sends a text message to the store manager you're out of mayo on aisle three position seven or the items in you know aisle four position seven are not displayed according to the planogram so it doesn't actually quote-unquote fix the problem it just identifies the problem maybe in the future i could fix whatever but right now it's just an interface

between I'll do a job that the humans don't want to do. I'm really accurate at it. I basically can then alert the humans and the humans will come fix the problem. Yep. I don't get tired. I do it all day long. I don't complain. Yeah. I'll do it at nighttime too. Um, it's an automated counting machine, right? Uh, and humans do like computer vision or computer vision. It's got a lot of sensors on it. Obviously it's got to know how to drive around, can't hit people, can't run over kids, you know, shouldn't bang into carts, got to avoid spills. Um, you gotta be able to function in the dark like a lot a lot of stuff and so it's pretty sophisticated robot uh but it does really well and it really solves this problem brilliantly yeah um and what's happening is retailers are not

firing the labor who used to do that the labor doesn't want to do that job they go do something more useful like interact with customers to your point what humans are good at yeah and and so take that system for example uh if i'm a grocer i want to use this do i have to like retrofit my entire store with sensors on the ceilings and all this kind of stuff or is it no that's what's cool about this company actually it's part of the investment thesis is you literally plug it in yeah it's got a little self-charging pad you got to set up a little area in the corner of the store and it's got to integrate with your inventory management system so it knows uh it can tell you like this skew is missing but it drives up and down for a day or two orients itself builds

a map and then you tell it how often you want it to run and it works so it's meant to to solve the problem without you needing to retrofit at all yeah it kind of operates in the environment the way humans did. Yeah. It's pretty interesting. Um, what's your thoughts on, uh, on these like Amazon go stores and what I'll call kind of the, the ultimate culmination of a lot of the different things we're talking about here, where it's software, computer vision, robotics, et cetera. How do you see that kind of plan? Yeah. So that eliminates the, uh, cashier experience. And, and the first time you went into an Apple store and like, you didn't have to go stand in line for a cashier. You were like, well, that's a better experience, right? Uh, this is even better

than that right because you don't have to find the person in the store you just leave um so i think that's a great you know going back to the netflix example that is a better version of retailing than legacy retailing so yeah we're going to do more of that but that requires a really major retrofit and a big expense so this solves an interesting set of problems without requiring like you got to build a store very differently so i think for instance i think simbi would be adopted is being adopted way more quickly then let's rebuild the store to be cash cheerless although that's happening too i think just much more slowly yeah so broadly i think we're looking at a lot of robotic solutions that do the jobs that humans don't want to do

or aren't very good at so you can lower the cost of production and humans can be deployed to do better things i um i forget the exact stats so no one tweeted me and tell me i got it wrong but it was something like uh maybe 200 years ago or something it was like 80 percent of people were farmers right well you know some long period of time ago a majority of humans worked on a farm and it's because we need food to live etc as quote-unquote automation or technology got incorporated so literally just a tractor or things like that uh those humans didn't all of a sudden you know die off because of starvation or not having a living they went and did other things that you're saying and so um this happens kind of you know periodically uh in these technology

shifts and um it's interesting to see i think uh it's happening to more white collar um employees than probably previously uh but it's also still happening in you know a grocery store or elsewhere um and i actually believe that the white collar populations are not expecting it whereas i think you know you talk to a truck driver talk to these people like they know this stuff's coming and so they're kind of quote-unquote prepared uh the white collar folks i don't think are nearly as Yeah. So I think broadly, we agree with your point of view that something like, you know, 30 to 45 percent of all jobs are automatable. And that's 30 to 45. It's a pretty big number of today's jobs. Yeah. Some are blue collar and some are white collar. And over the course of 20 or 30 years.

And so it doesn't mean that it will all be automated over that course of time. But but the job function itself was likely can be automated. Again, I think that history shows us that that doesn't throw a lot of people into unemployment. It makes systems more productive and lets humans move on to higher order tasks. So I am confident that over the long term, it's a good thing. It can create a bunch of economic disruption during the transitions, which can be pretty painful. And I think we're destined to experience some of that. Some political candidates recognize that.

They're talking about it a lot. And some ignore it. but it is a potentially societal issue. But it is inevitable because the march of technology is largely unstoppable. I mean, look, here's a good example that most people read in the New York Times over the weekend about this company Clearview AI, which is doing something with facial recognition that has been possible for a number of years,

but Google, Amazon, Facebook chose not to do. For those that don't know, essentially in an overgeneralized manner, uh, if there's a photo of an individual, whether whoever has it, they can essentially put it into this system and this system will then go match it to photos all over the internet and come back with, here's all of the data associated with these photos. So you can find people's names, their locations, you know, information that, uh, actually when I read the article, uh, I was kind of like, well, duh, most of you people put that information on the internet anyways, the computer is just being able to find it better. Um, but obviously, you know, if you look at it from like

a law enforcement standpoint, the fear mongering scenario is I take a picture of somebody on the sidewalk and next thing I know, I know where they live. Right. And I don't like that system because that feels more like a surveillance state. Yeah. Or a government that doesn't like people who are peacefully protesting outside can take pictures of everyone, find out exactly who they were and have the IRS audit them. Right. Without you realizing that it all happened. I mean, there are a bunch of scary scenarios that happen. My point in bringing it up was a lot of people possessed the technology to do this for a number of years and chose not to until one entrepreneur came along and chose to do it and so that's my point about like when it becomes possible

it will happen yeah do you think it's bad not necessarily that specific scenario but just do you think it's bad when some companies have the technology they choose not to pursue something for whether it's ethical reasons just not good for their business their opportunity costs whatever and then somebody comes along and does it or do you think of it it's less good or bad and it's kind of that's how technology progresses i view the march of technology as inexorable and you know it's like cloning right like of course china was going to do that right so somebody was yeah right so someone will do it whatever technology is possible someone will do it so it leaves you with two choices you can legislate this is where governments come in and you set hard and fast

rules about what society will accept and what it won't or you try a more delicate approach in getting um responsible companies to self-police right and there are examples of that that work both of those approaches working and there are examples of both of those approaches failing this is a debate the society should have i think most of the you're hearing this now i mean like um uh ceo of google just said last week like we we are seeking legislation from the government on ai that is basically him saying the government and society should decide what are the limits of ai because we don't want to be the arbiter. We will adhere to the laws, but the government should set the laws. Yeah. This is happening in a lot of different aspects of intersection of kind of

society and technology, right? Censorship, I think, is the easiest one. I actually don't have a view because I haven't spent enough time thinking about it. But one of the most interesting debates that seems to be kind of percolating now with the U.S. presidential election coming up is should platforms have a say in what political ads are run and are not run so everything from accuracy to just should we have political ads should we not and it seems like each platform is kind of taking a different approach and um you know given the in light of 2016 and and kind of all the accusations from every nation state in the world at each other uh you just get into this weird world of like uh one who do you trust right in terms of the information and then two like who

should be making those decisions and i look there's 330 million people in the united states i don't think you're gonna come up with an answer that everyone agrees on right and so this is going to take us to crypto because really ah you you see my lead yeah i mean this is the root of why a lot of people are excited about crypto is that what happened in the arc that we just discussed from 1995 until you know 2019 um we the internet enabled incredible things everything went digital every bit can move around the world we've wired it the world we're three and a half billion people we're going to get the rest of them on the internet as well everyone's connected it lowered the cost of publishing freedom of expression not in all countries but in many um it democratized

access to a lot of information and uh and it did two other things that are the sort of unintended consequences or at least consequences that maybe not everyone foresaw one it allowed for manipulation Because if the cost of producing information is really inexpensive and access is free, then you can put out non-information, disinformation, right? And people are doing that. And guess what? That's working. And two, I think what we didn't really foresee is that the new platforms would combine data with access and the data would become centralized and powerful. and so now we democratized access with the internet um but what what crypto i think hopes

for many of us think about as web3 is can we democratize data so no one can have a monopoly on your data and and so you sit inside of um you know one of kind of the the most respected venture capital firms in america right and you guys are investing in all kinds of different industries and maybe we just talked about why pay attention to crypto right is it a thing where hey we see opportunities to make money and our job is to take lp's capital and go make money um is it this is gonna touch everything and so we gotta pay attention to it uh is it personal interest like when you guys first start to look at it i don't think there was a lot of venture capital firms that were like you know i'm starting a venture capital firm focused on crypto i think

it's naturally been this progression it's like what drew you guys into it well for for real we are followers we are looking for where where is the future going irrespective of what we decide to do and we want to be on that path and what what we're looking for are waves that are not just small changes in the future, but big ones that can disrupt legacy industries and reorder value chains and be vectors themselves for progress. So we've seen a lot of those. In my life, I've seen a bunch of them, right? I saw sort of personal computers. I saw internet. I saw mobile social and cloud and now we might be on the cusp of seeing another ai not clear if that's going to be a wave or really a sort of evolution like augmentation right it's definitely something

um and there's good debates on both sides and crypto right decentralization do you have more confidence in ai being a wave or crypto or do you not look at it as like a comparative i don't know I never sort of stacked one against the other. I'm investing in both. When I say I don't know, I don't know. Like, I mean, really, my human brain is not capable of predicting. And so I'm looking for a signal. In AI, I think we are seeing very strong signal that it makes highly more performant software.

It works. Better than humans. It works. And so it's going to be applied everywhere. Does that reorder markets is an unknown question. And we're watching a couple indicators to see if that's true or not. Or is it just everyone gets good at that? just like everyone got good at cloud yeah um with crypto i feel like it is such a departure

both soft from a software architecture perspective the way you build things and from a philosophical perspective and the last time we saw as big a philosophical change as sort of web3 was um really open source uh and there's sort of this um fred wilson calls it uh internetness right there's there was this almost culture among the people, really, not almost, there was a culture among early pioneers of distributed networks, of the ARPANET and the internet, that information should be more democratized and we should remove gatekeepers and everyone should have the power to publish and software should be not controlled and behind walls, it should be open source and intellectual property should be more distributed. And that was the ethos of the internet, right?

It's kind of one of the things that led to Wikipedia. These are amazing attributes of the millions of people who worked on rebuilding the entire tech stack for the internet. There's a question now, is crypto that? And certainly if you talk to many, if not most, of the early developers working on crypto related projects across the whole gamut of crypto, There is an ethos there, right, of we do not trust middlemen, we do not trust banks, we do not trust governments, we may not even trust the media anymore. And we need to find software systems that enable people to communicate where you can trust that there is no more middleman, so we can trust the communication is authentic, and we'd like to exchange money or value or do transactions without an untrustworthy middleman as well. and we want to build that architecture in a way so that your data is your property and you can revoke it at any time and it can't be amassed by a single party anymore and though there's an ethos

there and that's a very powerful ethos if it sticks like I mean I think there are somewhere between tens of thousands and hundreds of thousands of software developers working in crypto right now it changes many of them have some piece of this ethos if not all of it and that's going to drive major change if it holds. So we're watching that closely and that could be a wave and that would be a very big wave. Yeah. So this brings up a question that I think a lot about. I do not have the answer because I don't even know if there is an answer, but it's this balance between society believing something and the technology being available to make it reality. Right. And what I mean by that is it's pretty well documented in these Pew surveys, et cetera, of the degradation

of trust in institutions institutions and centralized authorities right so yeah you see this in the media government kind of a lot of things you just named and that's not a hey in the last two years that's happened right i mean this has been happening for a very very long period of time um and to me that overlaps very uh interestingly yeah to the internet right of just um you know i used to believe everything that i saw on television because that's the only source of information i had now all of a sudden i got this other guy who's you know my favorite video uh probably ever on this is uh have you ever seen the weatherman uh and he's standing yeah and it's like a it's a clip of a weather guy okay and he's standing in a parking lot and it's raining

hard and he's standing and it looks like the wind's about to blow him over right he's struggling to stand there and in the background there's these two young kids in shorts and a t-shirt walking by her and they're just walking like they're walking to their car after they eat lunch Right. And again, I get it. Right. You know, some people would argue that television is entertainment. Right. It's not just news, whatever. But it's a perfect example of like, come on. Right. And so over time, this has happened. But could we ever do anything about it? Right. Yes, I trust this less and less, but it's still the best source. Right. It's still the only thing that I really have. I think the Internet kind of started to expose some of that. But now really what we're seeing is software engineers who say, wait a minute, I can use all of these technologies that are available to me and I can build something that can actually give us a different solution. Some argue it's better, some argue it's worse, some pros, cons, whatever. But is that now just happening because there's a technology point in time that makes it possible? Or is it that societal like shift in ethos where people say, no, no, no, no, this is a really big problem.

Let's go focus on solving it. I think it's the latter. but okay it's more the society component oh if you just ask the developers working on decentralized technology why are you doing this and the answer usually corresponds with the decline in trust of institutions in middleman many people who are really excited about crypto come from eastern european states with highly distrust you know untrustworthy governments and media state run media um surveillance states uh where people disappear and when you grew up even some of your life in uh which i did not um but if you if you did you could appreciate how your perspective

um and your motivation to want to build technologies that make it really hard for people to surveil you uh to people disappear you to have your data and know who you are and where you are um you could see why that might be something you want to spend your time on right And you believe the world might be better with some of those things. And it's no surprise that encryption, you know, is the basis for all this stuff. So I think that's a pretty powerful force. And I think your observation that Bitcoin effectively coincides, its rise coincides directly with the Pew and other data that shows a decline over 10 or 20 years in trust in institutions is not a historical accident. Yeah. Let's talk about Bitcoin and then let's talk about kind of crypto and the impact of that. So Bitcoin specifically, the way I think about it is it was the first real application of a lot of these technologies. It's got this kind of immaculate conception story, right? And it's had this incredible rise in US dollar price. how do you guys think about it um in terms of an investment opportunity given that you've got a venture fund do you guys do anything there uh and then also how do you think it in more macro just this is a piece of technology that could have a big impact on the world but but is that likely

do you put kind of some kind of handicap on it etc um i think a few thoughts one with bitcoin in particular i appreciate uh many things about it but two things in particular is one and they solved a hard technology problem a lot of computer scientists would tell you you know we had merkle trees and there it wasn't a major innovation but um whether you call it major or minor it is uh been around for a decade and has been unhacked um unhacked successfully and it um it solves a the problem of uh you know single transaction no double spend and uh distribute distributing trust on a transaction network. There's a bunch of improvements that can be made to it for scalability. The way Ethereum looked at that and said, well, hang on, we can do more than just

manage a ledger. We could actually execute code. We'll call them smart contracts, but they really can have logic to them. And we can do that decentralized. And you can be sure that it's secure. That really blew us away. That was the moment when I read the Ethereum white paper. I wait a second this is this is not just like some uh currency for evading uh you know for drug dealers like this is this is really significant and and so that's informed my view of how my little brain thinks a lot about crypto there there are use cases related to transactions whether that's store of value or you know money um and then there are then there's an entirely different uh opportunity around rebuilding the web stack where applications and data are

decentralized uh which is a as i said sort of a fundamental shift to the way we architect apps and i really think it's important to talk about them separately because in my mind i have to separate them because i think they're very different investment thesis theses around i agree with you yeah um on the currency money side uh do you guys do anything there or are you more focused on the the latter of kind of the computing platform and the decentralized the web yeah so well personally um i've been you know buying cryptocurrencies for a while also mining them i i mined bitcoin very early on just as an experiment when you could still do that i mined crypto for two years sorry i mined ethereum for two years and a couple other coins as well i've got a lot

of spare gpu if anyone wants to use them i've been training neural networks with them um and and so i did that really just to learn. And, you know, personally, you know, I've been an investor. We chose to do early on was to just try to get smarter. So we weren't smart enough here. There were other investors and entrepreneurs who were way smarter than we were, who saw the opportunity earlier than we did. So we looked for a partner. And what we found was this great group in Brooklyn called Coin Fund, who manage really a venture fund that is a combination of cryptocurrency that largely usually is illiquid at first. So they get involved very early in projects.

They do protocol design and digital economy design. And those currencies eventually become liquid. And also equity investments. But they're sort of an early warning indicator for us that really help us sort through the thousands of crypto projects in a lot of different dimensions. And so we invested in them, and they're our partner. They've been very helpful.

The partnership's been very successful. And they now have multiple funds. Some are liquid only and some are non-liquid. So I think we participate that way. That's one way. And separately, we have made investments ourselves as a fund in both in companies that denominate their value in tokens and companies that denominate in equity. Oh, interesting. So we are comfortable doing both. And so when you think through the value of the equity versus the tokens, kind of the value accrual, how have you guys thought about this? Because I think this is probably the biggest question I get from the institutional world of institutional investors realize something's happened. Too many smart people looking at it, working on it. I got to do something.

They understand equity, right? I'm going to get an investment. It's going to go public. It's going to get acquired, whatever. there's also this token accrual model where uh it kind of looks like equity but sometimes it's not sometimes it is right it just looks a little different do you guys separate those or do you look at it as there's companies and products being built and we'll do both and we frankly don't care which one is which yeah i think i kind of think of three actually you got equity you have security

token you have utility token yep okay and the reason it's important to separate that is the security token is really trying to be a proxy for equity ownership, right? It may have a different set of rights, and it's either equity in a company or equity in a piece of property, right? A physical asset. So there's a whole category of people in crypto focused on fractional ownership of real estate or art or sneakers or where you're taking a physical real world asset and you're trying to fractionalize its ownership into a liquid market that otherwise wasn't liquid. We are fortunate that RLPs allow us to take a broad view. We are used to having more than just ownership in our companies, but having some rights that protect our ownership. And you often don't get much of

that if you're doing security token as a proxy for equity ownership. So we tend to be, I would say, even in the crypto projects, the world really swung after 2017 and beginning 2018. A lot of projects went back to raising venture capital in the more traditional model. So the overwhelming majority of our crypto holdings if you will our equity often when we're buying a token there's an equity right associated with it or it's a combination so we're pretty flexible but i just think that's where the market is right now yeah absolutely um you uh you did an interview uh with my fiancee uh she's a much better interviewer than you are absolutely she gives me a hard time and critiques me all the time um she says i'm too nice to people you are she's she's tough

and one of the things that you said in that interview that i found really fascinating she asked you about kind of crypto replacing VC or competing with it. And I think this was more kind of the token investing. Anyone in the world can participate. And you actually took the view of like, we should encourage this is a good thing. Maybe talk a little bit about like how you view something that I think most venture capital would say, wait a minute, I like our business, right? We're pretty good at this and we can make money. Why be so kind of bullish on the crypto investing that that's been going well I don't like gatekeepers I just think they're endowed with an inauthentic authority right that somebody put them in this position through a set of fortunate

circumstances and they are making their best but usually arbitrary decisions about whether allocating capital or deciding who should be seen in an art gallery or you know who should what song should be played on the radio or you know whatever I don't what what books go on store shelves. I just don't like that. I just think it's sort of fundamentally unfair. So we should try to get rid of gatekeepers, right? And let consumers decide what they want to listen to, watch, et cetera. Give them a lot of choice. So VCs are gatekeepers, right? We allocate capital. To the extent that you can get more funding sources for entrepreneurs, if we can get more entrepreneurs to have more at-bats at trying their next great thing, I think that's great.

That'd be good. That's good for all of us. It's entrepreneurship drives economic growth, wealth creation makes jobs. It's probably the most important engine of our economy. You want more entrepreneurs. And why should VCs be a gating factor for having more? So I'd like there to be alternate mechanisms for funding, Kickstarter and things like that. Some have worked, some haven't. I think it would take a long time to perfect. So I'd like to see that. To the question of like, does it replace VCs? I'm skeptical of. And I think they can coexist. because I think that if you look, VC as a function has actually withstood a lot of different change. It's been around for a long time. I mean, the Lawrence Rockefeller started doing this type

of investing in the 30s and then Venrock, you know, more formally in the 60s. And it hasn't changed all that much, right? So there must be something about human, even in the face of a lot of technology and market change, humans still seem to be decent at being an allocator of selective capital and then helping build those companies. So I don't think we go away. I will steal a little bit of Mark Andreessen answer to this question where he's been asked this before. And he sort of says the nature of companies seems to be changing a little bit, right? We've got on one extreme like DAOs, right? And on another extreme, you know, just remote work or different ways of building enterprises. If that changes a lot, if they're token based and not equity based,

you might have to have a real change in the way capital allocation is done. And maybe VC has to change a lot. But if companies are still mostly kind of what they are today, then it might not change. Yeah. I mean, that's logical, right? Of just companies have been the same way for, you know, tens of years. And so capital sources doesn't change. This brings up the question, I guess, of kind of competition in the internet. And I found online that you many years ago did a testified around this idea of competition in the internet. And I think that every so often this kind of resurfaces, everyone wants to debate kind of monopolies, not monopolies and centralization of data, all this kind of stuff. But competition in venture capital has pretty much stayed the

same. Now we're seeing competition with the like average investor, right? If I can just set up a website, accept funds from anyone in the world, like that's pretty competitive with venture capital. How do you think about the internet empowering competition? And like, how does that intersect with some of the things we're seeing in crypto the argument that the internet giants have made around this is that our competitors are one click away right so we can it's very easy to change that might be the best way i've heard their argument summed up that's what they say the truth is that data right is the monopolistic fuel that empowers the sustained market dominance a lot of the internet giants personally i love the products of a lot of the internet giants so

it's kind of hard to hate them whereas like you hate your telecom company because like the service sucks right or you hate the airline because they treat you like your cattle it's sort of still hard to hate Google even though they're so dominant because that website is awesome the search engine is great the email is awesome thank you for YouTube so I think we as consumers are lulled into our joy in using the products, but there still is some market dysfunction that's happening because of their large exclusive access to our data. And so I think that leads us to crypto. And one of the areas of crypto that I've been fascinated by is can we build not just a decentralized compute architecture, but a decentralized data architecture. So just in a

small first step, we were talking earlier about the Facebook and Twitter algorithms and how their machine learning algorithms are deciding what we see. Well, suppose we could all take our data, which is like our tweets and the things that we've read and favorited and who we're following. And we can take that wherever we want to whomever's algorithms we like the most. Or reversed, Twitter could say, we'll just be the place where all the data is living, but anyone can plug a different algorithm in. So if you don't like the tweet stream we're showing you, let's use some other people's stuff who might not be so inclined to allow politically manipulatable speech to appear in your feed, right? Twitter seems to be, or Facebook seems to really like putting that in your feed. And Twitter is not so sure how they feel about it, which is sort of typical of everything about Twitter.

So I think there are other, you as a consumer may choose different algorithms, right, different feeds. And you can't do that today. Um, so there's a couple of different ways to re strike the, the dominance of the internet platforms. But the biggest way that starts is if you have a sort of revocability of your data and you can take it elsewhere. And that is made possible by decentralized architectures. And this is what excites me so much about the decentralized web. We invested in a company called Threebox. So if you try to write a decentralized application, a DAP, two things have to be really true for that to be decentralized. One is the compute, the code itself is run on a

network of computers, not on like your server sitting in AWS. All right. So it could run on Ethereum or any other different layer one smart contract platform. And two, the data that is exhaust comes off from your use of it photos you upload posts you write things that you like your profile all the things that facebook twitter etc would capture yep everything that they capture or google when i search for something and click on a link all that data really can't be stored on the server of the application company that made the app it has to be stored decentralized throughout a web, a decentralized web. And it needs to be revocable at any time. You need to say, I'm taking my data back. You can't have it anymore. I don't trust you anymore. I don't like

you anymore. The decentralized compute piece is being worked on by a lot of layer one smart contract platforms. The data part, which is in many ways even more critical, is really a hard problem to solve as a developer. So 3box says, we got you. You just write to our API and we run a decentralized data network and your data gets stored in IPFS on an Orbit database and with our schema. And we even will do an identity layer on top of it and it's all permissioned and you can be sure that you're essentially compliant with a decentralized architecture. That to me is really exciting because that enables the decentralization of the mainstream platform. So a couple of questions about this. One, the name 3box, is that because the data gets stored in three boxes or

more type uh mentality or is that a i think the three is more on web three got it uh-huh and uh because they're the data gets stored and you know an infinite number of places you got it uh and then in terms of uh if you have an application i'm a user i access your application you're using three box uh or you can't see kind of all the data right you're not capturing it uh what controls do they have over my data yeah so first of all the data is fully encrypted so no one can see it but people have you you've awarded permission to an application to be able to use it got it and you can revoke that permission so three bucks cannot see your data they do know which applications you have allowed to use it and they will help manage the revoking of that permission

but they're really a um they're a friendly agent they're really just an agent for you it really is they're operate a network of compute well they're not even their computers it's like anyone can put a node on a network just like you can spin up an ethereum um node yourself anyone could spin up a three box node and um and be a participant in the network um and so so they're really a protocol design you know they've really created a protocol and a schema and a bunch of software around that that in itself is open sourced which is sort of the ethos of web yeah Yeah. This brings up an interesting point. I'm really interested in kind of the increase in efficiency of hardware and software that my family comes out of the data center business, right? And

so that business is notorious for, they got a lot of hardware sitting around that's not being used, and they've got a sales force out trying to sell it, you know, whatever. All of a sudden, you get these, whether it's, you know, decentralized file storage, right, three box, whatever it is, you start to build these marketplaces of, well, what's my incentive to have my sales force go sell this hardware space or software, what can I make from that? Or there's persistent demand through one of these decentralized services.

If I just go ahead and point it that way, then there's revenue opportunities. And to me, it feels like old school versus new school a little bit, right? And the companies that can build those platforms or protocols that facilitate the decentralized component, it just feels like a really big opportunity that a lot of the legacy players don't see coming.

Yeah, I think one of the things you're scratching at there is that many of these decentralized protocols have an economic reward for participating, right? So we call that mining in one parlance, but other people have some different words for it, but it really means like contributing your resources to a network, whether they are compute resources or storage resources.

There should be economic motivation for you to do that. Otherwise, why would you do it? you have some material costs not just the capex you bought the computer or hard drive with but the power it takes to run it right and maybe the space you're paying you're paying to put it in and so the reward for running it has to be greater than those costs and um and i think there are some data centers who have a bunch of spare hardware that have have low power costs who can do quite well just adding their machinery to nodes but there are a bunch that are in new jersey where you pay, you know, I don't know, 16 cents a kilowatt hour and their profit's going to be

much lower than the 5 cent, you know, maybe in upstate New York. So there is some regionalization and other environmental dependencies to make, to see whether it's a good business. I think what you're scratching at is if these networks work, people may be able to take, you know, underutilized resources and throw them into the network. And that is the efficiency of it is that we are, well, you know, the data center business when virtualization came out, it's like, well, we bought a bunch of hardware, but we have this tiny little website running on the blade server. It's using 5% of the load. Why are we doing that? Let's slice it into 20 more websites and we'll share the load. Right. And so virtualization was an efficiency play. I think decentralization

is too. Got it. One other theme that you guys are investing alongside is crypto collectibles and kind of NFTs, digital collectibles. Maybe talk a little bit about what's the excitement there and kind of what have you guys done in that space so far? Yeah. What excites me in this space is again, it wasn't me who thought of it, you know, like, like all of us in 2017, I saw Crypto Kitties come out and watched it explode and break Ethereum in that month. And, and that kind of got me thinking what, you know, I collected baseball cards as a kid. I saw my kids, you know, play with pokemon cards they were just as big if not bigger than baseball cards i started looking into the size of collecting and it's a massive market like humans do it humans not all but many humans

collect something art shoes wine baseball cards antiques um there's and then i started reading up on the psychology of of collecting why do humans collect and there's a psychologist who have documented like seven different reasons why we do it and it's like a thing humans do so it's really big well i like to see pre-existing behavior as a a proof point for why something can succeed so collecting is big what we don't really do is do is collect anything digitally yet there's some evidence we do like um in-game items in but but those are centralized so and they have fortnight fortnight yeah cape or something but we can't it's very hard to sell that it's so collecting one of the reasons why we collect is like for investment reasons we want to hold it and then

profit from it later. You can't really sell your Fortnite account. You kind of can, but you have to do it on eBay and you lose your whole account. So you can't really sell the items. Most of the most valuable items in games are not liquid. You can't trade them outside the game. But what CryptoKitty is, what Dapper Labs demonstrated is you could build a crypto collectible. You could build a digital collectible that has provenance. You know who owned it, has authentication. It is real, can have scarcity. There is only one. And it can be traded decentralized on exchanges everywhere. So, wow, I could buy a piece of a digital item or digital art. I know that I own it. The company who built it could go out of business. It still functions

because the code is totally decentralized running on an open network. And I can sell it later if it increases in value. That to me feels like that could unlock very large digital collectible market. It hasn't happened. It hasn't crossed over in the mainstream, but that's the thesis for our investment in Dapper Labs, and they are trying to take what they built around CryptoKitties and now do that with the NBA. They've announced a big partnership with the NBA. They're going to have a game come this year that allows for moments to be collected. And those are, there can be rare ones and not rare ones. And you will know which ones you own and you can use them in games themselves and the way you use Pokemon cards in games. And hopefully they'll go up in value. And

so that's a pretty interesting experiment I'm excited about. There's one other area that I'm not doing much in, but I'm watching, which is the fractionalized ownership of physical collectibles. Like, would you take a great painting and slice it into a thousand tokens and you could sell the tokens and those tokens are liquid? Or we're staring out at a giant piece of real estate. It's not really a collectible, but there are people who would argue that we should fractionalize real estate and other large asset ownership and have them trade on exchanges. So it's a little bit less about collecting, but it is about fractionalized ownership, which is also made possible by crypto. So this is really big if it happens and we're watching it.

So my last question before we get into the rapid fire to wrap up is, is it happening, right? Kind of, I like your way of talking about AI, crypto, et cetera. It's like, we're looking for these waves, right? There's early data points that say, hey, something interesting is happening. A lot of smart people, a lot of capital going into this. How do you evaluate for those that ask like, hey is this real is this happening is it a wave like what's your answer yeah so i've recently changed my mind a little bit on this oh interesting um i was of the mind saying yeah the i need to see a killer app be built by decentralized app developers i'm looking for that moment in the way that there was a killer iphone app that proved that mobile was a thing and then it then it'll

convince me because this trading stuff is speculation. And I've been convinced that that is the wrong way to look at it by Tom Lee from Fundstrat, who he says, I'm going to get the numbers wrong because I haven't looked at it in a couple of weeks, but I believe he says a bit, I'm going to, I may have them inversed, but for every barrel of oil that is used in, consumed in transportation so like burned right it's traded 16 times for every one time that it's bought interesting and for every dollar u.s dollar that is used in a transaction it's traded 30 times wow now i may have them reversed i think the numbers are 16 and 30 i'm sorry tom if i'm getting it wrong but i am giving you full credit but the whole idea is there's a lot more financial

So speculation is the superset. It is the killer app, actually. And it precedes demonstration of other applications. So now I don't believe that in order for crypto to be big, there needs to be mainstream non-speculation use cases. I think they may come. But I think that speculation is a killer app. And what's nice about that is that is the behavior that's happening right now. Right. And in big numbers, right? I mean, it's very small compared to other asset classes, but millions of people, right? Hundreds of thousands to millions, depending on the chain. So I think that is reoriented a little. And the investors who are smarter than I was, who bet on, let's call it trading infrastructure, custody, wallets, all the things you need to enable speculation are going to do just fine. And not every one of the companies will succeed, but that infrastructure has to exist. If what that means is, you know, 16 to 30x the volume will be in speculation than is in transactions or other uses. And that was a really smart bet of

theirs. So I think the answer is it is here. The question is how much bigger does it get and what other use cases emerge and where do you invest? And so there I'm trying to be a realist. I'm like, show me the money, right? Or just show me evidence. The evidence I see is developer demand, which is what led to the three box investment. You know, they've had more than 20 or 30,000, uh, download and installs of our three box. Um, that's really big. It's really big, right? Might even be higher than that right now. Um, you know, the hundreds a week, sometimes hundreds a day, if it's around East Berlin or East Denver and, uh, and they are clearly dominating this use case. So that's an early warning indicator that there are devs who

are building apps, dApps and need to solve decentralized data. So that's, those are the type of indicators we're looking for got it um rapid fire questions what uh what's the most important company in all of crypto in your opinion most important or um i mean i just admire coinbase uh they're just run so well uh they're solving a real customer problem they do it elegantly they're very user-centric i'm not of the crypto mind that you could never have someone manage your keys, centralize, although I appreciate why you may not want to, but I think they're really a company to watch, just their execution is awesome. So exchanges are by far the most popular answer to that question. And I don't yet know what

to make of when that popular answer changes, right? If people start saying something other than exchanges, does that mean the way people are interacting with the assets is changing? Is it signal something in the market or whatever? But I've been asking that question for almost two years now. Everyone says Binance or Coinbase. exchanges. And so it'll be interesting when it changes. I think it'll tell us something. That's, yeah. Call me the day that happens. Absolutely. What's the one law or regulation that you would change or improve kind of related to

crypto if you could? Yeah, there's a lot. But the one I would change right away is that companies who follow the sort of Reg D approach to effectively listing a coin should be allowed to have them trade on exchanges in the US. And that's not possible yet. So basically just keep the Reg D rules, but change what happens if you follow those rules. At very least, you should, if you went Reg D, you should be rewarded by your token being able to be liquid. And they're not letting that happen in the US. But I would change so many other things about the way the SEC has approached this. For sure. What do you think your most controversial thought in crypto is? So something that you believe is true,

but most people will disagree with you on. Well, what's I think actually still, if you look broadly at tech, being pro crypto is controversial. Right. Right. Like believing that this really is something big is non consensus today. Yeah. It almost feels like pre like second half 17. A lot of people were just like, I don't know what this thing is, whatever. Then second half 17 also be like, wait a minute. Like, did I miss it? Right. Or am I making a mistake? And then it was like they all got suckered. Right. and the price crashes and now everyone's like oh that thing's written off again yeah so i think the hard question to answer right so an lp asked me what is the advantage of building a dap what was your answer well so like the because in 1995 if you said well what's the advantage of doing things on the web like why it's like let me count the ways man i mean like you get it instantly and there's no middleman and right low cost of publishing and like you know no trucks

better the environment like i it was too many answers um why should you use adapt because it's slower and right and it's super super clunky and the thing you got to do with your moat with the um with the wallet is terrible right yeah like not good answers so but the the answer is a developer friendly answer today right it's because you can trust the platform they won't revoke it later but that's not a that won't work that doesn't answer the pew question of why are you using adapt and yeah but i'll push back on that right like you still had dial-up internet right so like hey the newspaper gets delivered to my door or like i can take up my phone line and dial up to the internet to read the news like there's some comparisons i don't know if it's quite as

clear but but i think there was still some user experience stuff that that wasn't as good yeah so so in the answer of um why will you tolerate the inconveniences there still has to be a good thing you're getting out of it. And I think with crypto collectibles, it's like, well, I have a digital good that is going up in value or may not, but I can use it in games and I'm trading it and collecting it. So there's a real answer like for why did you tolerate the MetaMask brain deadness to buy your CryptoKitty? And I think Dapper wants to make sure there is no brain deadness when you buy your NBA moments. So I think we'll have a good answer then. But why would you, you wouldn't build a word processor today right decentralized and the consumer would not walk away from that

experience and say well that's better than using word or google docs right it's not it's slower and and the benefit wouldn't be clear but when you can say google doesn't have my data that's what that was my benefit like i now i don't have to worry about my photos leaking to someone else because they're encrypted and decentralized i think that will be a user benefit yeah makes sense? What's the most important book you've ever read? There's so many. Um, probably sapiens really helped me think longer term, um, that there's somewhat of an inevitable arc, uh, enlightenment now by Pinker made me get a little bit less depressed that, um, you know, that, that the arc of justice bends, you know, the arc of history bends towards justice and that like

things are getting better if you look over long term. In the last year, I've just switched from reading. I'm basically only reading textbooks, science textbooks. I'm just so annoyed by the fact that I can't trust any source anymore because I'm being manipulated by everything that I'm just going to science textbooks. I've been reading about electromagnetic, you know, rf actually the problem we were talking about earlier about why does uh do we get bad cell coverage here on the 54th floor yep i've been reading about that really yeah what uh what's the main takeaway like like as you consume what i'll call more dense scientific content is a non-disputable information well it depends on what's the in uh what is it uh in texas there

was a yeah i'm not reading any texas textbooks no um no but uh is it because you want to learn about that one specific subject and it's just fact it's pleasant to read fact and know that there's no dispute about it and like i'm learning and it's i'm getting smarter about something and i don't have to worry about it's the opinion of someone for sure uh i wrap up by letting you ask me a question but first we talk about aliens and given your uh you're bent on uh the scientific textbooks uh you think they're real you believe um there's no proof of extraterrestrial life thus far. So agreed, but do I think that will change? Yes. Um, well, I'm not sure where I am on the Fermi paradox, but by the numbers, you know, the number of, um, places where life could

exist feels like there should be life out there. Uh, but caveated with the Fermi paradox. So here's the, the two things, uh, that blow my mind. Right. So, um, I recently watched, uh, an explained series on netflix they have about uh basically intelligent life outside of earth right yeah they don't use the term aliens which i think was intentional to make it more uh you know legitimate um they claim that there's for every one grain of sand on earth there are 10 000 planets out there right and just like to even try to you know we really said like 10 000 planets incredible the second thing is uh i think a lot about the comparison between space and the ocean right and um and there's some accounts that i found on instagram and twitter and stuff

where they basically show things in the ocean that you when you look at you can't see right you just don't know what they are right you've never seen this before um and it just reminds me like we know so little about our own oceans let alone what's out there in space and so it comes down to like this mathematical probability but to me the part that um i think has shocked me as i've asked this question so many times is everyone looks up they look out right they want to know what's out there how many people are actually interested in the ocean how many people are interested in what's going on in here in earth and uh there's not that many people and we but we don't we discover like 20 new species a year just by going down in the ocean yeah crazy there's

a guy so uh my one of the podcast episodes i want to record is there's a brazilian billionaire who uh he's built the uh it's a one-man submarine basically right so kind of a capsule uh that is uh built to go to the depths of the ocean and he has visited now i think it's two or three of the ocean's deepest points so he's you know indian ocean whatever uh and you see him he's in this little thing right and you know rich guy so he got all the things he could be doing um but it's fascinating because he's the first person to do it right and you're like look like oceans right there right and uh and to see people kind of just now exploring um is cool and my question for him is going to be why are you in there yeah right you know it's kind of like hey there's a lot of

people like could just not find anyone else or are you actually that personally interested but we'll see okay i get to ask you a question you got it you've done 220 of these something like that yeah 220 221 something like that um what is the one thing you've heard that was the most controversial at the time to you you were like there's no way that's true that's proven to be right well that's proven to be right um all right before i answer that i just got a funny one for you so i don't know if you know uh josh brown from ritz holt uh wealth management um it's a uh basically wealth management like financial advisor he came in here he was maybe one of the first 50 interviews and uh he was very early when i started asking the alien question and he said to me

without a blinking he said uh i believe that ghosts are more real than aliens and i remember thinking like that's blasphemy right yeah but but that always sticks out my mind i would say the thing that somebody has said that i didn't agree with it ended up being true um this is recency bias but uh we had a guy mark schneider come in who a nuclear engineer and one of the things that um he really explained was uh a lot of the causes for the you know chernobyl fukushima etc uh how much of it was uh human error right compared to uh more like technology or just the actual nuclear so kind of as i looked at that that was pretty eye-opening um and then another thing that i think uh has been pretty interesting is um it's not so much

that it was controversial when he said it but uh jim o'shaughnessy um when he came in uh he spent a lot of time talking about just like look i don't know but that's how i look at every market right i'm a quant so like i don't know the answer in stocks i don't know the answer in bitcoin i don't know any of this stuff um and he's really like spent a lot of time both publicly and and helping me see like humans are really really biased right and like we think we're so smart and we know nothing um and uh it was backed up uh i read um two books uh you probably read them uh, super forecasters and then, um, algorithms to live by, I think it's the book. Um, and both of them, they hit on similar things, right. It's just like, we know nothing. Right. And, uh, I

think when you take that perspective at a lot of this, uh, all of a sudden you turn into more like curious. I want to receive the information rather than like, I know how the world's going to move and let me go invest. Bloomberg says in God, we trust everyone else bring data. I've never heard that. That's pretty good. All right. Listen, uh, where can people find you, on the internet, Venrock. Follow me on Twitter at Pacman, P-A-K-M-A-N. DM me, talk to me. Also, find me on the Venrock website.

Reach out anytime. Let me know if you're working on something great. Awesome. I really appreciate you coming to do this. Great honor to be here. This is probably one of the more fun conversations I've had. Hey, that's good. I'm going to tell Paulina you said that.

She will think you're lying. I appreciate it very much. Yeah, thanks for having me here. I appreciate it. Hey, everyone. Pop here. if you like this episode of off the chain and want to help us take crypto to the top of the apple spotify and other podcast charts please do us a favor and rate review and subscribe to review

simply go to the off the chain home page scroll down until you see the five blank stars taking 15 seconds to fill those stars in and leave a quick review goes a long way in helping us take the entire crypto ecosystem to the top of the charts i appreciate you listening and see you next time on off the chain