Introduction

Patrick: [00:00:59] My guests this week are Kyle Samani and Tushar Jain, both managing partners at Multicoin Capital. I've taken a bit of a break from crypto because I hadn't sensed many new angles to explore in this forum, especially from an investors point of view. I felt that while things kept evolving, the major investment thesis had been established and explored. Kyle and Tushar are interesting because of their often divergent views. For example, Kyle has been an outspoken supporter of Ethereum relative to Bitcoin. This conversation, which is meant for those still curious about crypto, offers lots of new food for thought. We discussed smart contract platforms, network effects, the coming platform wars, and why blockchains may not matter at all in 10 years.

Path to $10 Trillion

We'll start at the very broadest level, which is this idea of what the path may be for the crypto, beta, let's say, or total network value to reach five or $10 trillion, something massive that would result in a huge, multiple invested capital for people that are putting money into the crypto world today. So maybe we could start very broadly with that idea, what might be the path from here to there and where, if I don't know how many paths there are you tell me, along each of those paths, what kind of cryptocurrencies would represent the winners?

Kyle: [00:02:50] This is Kyle. So we wrote a post about this a couple of months ago, it's called Paths to Tens of Trillions. And in this post, I kind of lay out among the crypto people, what we kind of debate about as the fundamental hypothesis and the ways by which we think different crypto assets could become worth tens of trillions of dollars. So if you buy into the general thesis of global digital reserve assets and things that governments can't inflate and the kind of peer to peer model, if you buy that, there should be some global digital currency there. And the question is, well, what's the path by which we get there? So one path is what I call the store value hypothesis. This is really espoused by the Bitcoin community. And this basically says that you need very few features in this global digital reserve asset to get to that end state. You basically need just a handful of things. You need it to be very secure.

You need to make sure it's self sovereign. You want to make sure it's censorship resistant. You want to make sure that governments, no matter how hard they try, can't shut it down. And that's kind of basically all the valid hypothesis people say, oh, and they also mandate very, either fixed supply or very predictable, very low inflation schedule. And that's basically the only requirements to the store of value hypothesis. And so Bitcoin really is kind of the purest representation of this hypothesis. And these folks basically say, look, just make sure it works and any other features are fine, but if they come at the expense of any of these core characteristics, then they're probably not worth having, and you can add them in layer two or some other place.

Then the other major hypothesis is what we call the utility hypothesis. And this hypothesis basically states that the cryptocurrency is that it's the most useful, will become the most valuable. And so this is kind of a very Silicon Valley oriented thinking, basically saying, hey, just get people to use the thing, solve problems, get people using it. And as they use it, obviously they'll buy it. And at some point the market will kind of crown whatever the thing that's being the most used as the global digital reserve assets. There's some sub hypotheses within there, but those are kind of the two large buckets.

Token Velocity

Patrick: [00:04:37] Let's talk about the second one, because I think that the Bitcoin one is fairly straightforward and it just so happens to be transparent. That's the only one that I'm not convinced of it, but it's the only one that sinks with my mind after spending a ton of time in the crypto world. So it's fairly straightforward. The features seem to be there already. Maybe that's why it's easy to understand. The second one is very different. So let's talk about things like token velocity, things that matter for how much value will be held in the tokens themselves. And it would be helpful for you to use maybe a few examples of other cryptocurrencies that people might be familiar with, out there, some of the bigger names and talk about what you mean by use and what the various types of use might be today, given the options we have to buy as investors.

Tushar: [00:05:19] Yeah, for sure. So this is Tushar. And in order to make this debate easier, let's just abstract away some of the details. So let's just frame it as Bitcoin versus Ethereum. So Bitcoin is the store value hypothesis and Ethereum is this utility hypothesis. And so I don't know if you've read John Pfeiffer's paper, Institutional Investors' Take on Crypto Assets, where he basically makes the claim that the velocity of Ethereum or ether will be extremely high, and Ethereum does not capture any value. All of the value lives in Bitcoin, even if the Ethereum ecosystem is a thriving ecosystem, which has a large number of users and DApps and such built on top of it.

Patrick: [00:06:00] Just as a quick pause, can you just describe what velocity is and why high velocity means that?

Tushar: [00:06:05] Yes. So there is a fundamental macroeconomic equation, MV equals PQ, basically, monetary base times philosophy equals GDP. PQ is price times quantity. And the idea is if you have a certain monetary base and it turns over V number of times every year, then that is the total amount of economic activity that has happened in your economy, so that is your PQ. So if you have a higher velocity asset, then you need less money to accomplish the same GDP, basically. So when you now, let's abstract that away from the macroeconomic view to something in crypto, let's say, if you have a tremendous amount of GDP quote, unquote, happening in the Ethereum ecosystem, if the velocity of ether is really high, then you need ether to be less valuable in order to still facilitate that same economy.

Patrick: [00:06:56] Yeah. Okay, great. Keep going.

Tushar: [00:06:58] So those are really the two fundamental debating camps Kyle laid out, but we can't say for sure that there is one clear winner here. To look at the world probabilistically, and I think it's path-dependent. So one heuristic that I think really helps the utility side of this argument is to look at money, not as a noun, but as an adjective. Look at the moneyness of certain assets. Some assets are more money than other assets. It's hard for us living in the United States with a pretty stable currency, the US dollar, to really consider this, but think about this in all our nations, or think about this pre US dollar, maybe gold is more money than silver. So what increases the moneyness of an asset? What makes something more money? Because what we're seeing right now is this free-market competition in crypto for what will be money or what multiple things will become money.

So when thinking about what makes things more money, there's a hypothesis that things become money by being used as money. And when you look at the historical examples, let's just look at the US dollar again, and Nixon removing it from the gold standard. The reason why that worked, in my opinion, is because people were already used to using the US dollar as money. No one was actually really relying on the fact that it was convertible to gold to think that it had value, they were already using it and so therefore it was a very high moneyness asset and so that transition was able to be pulled off. So I actually think the US dollar falls into the utility hypothesis, not the store of value hypothesis, and so do other major currencies. So that's one major point in favor.

Patrick: [00:08:49] When we first met, one of the things that was really interesting is that you were the first person that was, I don't want to say use the word negative because you're probably positive on cryptocurrencies in general, but relatively speaking, negative on Bitcoin and very positive on Ethereum. This was, I don't know if it was a year ago, but it was a while back. So maybe talk through that idea as being a bit of a contrarian on that relative to the broader community.

Kyle: [00:09:09] Yeah. So I am intellectually short Bitcoin. Our fund is generally long crypto. We were short, I think at various points this year, but we are generally long Bitcoin. I believe with a high degree probability Bitcoin will break 50K at some point just as kind of the market evolves and just assets flow into the ecosystem. My reason for being intellectually short Bitcoin is basically this utility hypothesis versus store value hypothesis in which I think on a long enough time scale, at some point, the bubble will pop. And you asked about when we were down 60, 70% from top right now, but the market is still very much thriving, the amount of activity and energy is unprecedented. I think there will be a much larger bubble happen at some point as the real asset global asset flows into crypto.

And at some point when expectations are not met, at some point, people are going to look around and say, well, at least this one over here, these three are being used every single day and people are doing stuff with that. And I believe that will create a floor, price floor, ultimately that will kind of allow the most useful assets to capture the most long-term value.

Cryptocurrency Use Cases

Patrick: [00:10:06] So let's talk in some detail about uses. So this is the thing that always frustrates me, is that that makes a ton of sense but when you actually go investigate who's using Ethereum or whatever other platform to do stuff, the number is very small. The number of active users is tiny and it's been very difficult to scale any sort of interesting use case. So talk about the challenges there, where you see that evolving, will it be something other than Ethereum that captures that marketplace utility idea? What are your thoughts there?

Kyle: [00:10:32] So Ethereum was the first smart contract platform and very much kind of represents, I think, the utility hypothesis, but what we're seeing happen right now is there's just a wave of Ethereum competitor showing up. And these competitors all have fundamentally different ideas about the types of trade-offs they should make in the design of the system. And those trade-offs have pretty material impacts on end user usage. So say, I'll talk a little bit about EOS. There are a lot of other Ethereum competitors that we are invested in Dfinity Tezos, Kadana, Solano, a bunch of them. We think they're all interesting for various reasons, but let's just talk about EOS because A, it basically is launching today and B, it kind of represents within the utility app offices, the extreme away from Ethereum. So Ethereum architecturally was very much constructed in the model of Bitcoin, which is to say that it's focused on being as decentralized as possible.

So having as many nodes as possible, recalculate the exact same computations as a guarantee, and that makes the system trustless. Well, this is extremely technically inefficient and as a result, the throughput of Ethereum is quite low. It's relatively well-known that Ethereum can do 15, 20 transactions a second today. And so if you're going to have any sort of global usable application, that's kind of a joke, but actually there's a second performance characteristic with Ethereum that's also extremely limiting and that's latency. So in the Ethereum system today, every block is about 15 seconds, which means for all practical purposes, if you're going to have any sort of application that's dynamic and engaging and you click stuff and things move and whatever, and you refresh the screen and do stuff, you have a 15 second latency right now and doing that in a meaningful way, and that could be longer.

And consumers in 2018 will not accept that, period. We're used to LTE and everything. And so Ethereum has very novel ambitions, but the current architecture of the system really lends itself more towards being a store value that just happens to have the smart contract thing on top of it. And when we look at EOS as a counterpoint, EOS is saying, look, our goal is to make applications as usable as possible. So the system is a little bit more centralized in Ethereum, that centralization allows for both more transaction throughput and reduce latency to make the system actually meaningful, usable. And on top of that, the EOS development team is really focused on a lot of other, I'd say pragmatic features to make the system more accessible to end-users. So one of the key ones is the fact that the system imposes no transaction cost on end-users, which is a very cool concept.

Tushar: [00:12:40] And EOS isn't the only platform that is competing with Ethereum, but to go back to why are there so many Ethereum competitors, what is the possible utility of these? What is being built? Because actually like if you look at Ethereum today, I think there are maybe four DApps with over 1000 daily active users. I mean, that's nothing. 1000 daily active users is, it's so small. But this brings me back to a couple of anecdotes that are just helpful anecdotes, aren't data, shouldn't just be used exclusively for decision-making. But helpful to think about. One is, we have this huge design space that is being opened up by the smart contract platforms, by some of the features that Kyle just described, by just some of the broader things that are uniquely enabled by crypto assets, like scarce assets, et cetera.

And this design space is going to enable humans to express new creative outlets and create things that other people want. So I'll give you an example. I was a kid in the 90s and I remember growing up, we had two rules that were really important. One was, don't meet people from the internet in real life because you don't know who they could be. And two was, never get in a stranger's car. And when I leave here today, I'm going to summon someone from the internet to get into their car. It's fundamentally impossible to project out how people will design in an unbounded new design space like the internet. That was uniquely enabled by the internet. It never would have been possible to have something like Uber or Lyft without the internet. But it's just impossible to think of that. Another anecdote that I'll give you is, I think there was a quote from Paul Krugman.

I forget exactly when he said this, but he said something about how the internet isn't going to be worth very much because eventually, people will realize that they don't have all that much to say to each other. So once again, this enormous new design space of the internet that was created really enabled people to create things that they didn't even know they wanted yet. And I think that that is what we're going to see with these smart contracts platforms.

Patrick: [00:14:44] So talk more about why this is a new design space. Is it the ability to exchange value seamlessly? What is the major breakthrough that cryptocurrencies represent that creates some new space that doesn't exist in just the normal internet?

Tushar: [00:14:57] So one thing that I think is really novel here is the removal of the central party. The reason why human beings are the dominant species on the planet is because we are able to coordinate in large numbers very flexibly. And what crypto assets enable is more people to coordinate in this large flexible way than was ever possible before. We are able to coordinate flexibly in large numbers because we can believe in stories. And when you look at almost every institution out there, everything is a story, it doesn't physically exist. It's just something we tell ourselves in order to be able to coordinate. O'Shaughnessy Asset Management is a story that enables all of your clients and your employees to coordinate, to create value. Same thing with Multicoin Capital, it's just a story. But these stories have these middlemen like us and yourself that really are the ones driving that story forward here.

Here, what's opening up about this design space is that you can coordinate all of these various actors in a structured way. You can create a smart contract that actually defines a structure that can be amended, but without someone in the middle dictating those terms. So it's kind of this emergent structure, in a sense. And that's something that has never existed before. And if we look back in history as to the last time this type of coordination design space was really opened up, the most recent thing that I can think of is the advent of the publicly traded corporation. Because before that, businesses couldn't get all that large, right?

It was people owned businesses and they were passed down in a hereditary manner. And you couldn't really give out shares in that business to employees and have liquidity and have a whole structure around it. But as soon as they invented publicly traded corporations and had liquidity for that, it enabled this whole new design space for business, basically where you could create new businesses that didn't exist before. Same thing with like the limited liability as a concept. So here we have this new concept for businesses.

Kyle: [00:17:01] I'll follow it up and give two concrete examples. So there's a platform, it's called Augur. It's somewhat well known in the crypto space. It's launching in the next couple of weeks here. So Augur is a prediction market. And so the most simple example of this is, think about Vegas, go bet on a basketball game. I think tonight the Cavs and The Warriors are playing. So if you want to go bet in Vegas today, there's not really a problem with that other than the fact that Vegas takes about 10% or so for mediating the bet, also just regulation and accessibility and a bunch of other problems. So Augur has created a decentralized prediction market. So the idea is like they have a smart contract, basically you can imagine you create a contract that says Warriors versus Cavs, two people. There's an order book somewhere that sets the odds, people commit and buy, right size of the market.

And now the money's locked up in a contract that at some point after the game, the system needs to resolve the bets and pay up the winners. So it was pretty simple concept. Within the world of blockchains, blockchains themselves are not able to go query outside data sources and to get access to them. It's kind of a technical limitation. So the question is somebody has to go to nba.com after the game, get the score and send the score to the contract. Well, let's assume you can trust the person to do that, the problem is like let's say nba.com is compromised or has had the person whose job it is, is to kind of read it nba.com and send it over is compromised. You need some sort of fallback system in this model to make the system work. And how do you do this in a globally trustless way with no central party whose job it is, is to flip the switch after the game is over.

So Augur has solved this problem. So they've created basically a way that they have their Augur tokens and then bets are not the nominated in Augur tokens, bets are nominated in ether or something else. So the idea is if there's a dispute after some event happens, this system will get some subset of Augur token holders, something like 30 or 50 of them anywhere in the world and basically say, hey, it is your job to go and mediate this dispute, go Google the answer and vote correctly. And it turns out if you end up voting in the minority, on the losing side of the bet, then you actually get some of your tokens, get programmatically burned or taken away from you. And so the idea is basically as long as you're voting the truth and everyone else is voting the truth, you'll get to the right outcome.

There's a couple levels of appeal in the system and you get to increasing the larger numbers of people and the final, the Supreme court, so to speak, is like all the Augur token holders anywhere in the world. And the idea here is, is if you were going to lie about some sort of event, you'd have to have global coordination among millions of people, and it just wouldn't work. My point is, is coming back to your original question, what kinds of new interaction models can you enable with crypto that were not possible for, and Augur is a very fantastic example of this, where you can now have literally let the world arbitrate what happened and then create any prediction market for anything out there. I think this is going to be really... I mean, the first very simple examples are like online gambling and making that more accessible, politics, elections, whatever, but you can imagine a lot more interesting types of new prediction markets that really aren't available today.

So things like predicting flood markets and insurance things in various risk zones, all around the world. The Augur team has published all kinds of really interesting research of people who've come to them and said, we want to create these kinds of interesting new prediction markets to hedge risk and all kinds of crazy avenues. Where we are able to say like, what's the new design space? Augur's a really good example. That's very pragmatic. Augur's launching in the next week or two or three, it's right around the corner. I'll give you now something that's a little bit more hypothetical and it makes a ton of sense. So basically think of a decentralized Facebook or decentralized just social stream. So if you think about Facebook today, or you can upload your photos, your status, your locations, you go X places, X, Y, and Z, whatever.

But like all this stuff lives in Facebook, Facebook does have some APIs and they say within limited access and go read things, look at the newsfeed and different apps can plug stuff in the Facebook. But the actual Facebook database is extremely locked down and the number of APIs and ways that you can read or write from the database, are by definition, limited based on what the Facebook teams say the APIs can and can't do for both read and write access. So if you kind of took this concept of basically just think of it as a social stream and you put it on a blockchain, well, now there's no one controlling read and write access. Obviously you need to kind of deal with identity and making sure it's your photos and you can control privacy. Those are complicated questions, we'll come back to those in a bit, but just think about the ability for developers to read and write to a single social stream.

While imagine you could have Snapchat, or you could have Instagram, you can have all these different apps reading and writing to the same set of images and then doing different kinds of filters and then different kinds of games on top of those. And you could challenge your friends to games where you draw something funny on their face. I don't really know where this goes, but I think there's a lot to be unlocked from having this concept of a shared database where you have many different applications with all kinds of crazy developers, reading and writing from the same underlying data structure. I think that's a very, very powerful concept. This has basically never existed before, and I think there's going to be all kinds of really interesting ways people leverage this for gaming and for social, for certain. And we'll see where that goes from there, but there's going to be a lot of really cool emergent things.

Patrick: [00:21:31] So I really liked this idea of design space and things that we kind of can't foresee and that some or one or several of these platforms I'd actually be curious to hear. If you think it's going to be a winner, take all scenario or not on the utility side, we'll just create a ton of value in the world. And there's value creation, there's value capture. I've been talking about this idea a lot lately as it pertains to investing specifically. So the value capture portion of this becomes extremely important as an investor if I want to buy cryptocurrencies on the premise that you guys are right, or that I can somehow pick, let's say, EOS as the winner, and I can pick that ahead of time. Walk me through then the next step. How do I know that value will then accrue to, and maybe, I don't know, you said probabilistically a lot, so maybe I don't know and it's just a bet.

But how do I gain confidence that it's a good investing decision beyond just something that's a really neat technology and a great way to create marketplaces where the value just, it's like consumer surplus, like the value just accrues to the people selling stuff that other people want and the buyers of that thing, because it's a more seamless, effective capital coordination exercise.

Tushar: [00:22:30] One of the things you're asking is... okay, great. We have this new design space. We have all these smart contract platforms. We create a bunch of stuff that people really want to use, but why is that valuable? Why does it actually deserve to be invested in, or why will it drive a return for investors? And I think in order to answer that question, we need to zoom out and not look at just the people who are already in crypto, but look at the 99% of people in the world or higher that are not in crypto yet. So let's just go down a series of assumptions and agree or disagree on those or debate those and then we'll get to a final point. So the first point is that I think there will be a global digital money or monies that are not controlled by governments, that are powered using some sort of blockchain type technology.

So then the second assumption there is that things become money by being used as money. The reason why this green piece of paper in my wallet right now is valuable, is because I can give it to you and you would provide me a good or service in return, or I can go down the street and literally anyone in this entire city or country, or probably around the world would give me a good or service in return because they also believe that it is valuable. So then the final assumption is that there are economic models that are designed for these smart contract platforms to make their base layer tokens, money-like objects, or to increase the moneyness of their base layer tokens. So the model that Ethereum uses is that ether is used to pay for gas on the network. So if I ever want to interact with any sort of DApp that lives on the Ethereum ecosystem, I have to use ether to pay gas.

And so therefore I am using it as money, I'm using it to pay for a good or service effectively. And so in my opinion, that reinforces the idea of Ether being money in people. As I think that most people will first interact with crypto, not through Bitcoin, but through some DApp like Augur or some social media DApp, something like that, or EOS has another economic model that is very different from Ethereum where your ownership of EOS tokens is a claim on the resources of the network. So if you own, let's say, 0.5% of all EOS tokens, then you have a claim on 0.5% of all transaction throughput or memory, et cetera, in the EOS ecosystem. And so if you want to play a game and it's going to take up to 0.5%, you're good to go. That's another economic model.

There are other economic models out there too among smart contract platforms, those are the two best that I've seen so far. And so when you go down this list of assumptions, that there will be a global digital money or monies that things become money by being used as money, because it reinforces a story. And that the developers of these smart contract platforms are specifically designing economic incentives and economic designs into the base layer protocol in order to increase the moneyness of their token.

Different Types of Network Effects

Patrick: [00:25:26] Great excuse to talk about the different types of network effects that you guys have just recently written about. I think published an article fairly recently on the different, I think there are three different types of network effects and how you think about those as it pertains to the value of the underlying cryptocurrency. So this sinks kind of nicely with everything you just said. So maybe describe those three levels and how people maybe are getting... network effect is one of these phrases that has become so ubiquitous as to be like almost meaningless. So maybe describe the actual math behind this and how you think about it.

Kyle: [00:25:55] Yeah. So our post a couple weeks ago, I think it's on the network effects of stores of value. And in the first part of those are going to define what our network effects, and there are a few different types of network effects. For the purposes of this, I'll skip the types of network effects and let's just assume that they are in fact a thing. The next question then is, well, how strong are those network effects? And there are a few different kinds of models for how strong those network effects are. The most commonly cited thing is the thing called the Metcalf's Law. Bob Metcalf is a guy who invented basically ethernets and Bob Metcalf stated, I think in '82, '83 or something like that, he was referring specifically to telecommunications networks and he said, the value of the network is the square of the number of participants in the network.

He then very quickly caveated that and said that assumes that all connections in the network are worth the same amount, but that kind of has been... people have forgotten about the caveats and basically run with it and said, the value of all networks is N squared. And you can also think about it as maybe the defensibility characteristic of the network is N squared, N being the number of users. And that is basically false in every capacity. It is theoretically an upper bound, but it's so far from reality that I think it's kind of meaningless. Basically nothing grows quadratically forever because you get to be too big, too fast. Instead of what we see happen is in practice using networks, that there are some marginal benefits accruing visional users to the system, but each additional user to the system is not in fact worth the same amount.

So more reasonable proposals put forward by people basically state that the network effects of most types of call telecommunications networks is probably something closer to N log N. And so if you think about that, that curve is super linear, meaning it is always accelerating upwards faster. The rate of acceleration is always slowing down, but the curve is actually always bent upwards. Over time, this is pretty good for describing the earliest stages of network, and this explains why it's hard for people to compete with something like let's say, Facebook or with Google or Uber or whatever. But if you kind of then look a little bit further down the line, you realize, for example, well, Uber and Lyft both work quite well. And the reason it turns out is people don't care if the wait time is under two minutes. And if you look at the global texting market where we have WhatsApp, we have Telegram, we have Kik, we have... whatever, there's a bunch of these, the iMessage, right?

And clearly there wasn't one super big winner here. There were lots of big winners, but that very clearly tells you that if the curve was in fact sloping upwards in perpetuity, that you couldn't have seven substantial global players, that wouldn't happen. And so these curves have to start kind of tapping out at some point, like the value of each marginal user has to start decelerating at some point. And so basically we kind of modeled that out as an S-curve and I don't put any particularly strong bounds on how steep those S-curves rise or how quickly the flat part is at the beginning. But basically is to realize that these things are in fact, S-curves.

I think that's one key thing to understand about when you talk about network effects that are derived from numbers of users using using the system and some sort of interaction mechanism, whether it's Uber and Lyft or Apple and apps and users, or calling the car, whatever. The kind of fourth then interesting type of network effect to think about is, well, what's the reflective liquidity, or because you think about using, let's say gold or something, some other store value asset, what does the network effect?

And so if you think about, let's say you have a bar of gold in your vaults, and at some point you want to go buy a house or buy a car or whatever the only thing that really matters is you need to be able to sell the bar of gold for whatever the market price is. And so that is basically a function of liquidity. So let's say you have 0.1% of all of the gold in the world, if you're going to go sell that on the market, the question is, well, basically what's the liquidity as a percentage of... what percentage of gold turns over every single day? And if the numbers 0.1%, then okay, it's going to take a little while to sell your gold without too much slippage. If that number is 5%, then you can sell your gold probably in an hour more or less and it's good to go.

I mean, if you get the 10 or 20% at that point, the marginal gains are basically nonexistent because it just like, there's so much liquidity in the market that you can't... doesn't really matter. You can just market sell at that point. And so my point is, if you think about this curve from very small numbers to very large numbers, the rate of marginal utility of liquidity very quickly drops off. I mean, it doesn't even... I don't think, it never turns exponential at all. I think this curve is actually logarithmic from inception. So it's not N log N, I think it's just log N. And so there's value in having some liquidity early, very clearly, but the value of each marginal dollar of liquidity very quickly falls off.

And this actually makes a lot of sense. If you look at the crypto exchange market today, the current state of the market very much implies that this is true, that there are so many crypto exchanges. And the reason that it'll work is like a crypto exchange that does $100 million of Bitcoin volume today, well, if you're only selling half a million dollars of Bitcoin, do you care if there's $100 million of volume or $1 billion of volume? It turns out the answer doesn't really matter. And so we already see this happening quite pragmatically in the crypto exchange market. And so bringing this full circle to Bitcoin and the whole crypto ecosystem, one of our views is that if Bitcoin becomes digital gold, which is it currently is on the path to becoming digital gold, our view is that network effect of digital gold is not every strong, it's basically network effect is more or less log N.

And if Bitcoin becomes a digital cash, then network effect is probably something closer to an S-curve and you having something that's exponential even for the first half of the curve is a lot more valuable than something as logarithmic from inception. And so given the still less than 1% of the world's population is using crypto, if you look at the very left most part of all of these curves, they all kind of look the same, but as you grow out in the future, what clearly matters in the long run is having the strongest network effect. And our expectation is that something that has the model that's used as currency, where the network effect looks something more like a telecommunications network, we expect that in the long run, will be the most defensible and ultimately engender the strongest network effect.

Censorship Resistant Store of Value

Patrick: [00:31:19] Really interesting. So there's two layers that I'd like to get into of the story that you tell to two different groups of investors. One, let's call it institutional investors that are not convinced on crypto as a whole. And then the second layer will be people that are convinced that crypto is worth investing in, but are making a decision between you and some other asset manager, right? So the differential strategy that you think is interesting. And we'll start with the highest layer. So when you're talking to somebody that is smart and maybe familiar enough with what's been going on, but has not yet put any money into the crypto ecosystem as an investment beyond the path to tens of trillions that we talked about at the very beginning, what else do you talk about? What are the key points to try to convince somebody of that type, that it is worth a harder look or even their dollars?

Tushar: [00:32:05] I think the first thing that we share with those investors that is easiest to kind of absorb is, do you think there is a market for unseizable, uncensorable wealth? And the answer is invariably yes, because it exists already. Like we saw the Panama Papers recently, we saw... and I think like over the past 10 years, there's been a lot of crackdowns on Swiss bank accounts, but that had existed for a very long time as an unseasonable, uncensorable store of value or store of wealth. And this is an analogy that I'll borrow from Ari Paul from Block Tower, where it's a Swiss bank account in your pocket. And a lot of investors kind of really click with that instantly then after that, we start talking about, well, what asset is going to be in that Swiss bank account? Is it going to be Bitcoin? Is it going to be ether? Is it going to be EOS? Is it going to be Monero?

Even if you don't think it's certain that a crypto asset will win this use case, if you think that this use case has real demand, and you think that there is some chance that it will, then it deserves that level of allocation in your portfolio, if you were being entirely rational. So with Bitcoin and the thesis there of just digital gold, what ends up happening or digital store of value, what ends up happening is it's recursive up and recursive down. On the way up, it's a great store of value. Everyone's really happy. And it actually reinforces the idea of it being a global censorship resistant store of value that people want to buy into. And so therefore feeds into these extremely volatile moves, but then on the way down, it's a really shitty store of value if it can actually retain any value.

So that is a bit more sensitive to the entry price, but what we really like to focus on is the utility aspect and how many people are actually using this thing. The current prices of all these assets, just the total market cap is so small compared to the total addressable market here that yes, enterprise does matter because if you get in at 10,000 versus 20,000, let's say, your returns are half. So yes, entry price doesn't matter, but trying to time the market is not an easy exercise as you're well aware. So if you think that the utility hypothesis makes sense, if you think that people will use this stuff and that the economic design of these assets will have it capture value, then averaging in over time is the best thing for an investor who's not full-time.

Crypto Evaluation Framework

Patrick: [00:34:32] So I'd love now to get into the actual process of evaluating specific options for those six. So getting into the reasons why those six or the six. Maybe you could begin by describing a list of, or a set of dimensions that you care about when first approaching a new project, a new token, a new blockchain, and then how you sort of weigh those things as you're making investment decisions. So you've mentioned some of these ideas already, things like latency or throughput or censorship resistance, those are some examples of what I mean, but what's sort of an inventory of when you're first approaching something, what you care about?

Tushar: [00:35:07] This is a new type of asset class and it has these new rules that really change how decisions should be made. One of these new rules in this asset class is that everything is open-source, and that profoundly changes investing because what that means is none of your features matter. Not at all. Everything can really be copied. It's not trivial to copy. I do not mean to trivialize. It's not like command-C, command-V oh, we have all your features now, but it means that the architecture is out there, everything is out there and it is much easier to copy in an open-source world than in a closed source world. So what we really look for are trade-offs, where there are certain things that cannot be copied no matter what. So let's go back to the Ethereum EOS example, because we've been talking about that a lot here and it's on both of our minds.

So with EOS, they have chosen a different set of trade-offs than Ethereum chose. Let's just choose one dimension of those trade-offs for simplicity and let's say we have centralization versus decentralization and centralization enables scalability or more transaction throughput versus decentralization, specifically of block production. So if you want to look at the most decentralized system, that's going to be something like Bitcoin or Monero where it does very few transactions, but that's because every node on the network has to validate every block. And then you have the very centralized version of the world. This is like your visa or PayPal, et cetera, where it's really a centralized company that runs the whole thing. So we think value will definitely accrue on the very far end where you need to be totally decentralized. There will be some value accruing there. We see that happening already. But is there somewhere else along this trade-off space, that value will also accrue, right?

And so we see EOS taking up a spot somewhere in the middle, where they have a limited number of block producers, but it's not completely centralized, and the block producers can be replaced. Versus Ethereum, where every node still needs to validate every transaction. And that is something that fundamentally cannot be copied. That is a political or a design decision, a philosophical decision, that if you go to people at the Ethereum foundation and you ask them, they can bring over some of the features that Kyle mentioned earlier, like the name service. So you don't have to type in a 40 character Ethereum address that's impossible for human beings to really read, and they can bring over account recovery features that EOS has built in. That's doable. But changing that trade-off is not something that will really be considered seriously.

So now let's add on some more complexity. It's not just this one dimension of trade-offs between centralization and decentralization or scalability and not very scalable systems, there are many dimensions. Another dimension that can be interesting is how expressive is the language that can be used to control these assets. So you have something like Bitcoin with a very limited language. You can't really write very sophisticated smart contracts on Bitcoin. Funny story, Vitalik originally tried building Ethereum on Bitcoin and was told no, because of this specific trade off choice where they wanted to have very limited functionality to limit the attack surface on Bitcoin.

You can't do complex things, so therefore your money is safer. Gold works because it kind of just sits there, you can't do stuff with it. And then on the other end of that spectrum, you have a full turn and complete language that you can build anything you want and it's impossible to validate how the whole system will interact. And this once again is not binary. There are places along the middle of this spectrum. So what we do is we have identified about a dozen of these various trade-offs right now that we think really matter, and then we think about where along this right now, 12, but N-dimensional trade-off space will value really accrue?

So yes, there is some value accruing in the completely decentralized, very inflexible programming language and extremely secure places in this trade off space, but that's not the only place that value could accrue or will accrue. And so our model that we use is then trying to identify where in that trade-off space does a particular asset fit. And then what do we think the probability is that that spot in that trade-off space will capture any value. And then finally, and most importantly, look at how was the market pricing this, and how is that different from how we would price it? And if there is a difference, then we will invest.

Patrick: [00:39:32] It seems like a key in all of this. If you guys are right about whatever platform it is that designs the platform which is most attractive to people to build on top of, that two things matter. Tools that are easy for that group to use, like developer tools, or just user tools, producer tools, and just old-school marketing, somebody that's better at making the name of their cryptocurrency ELs or whatever else known to people like me or just the broader community out there. So how do you think about those two ideas of producer tools and just good old fashioned storytelling and marketing as an item that you're looking for when considering an investment?

Kyle: [00:40:08] We think go to market is super important and generally not valued by the diehards in the crypto space. Tushar and I both kind of come from the background of enterprise applications, so both of us previously started the companies and the enterprise software space building and selling these things to large complex bureaucracies. My favorite blog post of all time is actually... you would think a technical kind of given my public persona and what I like to think about, but actually my favorite blog post of all time is a blog post called, If SaaS Sells Itself, Why Do We Need Sales by a guy named Mark Cranney. He used to work for Ben Horwitz at Opsware and basically there's a key line in the post and he says, the ultimate realization is that sales creates value in the mind of the customer. And when that dawned on me, it's kind of the same thing as Angus, perception is reality.

It's kind of the other side of the same coin, but when that really dawned on me, I realized that's actually how you drive stuff in the world. And engineers like to think, well, it's like I have to build it and they will come, but actually you have to get that just in front of the customer, but you have to drill it into the customer's head, why is this thing useful? And so we spend a lot of time talking with early-stage teams and understanding how do you bring this thing to market? And depending on what you're building, there are different ways that it makes sense to do that in different verticals, and you might want to start with different early pads, different areas of focus, so to speak, and then broaden over time. So we think it's super important, how do you craft the story and tell a story that will resonate with people? And we very much value teams that are more well-rounded and not just hardcore tech people.

There is a place for certain hard core tech investments, especially if their only audience is tech people. But if your audience, especially is general purpose consumers, then you need to be thinking about how do you hire amazing sales and marketing people, et cetera. So we definitely think about that a lot, and it's certainly a key part of our investment criteria.

Opposing Views on Bitcoin

Patrick: [00:41:42] What do you guys disagree on most typically? Or right now.

Tushar: [00:41:48] Probably Bitcoin.

Patrick: [00:41:50] Can you describe how and why?

Tushar: [00:41:52] I'm generally more bullish on Bitcoin, specifically lightning network than Kyle is. This is actually one of the values within the firm is that when we started, we knew that our job is to make decisions. That is literally why we exist. So one of the best things for a firm that makes decisions is to be able to argue and debate well, and productively. We encourage that extensively within our team, and we will consider opinions and such from everybody within the team, as long as they're well backed up, et cetera, and really make sure that we are not subject to certain biases, like confirmation bias, et cetera. So we will open up debates like Bitcoin over and over again. Now specifically on Bitcoin, where we have somewhat of a disagreement is where I think network is really compelling, and you know what, it's a really fascinating technology that enables effectively low or no fee Bitcoin transactions.

And layer two could be... there is some probability that layer two could be the solution to scaling that lets Bitcoin really become usable for the vast majority of people, because right now it's not. The $50 transaction fees on Bitcoin that we saw in December, that is a default, it's a liveliness attack, or it was unusable, it's as if the system is down. And it's almost like if every time I go to Facebook, I have to wait a minute in order to use it. It's the same kind of thing. So how do we solve that problem? And I think lightning is really compelling. There are obvious technical problems there in terms of how do you route money and how do you open channels that have enough money in them, et cetera. But I think that those are things that could be solved and we just don't necessarily know the solution yet.

Kyle: [00:43:36] Yeah. So I think the Bitcoin debate is generally, yeah, we disagree there. I think the other is general biases. I'm kind of like the crazy one that's out trying to... I'm always trying to say what is everything that we assume is correct today that is not, and then try and figure out a place those bets. So recently I've been digging into some very interesting privacy technologies and some zero-knowledge proof systems, and I am slowly coming to the conclusion that the entire concept of blockchains, as we know them today could become irrelevant, relatively speaking. I'm trying to kind of formulate these crazy hypotheses and go off and do crazy stuff. That's generally what I do in the firm.

Patrick: [00:44:07] So can you say more about that? So what is that thread? That sounds interesting.

Kyle: [00:44:13] Okay. So for some general context, Nick Szabo who's one of the patriarchs of the space, so to speak, so he wrote a post six months ago. So it's excellent, a little bit long but excellent, it's called, I think Social Scalability or something to that effect. Basically in the post, he talks about the fact that Bitcoin trades technical inefficiency for social scalability. I mean, the idea here is that the more people run Bitcoin nodes and run validations and everything, the harder it is to modify the system. And it becomes more trustless because this system only will act according to the rules and no one else can modify those rules. And because those rules are well known and understood in public, that they scale socially such that 7 billion people theoretically could opt in to that set of rules if they chose to transact accordingly.

And so there's a kind of this very clear trade-off that he makes of you're trading technical efficiency for social scalability, and that's been largely taken for granted in the space. The Bitcoin people certainly espoused this very strongly. They are very focused on decentralization, decentralization, and then their tagline is don't trust verify. And so the idea is running on Bitcoin node and verify the transactions yourself. So in the last few years, there's been a lot of really interesting technical innovation in this area called zero-knowledge proofs and the idea of zero-knowledge proofs is someone can run a computation, given some input and given some function, they can run some computation, produce an output, and then they can generate a proof that they ran that computation correctly and give you that proof without actually verifying, without you as the verifier, knowing what the input was or even what the function was. But other than to say, I can assert that this is mathematically correct-

Patrick: [00:45:44] Truth state...

Kyle: [00:45:44] It's true, yeah. Cryptographically true. And this is now doable in a generalizable way for any arbitrarily large computation. It's really a mind blowing concept that this is doable and the math is very sophisticated. But if you take the idea of, well, some other guy over here can take a computation or a set of computations or otherwise Bitcoin transactions, so to speak, run them correctly and then give me a proof that all of them were run correctly. Meaning no one double spent money, no one broke the rules, so to speak. Then I don't have to trust anyone other than I have the proof and I know the thing is correct.

And this proof we're talking about is on the order of a couple hundred bytes,few kilobytes, tiny amount of data. The blockchain is 200 gigs today. And you can compress any arbitrary, large computation into a fixed, more or less fixed sized proof. And I'm just thinking 10 years from now, what if we all look back at her like, yeah, man, remember when we all thought the right answer was to replicate everything 100,000 times across all these computers, when you can just generate a proof that gives you the same mathematical guarantee.

Patrick: [00:46:39] How does that tie back to a simple idea of a fixed supply of a digital asset? I'm a little confused as to... because I understand Bitcoin fairly well and the nodes and the miners and everything, the whole ecosystem makes a lot of sense, it's very complete. So let's say we go that direction, then what is the digital asset? Who fixes the supply? How does it become valuable?

Kyle: [00:46:59] So to be clear in everything I just described here, this has nothing to do with supply schedule or inflation or anything else. What this would have a direct impact on is the people who maintain the network and not necessarily the supply schedule of the network.

Patrick: [00:47:09] Got it.

Kyle: [00:47:09] So rather than having all these miners mining and whatever, and rather than having to have a bajillion nodes out there trying to reverify the same transactions, the idea is you could basically have a very, very small number of people and you could trust them because the proofs they generate are mathematically guaranteed that they did not violate the truth. The whole point... the problem with the status quo basically is well, uncle Sam and the banks collude or conflate money and whatever, and there's call it, 10 of them in a room and they can make these decisions arbitrarily. And the whole point of Bitcoin is, well, if you enforce the protocol, if you make a million people enforce the protocol, then no one can collude. And so the idea here is you get the benefits of centralization in terms of scalability and throughput, but you get a mathematical proof that no one violated the protocol and that anyone anywhere in the world can verify the proof for a trivial amount of resources.

Compelling Trends In Crypto

Patrick: [00:47:56] So aside from that, I'd be curious to hear from each of you, what trend in this world is most compelling today. So what is something that you've recently spent time on that you think is exciting or scary or something notable?

Kyle: [00:48:08] I mean, so the thing I think intellectually that's the most interesting, and I think this will be true for the next few years, is the smart contract platform wars. I believe with a pretty high degree of confidence this will be the most epic technology battle that we've ever witnessed. iOS versus Android was epic in terms of, they were really at each other's throats, but there was only two of them, and it was very one-on-one, more or less, and the windows was on the side. The last time we have truly global, truly multi-party, multidimensional platform wars, of material was the LS wars in the 80s and into the early 90s.

But those teams were just much less well-funded than the teams are in crypto and the things weren't global, it was basically only in the US kind of, sort of, and we didn't have the internet and there weren't already 30 or 40 million developers across the planet. And so I just look at the stakes of what's happening here, and look EOS just raise four or five billion dollars. The Ethereum foundation has over a billion dollars. These teams are extremely well capitalized, and they all believe that these things could be worth $100 trillion. Maybe they're not worth 100 trillion, but they think they are. And so it's totally rational to spend $10 billion swinging for it. And I just think it's going to be the most epic technology battles we've ever seen.

Patrick: [00:49:11] How about you, Tushar?

Tushar: [00:49:12] I'm really interested in the infrastructure layer here as well, but really in terms of what are the primitives that we can create today, which will enable new use cases? Basically, how can we expand out the design space or make that design space more easily realized for developers? So I'll give a couple of examples to make that a little bit more real. One example, we talked about this a little bit earlier is Augur. I see Augur and global prediction markets as a financial primitive, because out of those prediction markets, you can basically construct any number of securities or-

Patrick: [00:49:50] Derivatives.

Tushar: [00:49:50] ... other derivatives, et cetera. And that is going to vastly expand out the utility of the underlying platform, which is Ethereum, in this case. And this goes back to my thesis that things become money by being used as money, and people will first interact with crypto, not through Bitcoin, but through some DApp, because I want to bet on the 2020 election or because I want to bet on the Cavs versus the Warriors or something like that. So I think Augur is a great example. I think another really powerful example here is Tokenized securities. I see blockchain if nothing else as just a fundamental upgrade to the entire financial ecosystem in terms of not only things like, oh, well, faster settlement and fractional shares and easy audit ability and lowering the costs of administrating large public securities in terms of dividend distributions or collecting votes, et cetera, that's all really easy and intuitive.

But I also see it as potentially creating a whole new design space for securities where perhaps if I own 10 shares of Apple for a year, I get a discount on the iPhone. Or if I own shares a particular REIT, I get a discount on my rent, is one interesting thing that I thought of. But there are other new types of financial products that can be uniquely enabled by this technology. And so these are places that I'm focusing a lot of my time on is like, what are the infrastructure level plays that will enable developers to make more stuff? Because we're not there yet in terms of consumer adoption or mass adoption that we should be focusing on the top layer. I don't think that the user experience is there. I don't think that there's been enough time for the vast majority of people to get into the crypto world.

And some of these things just take time. It doesn't matter how much money you throw at it. There's another one of my favorite startup quotes is, some things are not about resources, nine women can't make a baby in a month. Okay. It just, it takes time to reinforce that story in people's mind, get the user experience there, get that all there. So what I'm most excited about is the infrastructure layer that will enable that.

Best Resources to Learn More

Patrick: [00:51:51] I always like closing with some resources that people that are interested in all of this stuff can go and continue researching on their own. Maybe you could each pick a person or a resource, and I'll say maybe one each that's an investing specific resource and maybe another investor or group of investors or firm that you really respect. Then the second would be a non-investing resource or more of just like an ecosystem. Nick Szabo would be a perfect example of a resource. So maybe would love to hear from each of you on that idea.

Kyle: [00:52:27] So I think if you just want general, if you want the history of the cypherpunk movement and digital cash, and how all this stuff came to be, there's a website called the Nakamoto Institute and it is excellent. The resources on there, I think they have like an education or literature page or something, and it's probably 20, 30 hours of collective reading across all those. Actually the books are longer, but it dives into the history of money, how money has evolved, history of the cyberpunk movement, the history of open-source technology.

All the things that have culminated in crypto today, it kind of covers the history of those over the last 30 or 40 years of technology, and then all the money history stuff. So the Nakamoto Institute is excellent. And then in terms of, I'd say just general ecosystem resources, there's a website called Token Economy that's excellent. They publish once a week on Sundays. And then if you want a little more high volume, I would check out... So I curate three Twitter lists, one is very narrow, 60 people. One is like 600 people, it's a little bit more noisy. And then one is like traders and chartists and technical analysts. So if you go to my Twitter, there's three, go to my lists, there's three of those are public if you want a little bit more noise.

Tushar: [00:53:22] Well, first I want to kind of talk about our blog, I think that's a useful resource, multicoin.capital, but I'll give you some others as well. A16Z recently published this list called Crypto Canon, which is basically a curated reading list of a bunch of different, really high quality resources that help take you through that crypto journey. So I think that that's an extremely valuable resource, both for the investors, as well as for the non investing, like the developer community who is more interested in building stuff than investing in stuff. And then otherwise, I think that Twitter is really where the conversations happen in crypto. It's all real-time. There's a lot of noise that you have to filter out. You kind of have to be comfortable using Twitter in a sense, and realize that you are going to miss things, et cetera, but that's where all those conversations happen in real-time. And I find that to be the most intellectually engaging place.

Patrick: [00:54:14] So my closing question for everybody, and I'll ask each of you is for the kindest thing that anyone's ever done for you.

Tushar: [00:54:19] I would say, so I recently had a mishap at my apartment where there was some flooding. It was quite bad in that a bunch of my stuff was ruined, et cetera. And I wasn't home. I wasn't able to deal with the whole situation. And so this might just be recency bias, but the apartment manager was extremely kind in the fact that she handled everything for me. She went way above and beyond the call of duty in order to take care of things and help make sure that my life got back to normal more quickly.

Kyle: [00:54:49] Yeah. I'm pretty stuck on recency bias here as well. But the last March through May of this year was just the most hectic couple of months, I think, of my professional career, just with travel and all kinds of stuff. And I remember... my parents stay about 45 minutes from where I live. So I was home for the weekend visiting them and I just fell asleep on the couch and my mom put a blanket on me. I remember I woke up and I was like, this was amazing. Oh, and then my mom has a little three-pound Chihuahua and her name is Twinkie. She's adorable. And Twinkie was huddled up on top of me on the blanket.

Patrick: [00:55:20] The simple stuff. Well, this has been a lot of fun, guys. You inched me a little bit away from my pure Bitcoin idea in these developer kinds of platforms. And it will be very fun to see where this goes from here. So thanks for all the insight and the knowledge.

Tushar: [00:55:32] Thanks, Patrick. Big thanks for talking to us.

Kyle: [00:55:33] This was awesome.

Created by @pbrychcy1