Evaluating Cryptoassets

Patrick: [00:05:15] Chris, this is going to be our deepest dive into nascent strategies for actually valuing individual crypto-assets. You've got a brand new book that's sitting in front of us that's out, I think, literally today. I've read a ton of what you've written on the topic. And I think you are, at one point, were the only traditional buy-side analyst that was covering these things, trying to deal with them as you would deal with an equity. So we'll start by your overall method for evaluating cryptocurrencies and then dive into what you think maybe some of the ways that Wall Street, the broader world will collectively decide appropriate ways to think about underlying utility value of these things.

Chris: [00:06:00] Great. Thanks, Patrick. Since these are new assets or we're dealing with a new asset class, we have to create new valuation methodologies. And the starting point within equity space is discounted cash flow models. The problem with protocols is they don't have business models, they don't have cash flows. They have incentive models that catalyze a community to pull together and provision a decentralized digital good or service. And so evaluating how the native asset that incentivizes that economy to work is very different from a DCF. That said, it can all be boiled down to a similar mindset in terms of going year by year and what are the demands for this asset and what does the economy that it supports look like. Now, the native or the cornerstone model within the valuation is the equation of exchange, MV equals PQ. And we can get into-

Patrick: [00:07:04] Let's get into detail.

Chris: [00:07:05] Okay. So within classical monetarism, MV equals PQ is used to look at the money flow of a fiat currency within an economy. And so M refers to the size of the monetary base. And so with the US dollar right now, it's around 4 trillion. V is the velocity, so how many times a dollar, a yen or euro or whatever it may be turns over within a year. Right now, the US dollar is in this five to six range. P refers to the average price of a basket of goods and Q to the quantity of that basket of goods. And so together, again, we have MV equals PQ. On the PQ side where we just ended, that multiplying together of the average price of basket of goods by the number of these baskets yields the economy of a nation. And on the MV side, you have the size of the monetary base of 4 trillion times the velocity of roughly five yields a value of 20 trillion, which is about what the US GDP is.

So that's how we set these things up. The question then becomes, "Okay, how does that make any sense within crypto-assets?" And this is where, as I said at the start, you can look at crypto-assets and the protocols within which they operate as these many economies, many economies that provision a good or two. And so in the case of something like Filecoin, that is provisioning cloud storage, that is the good of that network. And so I can set up a very similar model to a DCF, but using MV equals PQ. So I start off with my price per gigabyte, so dollar per gigabyte. I have to have a cost decline curve for that, which I base on historical cost decline curves of cloud storage or hard disk drives, whatever it may be. So that gives me my dollar per gigabyte in 2018, 2019, so on and so forth.

I then have to figure out the quantity, and that requires assessing the total addressable market of cloud storage. There's a lot of assumptions that go into that, but again, pulling together the information that's out there. And then the tricky part there is actually creating an S curve based on the time the network launches, the time it'll hit 10% of its total penetration. And then the takeover time, which is basically the inflection of the S curve from 10% to 90%.

Patrick: [00:09:26] Is an adoption curve.

Chris: [00:09:28] An adoption curve, exactly. And that adoption curve will tell me, if I expect Filecoin to have this percent penetration, how many gigabytes will it have to be provisioning? And so at that point, in any given year, I can multiply the dollar per gigabyte in that year by the number of gigabytes provisioned to get a dollar value. And that is the economy of Filecoin in that year.

Patrick: [00:09:52] It's like the GDP of Filecoin.

Chris: [00:09:52] That is the GDP of Filecoin. Now, taking this equation, MV equals PQ, if I want to solve for the size of the monetary base, M equals PQ over V. So I have to take that economy and divide it by a velocity. And this is where things are going to start to get murkier because we don't have a good basis or enough data to say, "Well, velocity of 5 makes sense or velocity of 10 or 20 or 100 makes sense." We have hypotheses about it, and we can base it on some of what we've seen with Bitcoin, but suffice to say it's a big assumption. And if we have an economy, say, a Filecoin that's turning over 100 billion dollars a year, if that thing has a velocity of five, that means Filecoin itself needs to store 100 billion divided by 5 or 20 billion in that year in order to function within or allow that economy to function.

After all of that, we have gotten to the size of the necessary monetary base. And I divide the size of that monetary base by the number of tokens in the float to get the current utility value of a token in any given year. Now, that brings me to another part of this model, which is looking at, well, how many tokens are in the float? And that is a function of the total amount that has been issued at that point in time. But I need to subtract out from what has been issued, the percent that are purely being held at store value because those have a velocity of zero and are available to this economy. I also need to subtract out the percent that are being bonded by the nodes supporting the network. And not every network has that feature, but some do.

Patrick: [00:11:36] Can you describe bonding in a little bit of detail?

Chris: [00:11:39] Bonding is going to have different manifestations depending on the network. But if you take something like proof of stake, which Ethereum may move from proof of work to proof of stake in Q1 of 2018, proof of stake requires you as a participant within the consensus process of Ethereum to bond a certain number. And I think the last I heard it was going to be 1,000 Ether, which is a significant amount of money at today's prices, $300,000, and to bond that to be part of the consensus process. And the new are actually the assets that you have bonded proportional to the assets that the entire group of people participating in the consensus process that dictates how frequently you basically are able to participate in the consensus process and earn a reward, similar to how mining works.

And so it's basically, the idea with bonding is you get people to lock up a certain amount of asset so that if they behave poorly, you slash that deposit and hence you have an incentive to not be a bad actor, as opposed to Bitcoin, which is kind of saying, "We assume you're bad actors all the time and we're going to continually punish you by making you exert work."

Proof of Work vs Proof of Stake

Patrick: [00:12:52] And just really quickly, the motivation behind this, touch on how proof of work has become maybe less and less efficient in terms of the amount of just raw energy required to run proof of work in the network as the consensus mechanism and why proof of stake, this bonding mechanism, might represent a more efficient, or if it does represent a more efficient way of reaching consensus.

Chris: [00:13:12] Sure. So to reiterate what I said at the end, from a broad perspective with proof of work, the assumption there is everyone is a bad actor. And so I am going to force them to continually expand energy, to participate in the consensus process. And the reason people think of it as inefficient is as the price of Bitcoin rises, there is actually a greater and greater incentive to participate in the proof of work process or to be one of those machines. And because of that, more and more and more people are becoming part of that proof of work process, which makes it more and more competitive, which means more and more electricity is being burned. And so actually, the individual machines themselves have become massively more efficient as we went from CPU to GPU, small period with FPGAs and then ASICs, but it's just the scale of Bitcoin's mining network has become massive.

That is very different from the way in which a lot of these proof of stake systems will operate where... I've heard someone call it virtual mining before, where you're not actually burning electricity to be part of the consensus process. You are staking your assets to say, "Hey, I'm a good actor. And I am putting down this deposit. And if I turn out to be a bad actor, then you can slash my deposit." And so it becomes a much more virtual consensus process as opposed to this very hardcore expenditure of energy and expenses to support the network.

Patrick: [00:14:46] That's really helpful kind of side train of thought. So let's circle then back to why this is important, why the amount that's being bonded and also the amount that's just being held without transactions should be taken out of the float back to the equation.

Chris: [00:14:58] Sure. So, for example, let's say 100 million tokens have been issued in total, but 60% of those are being held as a store of value, which we believe given some of the analysis I did with Coinbase is roughly accurate. It could be higher but...

Patrick: [00:15:15] 60%.

Chris: [00:15:15] 60%. So 60 million in this hypothetical scenario is pulled out of the float. And then let's say 20% is bonded to be part of this consensus process. That means only 20 million tokens are on the float and all assets are priced at the margin and price based on future utility. And so when this calculation of the current utility value of an asset really sets aside what is just held with a velocity of zero and says, "That is not continually being priced, that is basically just being held," what is being priced at the intersection of supply and demand at the bid and the ask is if you have a certain number of assets in the float that can incentivize or lubricate this economy within the MV equals PQ calculation, only those assets should be taken into consideration.

So if I have, from the prior example, let's say 20 billion or as 20 billion in the monetary base that needs to be stored, I would divide that 20 billion by the 20 million tokens that are in the float to get the current utility value of one of those tokens in that given year. Now, I have to do that with the way I've set it up into the future. And so then I can look at the current utility value of a token in 2025 and I can discount that back to the present at whatever discount rate I want to use-

Patrick: [00:16:39] A very high one.

Chris: [00:16:40] A very high one, 30%, 40%, 50% to get, "Okay, this is a rational market price in 2017 based on the future utility value in 2025." That rational market price in 2017 hopefully has a certain percentage that is current utility value based on this whole calculation we've gone through. And then the premium above that, you could call it speculative value, but I can make the same argument for how stocks are priced and we don't call those speculative values. It's really just the expected utility value over current utility value.

Patrick: [00:17:14] So let's use Bitcoin and maybe the fairly simple example of remittances as potentially the PQ side of this to really put some meat around this idea. And maybe we'll even do a few of these because I think it's complicated and there are a lot of assumptions that go into this and changing velocity from four to seven makes obviously a huge difference in the outcome of the model. So let's use Bitcoin example and maybe walk through how you think about that entire equation for this hypothetical.

Chris: [00:17:41] Perfect. So current remittances market, just to make the math easy, roughly $500 billion. It's a bit north of that, but roughly $500 billion. If we make an assumption using this adoption curve in a future state, Bitcoin is going to be facilitating 10% of that market. That means Bitcoin will be facilitating $50 billion of transactional volume in order to facilitate the remittances market. So that's the GDP in that case of Bitcoin for that one service it's providing. So the monetary base necessary equals PQ divided by V if I have an assumption about velocity of, say, five. That $50 billion economy divided by velocity of five means Bitcoin needs to store $10 billion in order to allow that. Now, if it has a velocity of 10 for that use case, it only needs to store five billion because basically you're getting twice as much value carrying utility as you were with a velocity of five. And that's where we start to go down this rabbit hole of, "Well, how high could the velocity of these assets go?"

Patrick: [00:18:48] Let's talk about velocity. Maybe start with... I think you mentioned it earlier, but maybe start with a recent history, say, over the last 10 to 15 years of the velocity of the US dollar money supply and then talk about how you think about your assumption for that number for different crypto-assets going forward.

Chris: [00:19:08] I don't have the graph in front of me of velocity. I typically use the St. Louis Fed. But I know right before the '08, '09, crisis were around 10. And then we massively increased the size of the monetary base while the economy stayed roughly flat and so velocity has been cut in half. And so I think right now velocity for the US dollar is around five. Now, when we look at Bitcoin historically, that aggregate velocity has been in this five to six range. And so for example, to do that math, right now Bitcoin is floating with roughly $1 billion a day in transactional volume. So that doesn't take into account exchange-traded volume. That's just on-chain transaction volume within Bitcoin's economy, so $1 billion daily. If Bitcoin were to do that every day for a year, that's 360 billion roughly in that year.

And then Bitcoin right now is around 90 billion in network value, but you end up taking a weighted average for the whole year. And so let's say that weighted average actually turns out to be 60 billion because Bitcoin's inflected towards the late part of 2017. So that 360 billion in GDP divided by a monetary base of 60 billion yields an aggregate velocity of Bitcoin of six. Now, that aggregate velocity is composed of smaller instances in which Bitcoin is used. So, for example, going back to this idea of 60% of Bitcoin is held just as a store of value, that has a velocity of zero. So 60% times zero plus 40% times some mystery velocity needs to equal six.

And so doing the math, that means of exchange velocity would need to be 15. And that means of exchange velocity would need to be even higher if more assets are being held just as a store of value. So if we were to take it to 80% are being held as a store of value, then 20% times mystery velocity needs to equal six. And so this is where we start to see the means of exchange velocities of these assets could be much higher than the aggregate velocity.

Patrick: [00:21:21] This all makes me think that the term cryptocurrency may actually be appropriate with Bitcoin specifically. If it's the equation of exchange and we're talking about a monetary base, effectively solving for the value of a monetary base to get like a utility value or an intrinsic value, however you want to think about it, of something like Bitcoin. But maybe talk a little bit about... Your book's called Cryptoassets, not cryptocurrencies. So talk about your impression of the different ways of categorizing these things. Are some currencies, are some commodities, are they all just something completely different and new? How do you think about categorizing the different types?

Chris: [00:21:57] That's a really good question. And we specifically called the book Cryptoassets and not cryptocurrencies for a reason. Right now, the taxonomy that we're running with, which I have no doubt will change given how early we are in this industry, the taxonomy is that crypto-assets are composed of three verticals. So there's cryptocurrencies, crypto commodities and crypto tokens. And the cryptocurrencies are what people most commonly associate or confused with the term crypto-assets. And those are a Bitcoin, Monero, Zcash, Dash, Litecoin, Ripple. And the definition of a currency is that it serve as a universal means of exchange store value and unit of account. And so these cryptocurrencies, that is their use case.

And that actually starts to get a bit more abstract within the MV equals PQ calculation because of the pricing actually of the good, like I struggle and other people struggle with this idea of, "Well, there's not really explicit pricing of what it costs to send a remittance." And that's why you may have noticed when we were talking about the remittance example, I actually just lump things into PQ. So that's cryptocurrencies. What is actually easier to grok for most people I've been finding are the crypto commodities. And so when I say crypto commodity, that is basically referring to this idea that in our physical world, we have physical commodities, so oil, natural gas, so on and so forth, and they power our physical world, but we increasingly live in a digital world.

And so this idea that we have digital commodities, compute power, storage, bandwidth, transcoding, re-encryption, all these different things, there's actually a price for that. It's hidden within the vertically integrated structures of the tech incumbents that we use today, but there is a price for that. And so what we see happening with crypto commodities, Ethereum being the most well known one, which provides access to this decentralized world computer is this massive proliferation of the dis-aggregation of our tech stacks and actually creating markets to price these crypto commodities on a 24/7 global basis, which could drive a lot of efficiencies.

And then the third vertical, the crypto tokens, running with this analogy of... In our physical world, we have physical currencies and physical commodities that come together and form an economy which produces finished goods and services, which the end-user interfaces with. Crypto tokens are the end user facing products. They are the things like Steam or Augur or Swarm City or whatever it may be that is pulling from a lot of this underlying cryptocurrency and crypto commodity infrastructure, much of which still needs to be built to provision this finished good or service.

Patrick: [00:24:54] Can you describe Augur as one example just to get a little deeper into a crypto token.

Chris: [00:24:58] Sure. So Augur is a prediction market. And basically, what a prediction market allows you to do is to put up something that you would like to bet on and someone else can take the other side of that. Now, Augur and another similar crypto token Gnosis are setting this up so that it's an unstoppable form of this market. It's censorship resistant. So you can end up wagering on anything, which certainly that can become very dark very quickly, but there are other more interesting things that we can wager on that, again, could drive efficiencies. I think that in the case of Augur and Gnosis, this is something where because it is so consumer-facing that it needs to be very easy to use. And so with a lot of these crypto tokens, you're going to, I think, see them crossing a consumer adoption chasm before they get real utility value versus some of the cryptocurrencies and crypto commodities.

Crypto and Censorship Resistance

Patrick: [00:25:54] I want to talk about this idea of censorship resistance. Another way that this has been put in prior episodes where I've talked about cryptocurrencies is permissioned versus permissionless networks, that effectively the worldwide web was permissionless. Anyone can start a website. You don't need to get some approval to do it. I think that's another way of saying censorship resistant, meaning you can do what you want on top of the network. How do you think about... And this could even come back to the investing side of something. You said this would get dark really quickly. Let's say someone starts betting on someone's assassin, the odds of someone's going to get assassinated or something horrible like that, how do you think about how that might affect the movement or the support of these things? That could happen, right? And after that, my guess is there would a huge outcry to somehow shut this thing down. So let's talk about the pros and cons of this idea of censorship resistance.

Chris: [00:26:49] Censorship resistance. Well, on one hand, censorship resistance could lead to total chaos. That's where you were going with that example. On the other hand, you have more the libertarian side of things, where we will self-organize, where we will self-actuate and self-police ourselves. I think the reality actually falls in-between. The vast majority of these protocols are being built by really well-intentioned brilliant people that are working to decentralize the world's data, wealth, and power. There are social protocols baked into the software that dictate the way in which people interface within it. And so myself, maybe being an optimist, my hope and also as an investor in the space, things I look for would be things that work towards enabling a censorship resistance that can be self-policed, I guess.

And you look at, for example, the Bitcoin community, super focused on censorship resistance, but within themselves, constantly policing each other all the time. People getting kicked out of the community, all kinds of political strife, really. And so my feeling with these systems is that if they're not serving for the good of humanity, then the masses will come down on the protocol and maybe it should be wiped out. It all depends on what that protocol is doing.

Patrick: [00:28:14] So we've got one equation and a variable in that equation that we could solve for that could give us, with all of the adjustments we talked about, what the price should be today. This would be the same as an equity analyst building a discounted cash flow model, comparing what they think the intrinsic value of the business is to the market's current price, whatever that gap is, is the opportunity, long or short. So talk about that gap specifically as you're seeing it today. I don't know how many of these independent valuations that you've done, but what's the dispersion or the standard deviation of the... You've got like a stack chart or something where, I don't know, 10% of its utility, 90% speculation. I'm making that number up. How different is that pie chart for cryptocurrencies today and roughly how much is in your estimation utility or discounted utility value versus just total speculation?

Chris: [00:29:11] Right now, a lot of these assets, the majority are composed of more expectation than current utility value. And you can think of it as very similar to when a newly issued stock hits the market and trades at a pricey P/E or price to sales. I mean, look at Amazon, it continues to trade, I think, above 200 times earnings. This is what makes a market as people disagreeing about these different ratios and how something should be priced based on future expectations. To your question about the standard deviation or how are all these things shaping up, I haven't gone through and valued every single asset. It's actually not what I have to do as a venture capitalist. I just have to make sure that the assets that I'm choosing to invest in have some reasonable basis of what I have to believe for this to turn out to be a good investment in 10 years.

So I'm actually always boiling these things down to the physical reality of how many machines have to be connected to network, or how many users, and is it 100 trillion, or is it 1 million because there's a big difference there? That said, I fully expect this to happen, this idea of a lot more sell-side research, a lot more buy-side analysts covering these assets. Something I've been heartened to see is since I opened source the general framework, I've started to see other teams, the actual crypto-asset teams use the framework, create their own inputs, and include it as part of their SAFT or sell documents to explicitly show people what their assumptions are and why they're pricing this asset. And so my goal is, as a community, we can find some way to create, say, for the top 100 crypto-assets, these models that rise to the top based on what people feel are the most reasonable assumptions.

Network Value to Transaction Volume Ratio

Patrick: [00:30:59] That's a good way to transition into other emerging ways of valuing these things. One of the things I wondered aloud at the end of Hash Power was what will be like the price to earnings ratio or the price to sales ratio of these things. So is there anything like that that's being proposed or people are talking about? What are some of the other emerging ways of valuing these things like a common ratio, like a P/E ratio?

Chris: [00:31:23] Full credit goes to Willy Woo, who I always give credit to for this because he is the first person who tweeted out the idea because everything in crypto happens on Twitter. But it's this idea... what's called an NVT ratio, so Network Value to Transaction volume ratio. And what that ratio is, is you take any given days network value, which is synonymous with market cap, but I use network value because these aren't companies, these are networks. So you take any given day's network value and you divide it by that day's transaction volume. And so if you take something like Bitcoin, right now network value roughly at 90 billion, daily transaction volume at 1 billion, 90 divided by one yields a ratio of 90.

Historically with Bitcoin, we've seen that it likes to base at around 50. And then everyone's still trying to work out what exactly happens in times when the market really heats up and speculation goes through the roof. There are a few interesting things that happen. There's certainly a period after the bubble burst, where this ratio goes really high because the price of the asset is still relatively inflated, but everyone has gotten scared and they're not really moving their assets. So your denominator gets small, even though the numerator is still rather quite big. It appears that right before the pop, the bad pop, not the good pop, right before things collapse, this ratio can soften a bit. And the best way to explain that is actually some of Bitcoin's transaction volume can be shoveling Bitcoin between exchanges for arbitrage place.

And so as things are really heating up in a speculative fashion, you actually have more people shuttling their Bitcoin between exchanges, which increases the transaction volume, but not really from a means of exchange use case for goods and services and economy, more to support FX volumes. And so this is just a ratio that we have pulled together and we're tracking for Bitcoin. I would love to see someone do it again for the top 100 crypto-assets. I've been working with a couple of professors to pull together the data set necessary to do this. It's a heavy lift because you actually need to sync a node to all the networks and pull blockchain level data and parse it. But again, inevitably something that will happen. It's just we have to find the right people with the right appetite and resources in order to do it.

On-Chain Transactions

Patrick: [00:33:49] I want to just dig into this billion dollar a day idea of on-chain transactions, to have you explain what that means, what the difference is with someone that's, say, going on to Coinbase and buy... If I go into Coinbase and buy $10,000 worth of Bitcoin, how does that impact that $1 billion number? Because I think the question I get all the time is what are the underlying use cases here? What's it actually being used for? Maybe remittances is one example that are literally on-chain transactions, not me buying it at Coinbase. So maybe talk a little bit about that difference.

Chris: [00:34:20] There's a big difference between transaction volumes and trading volumes. And so transaction volumes are as you're alluding to on-chain volumes, which are basically moving amounts of Bitcoin using Bitcoin's blockchain directly, getting cleared and settled within the chain. Now, what a lot of people think of as Bitcoin volumes are actually trading volumes. And that is within the isolated liquidity pools that we call exchanges, so GDAX, Bitstamp and these guys. And that's just matching a bid and an ask within an exchanges order books. And a lot of that volume never even hits the chain. It's within that isolated liquidity pool. Where it would hit the chain is if at a certain point an exchange has had enough Bitcoin go off its exchange that it needs to net out the books or people, just individual users, are sending from their GDAX account to their Bitstamp account. Then in order to move that Bitcoin from one address to another, the exchanges actually have to, or one exchange has to send another exchange the necessary amount of Bitcoin.

The third thing that'll go on is within some merchant service platforms. So, for example, if I'm a Coinbase customer and I send Bitcoin to Coinbase merchant, that also doesn't have to hit Bitcoin's blockchain, it can stay within Coinbase. So when I did a bunch of work with Coinbase in late 2015 and into 2016, we basically did something where we knew from some of Coinbase numbers and this publicly in the paper, we just said, "Okay, Coinbase is..." Within its platform, at least at that time, transaction volumes were about 5% of what we were seeing on-chain. So we just multiplied on-chain transaction volume by 1.05 to calibrate for that. Really, we should probably multiply by something like 1.2 or 1.3 to take into account some of the other platforms. And I'm sure all these numbers have changed since then.

Impact of Time-Horizons

Patrick: [00:36:20] Changed a lot. Another interesting one that I've seen in terms of valuing or evaluating these things, and this probably bridges nicely into venture capital or the venture capital way of thinking about, like you said, a 10-year investment horizon versus a speculative trading horizon, which is the quality of the underlying community and developers. Are there any formulas or methods that are emerging for how to think about the impact of the team on the value of the network?

Chris: [00:36:49] So there's a gentleman named Nic Carter who's making a name for himself on Twitter right now. He actually did a master's thesis under the sponsorship of ARK when I was here with the University of Edinburgh quantifying different forms of governance. And looking at different ratios, for example, within equities, there's the Tobin ratio, I think it is.

Patrick: [00:37:11] Tobin's Q, yeah.

Chris: [00:37:11] Yeah, Tobin's Q and looking at are there analogies from that into the crypto space? I don't think anyone has really formalized that. Nic took the best stab at it thus far and I think he'll continue a lot of that work. But right now, it's still a very qualitative measure.

Patrick: [00:37:31] If you're thinking about a 10-year horizon, instead of a much nearer-term horizon, which I think is what a lot of people are thinking about when it comes to buying and selling different crypto-assets, what are the key variables that matter, or maybe those variables that matter more on a much longer horizon than on a shorter horizon?

Chris: [00:37:49] Definitely the starting point is always technology. What are the tech chops of the developers behind on this protocol? And are they trying to make a breakthrough in distributed systems that people have been trying to do for the last 20 years, or is there a much more achievable roadmap and how qualified are they to do so? Now, interestingly, technology is kind of table stakes because these are open source systems and I can download all the source code and spin up my own network and basically I've recreated everything at no cost, which then takes us into the next investigation, which is one of governance. All of this conversation and activity we've seen around forks really comes down to governance.

And if you are not properly governing your community, which takes into account things like what is the asset distribution, what is the participation process in software updates, who is getting the majority of the new assets that are being issued, all these kinds of things, if these teams don't have a fair approach to governance, then they can actually easily be forked and you can actually see a lot of value start to accrue to a competing protocol simply because they are better governing the members of their community. So that's a really key if not the most key component of long-term, is this a team that I can put my dollars behind? And that's very similar with a venture capital approach of, "Okay, bet on the team."

And then the third one that I spend a lot of time looking into is the cryptoeconomics. Is the projected graph for this token based on all the assumptions I have one that goes up into the right, or is it one that stalls out and actually goes down into the right? That starts to become a balance of the adoption curve, so the growth of the unit economics mashing up against the cost of client curve of the service being provided, the inflation rate of the crypto-asset, all of these things come together so that... Sometimes you have really weird looking current utility value charts. And so that's why I go through this exercise.

Patrick: [00:39:51] It's almost like this investing opportunity. If it is an opportunity, is if we could somehow bet on the GDP, if I could make an investment in the GDP growth of-

Chris: [00:40:02] Cloud storage.

Patrick: [00:40:03] ... of an individual country, but instead of a country, it's a contained network. I can make a bet on the value of a network over the long-term. In that vein, how do you think about when you're thinking about a network, whether or not centralization or decentralization is appropriate for that thing? So you mentioned cloud storage several times, those are obviously popular because it's a clear use case. How do you think about evaluating whether or not something is just better off centralized? What does it take for you to say, "No, that's something that belongs in a decentralized network instead?"

Chris: [00:40:40] Well, I would say it always comes down to is the service that it's providing in some way better being decentralized and centralized? Because for example, we take cloud storage, I don't think that all files will ultimately be stored in a decentralized network. I think that some files are better suited to being stored in decentralized network and some files are better being stored in a centralized network. So what is the competitive advantage? And so this is where we get into some of the things we talked about before in terms of being uncensorable, but also uptime. If you look at Bitcoin's network from an uptime perspective, Bitcoin, for all its flaws and as much as people complain, it doesn't go down. It's supported by tens of thousands of machines all around the world, independently running this code, independently verifying and settling transactions. That is much more robust, much more anti-fragile in the way in which it grows and unfolds than a verticalized infrastructure like AWS or Visa or whatever it may be.

Since the resistance resiliency just generally as a technology system, there's also a case to be made in some cases for a performance edge. So you take something like Filecoin. They actually think that they can provision style files more quickly than AWS because there can emerge these local biases. If I am storing my files in New York, then the farmers, the people storing files around me are likely to be able to provision that file to me much more quickly if they're in the nearby vicinity. And that allows massive scaling, but with local biases in a way that centralized operators cannot.

Patrick: [00:42:23] what are some of the... This always seems to happen in any interesting field that a lot of the discussion coalesces around a couple of key examples. So like right now, it's Filecoin. Everyone seems to be talking about Filecoin, and for good reason. I think it's the biggest ICO.

Chris: [00:42:40] It's a big protocol innovation.

Patrick: [00:42:41] It's a big one. So it's good that we're talking about it. What are some of the other, I guess, potential or in development networks that you are excited about that you think are some of the most important projects? Maybe a few that everyone isn't as focused on is something like Filecoin.

Chris: [00:42:57] I can go with the underdog that doesn't get enough love. And that would be Steam actually. Steam was launched, I believe it was 2016, although it feels much longer ago. Steam, you can think of as the marriage of Medium and Reddit, but incentivized with a token. And they came out, they had a hyperinflationary monetary policy. They fixed that. They've hard forked 19 times, they've changed lots of things. But similar to Bitcoin, say what you like about Steam, it's there, it works, people are using it, and they're just proving out its resiliency. What I like about Steam is roughly two-thirds of the block reward goes to authors of content, roughly one-six goes to curator, so people that are actually upvoting, downvoting, and helping surface the best content, and then a remainder goes to witnesses and people that stake and get interest.

So there's a really finessed monetary policy and way in which value is distributed going forward within the Steam network, which I think is great and it's using a consensus model called Delegated Proof of Stake, which as you asked before, it's decentralized enough. Some people would say, "Oh, that's too close to being a centralized system," but it's decentralized enough to allow Steam to function in this global manner. So that's one and people can go and check out Steam if they like and look at the content, all of that.

Company Deep Dives

Patrick: [00:44:23] Real quick, just so I understand, so I always try to, maybe this is a naive framework, but think about each token that this is like a matchmaking service that you can spend the token to get something, that could be a transaction fee to send Bitcoin, it could be an Ether to have access to this kind of global computer, a Filecoin to store something. And then there are ways of earning it. So you've already talked about earning it, which is create content, curate content, secure the network. How do you spend it? What's the idea behind right now if I want to go read a Medium post by you I don't have to pay anything? So in this model, do people have to buy Steam to then spend it to access content? What's the other side of the marketplace?

Chris: [00:45:04] There are a few sides to this marketplace, but one of the most common ones is basically you can exert more influence on the network, the more Steam that you have. So let's say in the future, you're an advertiser and you want to exert a lot of influence on the network. You can actually do that based on then on a Steam you have and push content higher to the top. So you're actually paying to get some content to the top, and that gives you leverage within the system. But it's very explicit leverage. It's not sort of in the shadows. It's very clear what's going on, which allows people to trust the content more.

Patrick: [00:45:36] And what about just like an average person? I want to go read something that's... Is it published in the Steam network? How does that all work?

Chris: [00:45:43] So you as a reader, as a consumer, don't have to pay anything. But the way that Steam, at least in the last time that I went through it, the way that Steam allows you as a publisher to get content on there, you actually have to pay for the bandwidth. And you're more paying with it by the fact that you have Steam. The amount of basically material that you can put through the steam network is proportional to the amount of Steam tokens that you have.

Patrick: [00:46:11] All right. Let's do another one, another maybe overlooked project that you think is appropriate for a blockchain-based network.

Chris: [00:46:17] So really like the Aragon team?

Patrick: [00:46:19] I don't know anything about Aragon.

Chris: [00:46:20] Okay. So, Aragon, I believe they did their crowdsale in May. They're a team that goes a long ways back actually in the crypto space. They worked on a company called Stampery, which basically stamping material within Bitcoin's blockchain and using it as this monument of immutability. And what Aragon does, it's an ERC-20 token, so it runs on top of Ethereum, whereas Steam its own full stack. And the best way to say it in a sentence is it's decentralized Delaware on a blockchain, in terms of it provides a framework for organizations to manage... well, first to register, but then to manage their cap table, to pay-out employees if they need to, to be very transparent about where funds are being held, how funds are being used.

So a number of components, and this is constantly expanding as it's under development, could very easily see it becoming a bit like a stamp of approval, just like Verisign works on the web, this ICO, or this token project runs their funds and administers things using Aragon. And then as we get farther out in the future and this whole idea of decentralized autonomous organizations and codifying much of what we see in existing organizations and running it transparently on a blockchain, that's more the direction that Aragon is going in the future.

Patrick: [00:47:44] Really interesting. Well, how do you think about, I'll say, holistically so for decentralized networks in general and then when evaluating individual nascent networks, how do you think about and categorize existential risks? What are the things that in your mind could... And usually, the answer is there's no thing that could completely destroy the whole idea that the technology is not going anywhere, but maybe you have an idea on this topic. So existential risks, what are they and how do you think about them?

Chris: [00:48:12] Well, so it's going to vary crypto-asset by crypto-asset. But just for ease of example, we can run with Bitcoin. So Bitcoin's capital base right now is supported by half a million to a billion dollars of machinery. And you can do that calculation by just looking at the hash rate of Bitcoin and dividing that hash rate by the average hash rate of your average machine. I think the latest Antminer is 14 terahash, and you get the number of machines that need to be connected to the network. It's somewhere between 200,000 to 400,000. And then the average price of these machines is $2,000. That's a lot of money, but it's not that much money if I'm a nation-state.

So then it becomes, "Okay, well, I could basically try and rebuild a whole network and I have the funds to do it." But what becomes a limiting factor is actually that Bitcoin is secured by custom hardware ASICs and the production capacity of ASICs is not that great. So for you to actually try and rebuild the whole Bitcoin network, would be a painfully laborious process. You would probably want to set up your own fabrication facility and just quietly be producing your own ASICs and then hope that over time, no one is going to notice that you're amassing this massive amount of hash power. You just turn it all on immediately.

There are lots of scenarios, ways this could play out. But at the end state, let's say someone is 51% attacking Bitcoin, and they have the majority of the hash rate. At that point, the core developers can release a software update, which everyone's going to be incentivized to move on quickly that changes the proof of work from SHA-256 to something else and basically obviates the existing ASIC base. So then all this work has basically gone into doing something that at that point is obsolete. So that's one avenue of attack.

I can go through a number of different avenues of attack. There's the regulatory attack, where basically just-

Patrick: [00:50:03] We're going to jail.

Chris: [00:50:04] Yeah, and killing all the Fiat on-ramps. That's something that people talk about a lot, like China did. And sure, that could happen in one nation or another nation, or maybe there's a conglomerate of nations. But what we've been finding is that it's a bit this game of Whack-A-Mole, regulatory arbitrage, one nation whack's crypto and the nation next raises their hand, they're like, "Hey, you guys can come here." And so because these are supernational protocols, getting supernational collaboration from nations on how to handle them, it would be pretty amazing. And even then there would be... Again, I'm going deep down a hypothetical scenario and I can play devil's advocate within myself, but that's another one.

I think, really, probably the one that seems most plausible to me is that you end up kidnapping all the core people within the space, be they the developers, some of the miners, and that community and you basically, I don't know, hold them hostage until everyone shuts down the network, or I don't want to get morbid here. But there is sort of that side. And so this is where Balaji wrote a great post on the different forms of centralization and decentralization and how to look at those to better understand and investigate the resilience of these networks.

Patrick: [00:51:25] Such fascinating, interesting stuff. As you think about... And I know of course your orientation is more towards 10 years or very long holding periods. So volatility of the price doesn't really matter all that much to you.

Chris: [00:51:39] It's a buying opportunity.

Patrick: [00:51:40] It's a buying opportunity. As you think about though that kind of movement, is it fair to say that the equation that we started this whole conversation with sort of forms a base below which the entire speculative portion of the stack could collapse to nothing but that that utility base would be the floor, or does the collapse of the speculation itself affect the underlying utility base? It's another way of saying, could these go to zero if there's a 90% crash?

Chris: [00:52:10] I think of it exactly as you said initially, the former example, where basically the market price of these assets is a certain component of current utility value, and then the excess above that is the discounted expected utility value based on discounting a future utility value back to the present. And so that current utility value is the price floor. Now, I can imagine scenarios where it trades beneath that just as stocks can trade below book. It's hard for me to see that persisting for long if the network is truly providing value to the world. And we have this very clear understanding of the supply side and demand is there or growing.

Now, the way in which these assets could start to fall more and more in value in this doomsday scenario is if we go back to what I was talking about, the percentage and the float. If more of your store of value holders capitulate, "I'm out and I'm not going back in and there's no hands that come back," then you will actually have more tokens in the float and increasing number of tokens in the float. And so that actually will drop the price of the token more, but it will ultimately find this equilibrium price where you've got the really strong hands holding the people who have bonded and this economy is still ticking along, still using this asset because the economy is not dependent so much on the price of the asset.

If you think of something like Ether and people are working on this, but Ether is working to dissociate the price per Ether from the actual cost to use the network. And it needs to do that because... Just because the price of Ether is going up, doesn't mean the network should get more expensive to use. It just means that actually need fewer units or a smaller fraction of Ether to use the network. And so therefore the expense stays stable and that's why I'm not overly concerned. I know there are some massive crashes coming. But so long as we're not deluding ourselves about the actual utility of these networks and these assets and what we're building, which I don't believe we are by any means, long run we'll be just fine.

Patrick: [00:54:24] Last two questions, the first is around velocity. It's where we started. It's a really important variable in evaluating the value of one of these networks. Is there a scenario in which velocity just gets so high that these things are moving so much, that it permanently affects and depresses the value of these things? What's kind of the worst-case scenario from an investor standpoint we want velocity to be lower and other ways that people are thinking about artificially reducing velocity or something like that so that the M that you're solving for here doesn't need to be, say, just a couple hundred million dollars or something?

Chris: [00:55:00] So there are all kinds of scenarios and ways in which we can see velocity progressing. If velocity is infinity, then M equals PQ divided by infinity implies your monetary base needs to be zero. And that's just a theoretical example. I would say this is where building-in mechanisms to the actual consensus process that require people to bond, not only for this value component, but also reputationally. Like if you're going to participate, you should bond in their protections against sybil attacks, where I can replicate my identity and actually become a threat to the network. It becomes much harder to do that if I have to stake $300,000 each time I want to replicate a node on the network. So we are seeing people build that bonding in.

I do expect this asset class to be a new store of value on a global basis. And right now, it's serving as a disaster hedge against global macroeconomic dislocations. But in the future, I won't use Filecoin, but let's say Golem for GPU flops and rendering and machine learning algos. If that's being used on a massive basis and very stably growing in its use, maybe that becomes a reserve commodity in a sense. Maybe we have these new massive protocols that just have this utility but have become stable and massive enough to the point where they're becoming a reserve asset of sorts.

Patrick: [00:56:26] Last question, has you most interested in terms of... I've read all your old posts and obviously your views have evolved as you continue to look through this. And when you were valuing Bitcoin using remittances, you mentioned in your post on crypto-asset valuation that there were some assumptions that you got wrong. So you're learning on the fly, just like everybody else is. As you think about valuing these things has you most interested, what's the frontier for you, the thing that you are most actively investigating that maybe you don't know yet?

Chris: [00:56:56] Well, we've talked about them a lot. It's formalizing our understanding of velocity, formalizing our understanding of discount rate. Discount rate within equities is the weighted average cost of capital. And we have formulas for doing that. For me right now with crypto-assets, I'm basically saying this is three to five times riskier than what I would use for risky equity, but I don't have a mathematical basis for that. As an investor, I can say, "Well, this is how much I expect to get compensated and-

Patrick: [00:57:24] This is what I need.

Chris: [00:57:25] Yeah, 30% CAGR for 10 years is a significant return. Yeah. But we don't have a formal basis for that. And then related to all of this is really the psychology of how these assets are going to play with the masses as they gain broader adoption. But someone is going to win a Nobel Prize on this stuff. We saw Black Scholes. It may not be me. There may be multiple Nobel Prizes that get one. I just want some crypto people to win Nobel Prizes, 10, 20 years out.

Patrick: [00:57:54] Well, this has been really, really informative. I've learned a ton in this hour. Thank you so much for your time.

Chris: [00:58:00] Thank you, Patrick.

Created by @pbrychcy1