Blockchain Mind Candy: Increasingly Esoteric Applications for Blockchains - Dhruv Bansal

Guys, can you hear me okay? This is good. Whoa, blinding. Okay, hi, my name is Rapunzel. I'm going to be giving this talk. It's called 'Down the Rabbit Hole: A Series of Increasingly Harebrained Blockchain Schemes.' I'm going to try and get through like five increasingly harebrained schemes. This graph is accurate. I'm going to have to really up your tolerance level for weird stuff during the course of this. It's all cultivated stuff though, which I believe is in the grandest bar tradition. I should declare, I suppose, when I start up here, I'm not a Bitcoin maximalist. I by no means agree with that label, but I do at the same time believe Bitcoin is sound money, or it's the best we have right now for sound money, and it will become the basis of the global economy. Feel free to disagree. I think there are three major lessons that I have learned from Bitcoin. These are my ways of putting it, and I'm going to try in these various schemes that I concoct or talk about over the course of... how long do I have? Oh god, I should have checked. All right, of course, maybe the next half hour. So I'm going to try and reason from these lessons, from these kind of teachings. The first one: simple money is made up. All right, it does matter, but it is made up. Stateless socialism, I don't know if that's where I got the phrase, but it feels like people are eager to own the means of production directly. Socialism is weird because you can't directly own the means of production, so let's give the state ownership of that, whatever the state is. Thank you for expanding my mind on that earlier. And stuff moves atoms. I like this one. Blockchains, bitcoins, we made up digital concepts, but they move atoms in the world. They move us to be at conferences like this. We're made of atoms, but I think more interestingly, they cause all sorts of real-world hardware networks, data centers, transportation, ski... like this stuff gets built in support of a completely made up electronic digital protocol. So I think that's really exciting when you think about how economic incentive can shape the world. They really do move atoms. So scheme number one, I call this the Nakamoto Point. I'll define that in a little bit. I love this drawing. It's a little silly. I don't necessarily believe some of the more macabre visualizations that my friend Daniel likes to put out, but I do like this drawing. It's turning into Bitcoin. We might actually all be made of Bitcoin, but that's scheme number five. I'll get back to that. Here's a chart of Bitcoin's value. It's a small fraction of today's money. That's true. I updated the Bitcoin figure recently. Some of these are probably old. I stole this from somewhere. Okay, so Bitcoin can probably grow by a factor of a hundred to two thousand X. Let's just sort of put some numbers out there for you guys. Bitcoin can grow by maybe a hundred to two thousand X before it becomes like a really appreciable fraction of world money. It's like less than a hundred to a thousand percent right now. That same story is true for energy. People complain sometimes about proof of work and how much energy it uses and how dirty it is. Jeremie and I had this conversation earlier, but actually it uses a tiny amount of energy if you think about it. I mean, well, a lot. It uses about a nuclear power plant's worth of energy. But if you compare the y-axis on these two charts, Bitcoin is at like 60-70 terawatt hours per year, and that's thousands of terawatt hours. Even renewables are huge in comparison to how much energy Bitcoin uses. It's literally one nuclear power plant. One large nuclear power plant provides the same power as all the computers roughly. There's some fudging in there. That's interesting from a global politics perspective. I think there are certain dictatorial fellows in the world that maybe have access to nuclear power plants. One wonders what they would do with that. But same fraction here, right? A hundred to a thousand X. We can grow by roughly a hundred to a thousand times in terms of its energy usage before it really becomes an appreciable fraction of world energy. That's just an interesting concordance. Data centers by the way use 10 times more energy today, just an artifact. But Bitcoin is an arms race. The hash rate is always increasing because there's always marginal utility in acquiring more of the relative hash power so you can get more of the reward. Bitcoin miners have even learned that they can mine at a loss for relatively long periods because prices may appreciate, and that puts them back in the green. So the emotions and the economics behind this get a little crazy. When does it stop? When does Bitcoin stop growing and consuming ever more and more energy? I think to understand that, you got to look at energy. How does energy work? Where does it come from? We make it somewhere. There's a power generator, there's different cables that transmit and distribute power. Eventually it gets all the way down to your home and your local substation and your neighborhood. It's a centralized infrastructure, sometimes semi-public or whatever. But let's average. This is just an average number, it's totally different in a lot of places. Why do I have a chart about that in a second? Roughly, the power companies are selling you power at like 10 cents a kilowatt hour or so. That's the rough price point. That means they're probably selling it to the distributors, the transmission people, at maybe 7 cents. There's a little bit of a share that the networks take. Bitcoin mining is 25 cents per kilowatt hour right now. Again, rough estimate, depends on the day, the hash power, it's going to change, there's capital investments, so on. But roughly that's where we're at. I keep saying Bitcoin by the way. I'm talking about proof of work here, but it's secretly I'm only talking about Bitcoin. So worldwide electricity prices. Roughly this is a chart. Not every country is on here. I think some interesting countries like Nigeria are left off this list. But look, there's a lot of population down here that has really, really cheap electricity. So those people are able to right now, if they suddenly decided, those people, the power producers in those areas, if they wanted to just throw their kilowatt hours at producing bitcoins, they maybe make more money. That's interesting. This is not just a China in the developing world problem either. This is in California, a fairly sophisticated power environment. What we start to see is renewables, solar energy, kelp, sunny area, oranges. You got negative power rates. Power companies will pay you to take energy from them. They can't drop the energy. There's a thermodynamics cost associated with things like that. You can't just get rid of it. It has to go somewhere, which means someone has to be paid to take it. So you're talking about literally negative some number of cents in certain times of the year. This is the worst graph. Don't they like this? The bars are consecutive years, I guess. Well, the point is, it's the best I could find, guys. The point is there are strong reasons for power companies to want to mine Bitcoin if they could do it. And why don't they do it? Well, they don't really know about it too, but they're starting to. I've had some interesting conversations with folks here where it is beginning, but it's a scale thing. If you look at mining revenues, they're like six billion dollars a year or something like that right now. Bitcoin is not perfectly efficient yet, but let's even assume in the worst case scenario that all of that is being spent on electricity. That's not, but let's even assume that it were. It's a tiny fraction of the revenues of electricity companies. They couldn't, even if they wanted to, just start throwing their excess waste energy or trying to game and start mining today. There's just not enough liquidity, not enough revenues in the entire ecosystem to support that. But Bitcoin is an arms race, so it's going to keep growing. It would have to grow by interestingly the same factor, roughly one hundred to a thousand times, in order for the revenues, again if you assume some sort of perfect saturation, to approach something interesting for power companies to actually want to get inside of. So this suggests to me that this is a temporal process. We are in this era today where they don't do it. Power companies don't mine it, but maybe they will. So I introduce to you this metric, the Nakamoto Ratio. I refined it as the fraction of world electricity used in proof of work to secure the money supply. Simple definition. That number from the energy slide, by the way, it's like really small right now. It's like 0.001. It's a fraction, so it's between zero and one. This is not a very precise scale as I get into it. In 2009 it was zero. Bitcoin had just come out, just starting out. Today it is like this is today, it's 0.01 roughly. It's a tenth of a percent of world energy. And Bitcoin is roughly ten thousand dollars per coin. I think there's a point, some future point here, that Bitcoin is worth quite a bit more. It's large enough where power companies, some power companies, decide that it is actually an interesting sink for our excess energy. This is an interesting way to perhaps directly, without having to even deal with power transmission or distributors, to get revenues from this resource that is dropping in value because of distribution effects of green energy. There's some point where they're going to start mining. And maybe Bitcoin is... I don't know if you want to reason linearly that if it's ten times the world energy, that it's there for ten times the price. I don't think that's a fair way of reasoning, but I don't have anything better. That's maybe say if it hits a hundred thousand dollars a coin, now we're talking about sixty billion dollars or so a year in revenue in efficiently priced electricity costs. So maybe that's interesting. So that's my claim that there are two phases, and we're going to go through very soon. My guess, I don't know, 2020s maybe after the next halving, when Bitcoin has another spike up, we're going to transition from a phase in which proof of work mining and energy production are two separate industries, and then over time they're going to pinch into each other. Energy companies are going to realize they can do proof of work mining, and proof of work miners are going to realize their margins are better if they produce energy. Where does this stop again? Where does it stop? I think it does stop. I think there's a saturation point, and this is what I call the Nakamoto Point. The saturation point here is the Nakamoto Point if you want. It's when the marginal revenue from spending a kilowatt hour on proof of work mining is equal to the revenues earned by selling that kilowatt hour on the grid. If we reach that point, Bitcoin's got to be worth quite a bit more. It's got to be using a lot more of the world energy supply. But if it does get there, my conjecture, no proof, just gut instinct, my conjecture is that it's a stable point. That somehow having saturated this bound, the universes of energy production and the money supply will be kind of concordant, and at that point we will just use that amount of energy forevermore to protect our money supply. Maybe we'll argue about it. It could be like an inflation rate. It'll be like a base economic parameter that people decide they want to try to influence through policy during periods of great expansion. If we go to space or whatever, I can imagine us spending a lot more of our energy budget on real-world things, but I think there's just a balance point. I don't know that this is actually the figure. It's just a conjecture that it exists. Okay, that's scheme one. Am I doing on time? That's pretty good. Oh, really unstable guys. Wow. Okay, scheme two, I call this Chain Lightning. I stole this comic from SMBC, which I love. It was originally about physics, but I repurposed it to be about blockchain mesh networking, which I do think is awesome and would try to motivate a little bit of discussion about networking and how I think blockchains help us get there. Remember, I'm crazy. I believe Bitcoin is going to be the base of the world economy. I think the entire energy production apparatus will turn to mining large amounts of Bitcoin. So I do believe blockchains have the ability to move atoms and really reconfigure the world. So I think some of the things I'm about to propose, while they're a little outlandish today, people are working on them and I think they are going to exist. But let me first quickly review with you what is Bitcoin like in terms of the data structure, because in order to have this conversation I have to route a little bit through some of what's happening right now in Bitcoin's development. So it's a ledger. What is a ledger? It's a list of, and I've simplified it here, but which transaction ID, which output of that transaction, script which is really address, addresses are just scripts simplifying, and some amount. So in this case, the address 1AB has like a bunch of coins in it. Type of some of those, and then you can cross entries out in the ledger as they get used up. Really all Bitcoin is, it is a consensus, a shared ledger of title to these bitcoins that are constantly being transferred amongst people. And consensus and proof of work is just a way to keep this list shared and current amongst all participants in the network. And it's extremely slow. Whatever you believe about the Bitcoin Cash utility whole thing, if my reading of the whitepaper, despite words like cash and peer-to-peer, is you built a system that is slow and prizes the censorship resistance and decentralization, distribution, redundancy of data over almost every other possible feature. So what you've built is the most robust way to share a list of information, and this is what you choose to put in that list. So I had a comment yesterday in Kyle's talk about layers and Lightning, and this is really part of why I believe that is. I think the Lightning Network is a really compelling extension to the idea of having consensus about a list like this. What is the Lightning Network? It is composed of two different concepts. One is the Lightning, and the other is the Network. It's kind of a funny haha finger cuffs, but it's kind of like that. A bi-directional payment channel is a relationship that two participants in the network enter into in which they lock funds together. That's like the finger cuff. They're locked, and they can poke each other. They're going to be like, 'Hey, I'll send you a little bit. Okay, cool. I'll update our IOU. I'll send you a little bit. Okay, I'll update our IOU.' A crucial part about it is there's no consensus mechanism. Those channels are private. There's no having to settle that to the chain. There's no proof of work mining. That's not how it works. It's just using some of the basic protocol features of Bitcoin that this ledger provides. It's locking some funds together and exchanging IOUs. That's the Lightning. The Network is a literal network of these bi-directional payment channels. So now when I get poked, I can just poke you, and then you can poke some other guys you have a channel with, and we can exchange a series of pokes, and that's how we move funds online. None of that has any consensus about it. That's a hundred percent just driven by how quickly we can exchange that information and telecommunications transfer time. So it can be extremely fast. It can be on the order of milliseconds to seconds. And then you still get the protection of the underlying chain because there's a feature built in. This is part of why we needed Segregated Witness to be able to do this. There's a feature built in that if someone tries to break that channel, because it's distributed, if they try to cheat and get out of the IOU network, you can flush it and you can have some strong guarantees that whenever your channel's sequence of IOUs settles back to the Bitcoin blockchain, it's going to be what you expected. I'm not going to get into any of the details of how this works. I think the Lightning part is uncontroversial. You could have built the Lightning part in 2013 using email. The core protocol just provides you the primitives to do that. The Network part, that's the interesting part. That's the hard part. Here's some visualization, just like a hairball. I don't know what that is. It's like the Lightning Network today. There's all these things out there. The Lightning Network is centralized. This guy, mathematical proof dude, if you publish your mathematical proof on Medium, it's not a mathematical proof. My goodness, that's my strong feeling. Whatever, this article makes some good points that some people do Lightning's orders do have this, I think absurd and false idea, that Lightning works like this totally distributed mesh. No, it cannot be that. In order to scale, it is making a trade-off. It is saying that we don't deal with consensus at the baseline. We flush out to that and we replace it with a certain degree of centralization in order to allow greater throughput. The miracle part is what is the right metric? The hard part is what is the right measure of that centralization? It is insufficient to say that it is or is not centralized. Define that term for me. Part of my own background as a business is trying to measure and define and think about how to construct networks with particular transport properties. There's no one definition of this term. So there needs to be an application-specific definition that we think is correct, and users need choice. Given that we are centralizing here, if we centralize with an inability for users to, let's say, not route through PayPal's Lightning hub because they know they're going to get all sorts of weird tracking or whatever applied to them, if they would prefer to route from totally private dark net Lightning hubs, they should have that option in my opinion. So the software being able to actually provide you this level of control on how you choose to route, who you choose to connect to, is I think extremely important. Furthermore, what are the economic incentives? The Lightning Network has all sorts of different fee structures. I'm not going to get into the details, but there are some details. Are those details correct? Are they going to encourage the kind of growth that someone like me is going to want to see, which is to say centralized but not totally controlled, decentralized as compared to distributed? It's interesting that all the existing implementations today of the Lightning Network concept, none of them really tackle this problem. Where we currently stand with Lightning is we've got channels and they seem to be working. We have a basic ability to network together, but it's extremely inefficient and not scalable at the moment. It's okay. It's a first draft. It's fine. But I'd like to see the clients develop a more sophisticated ability for users to, one, just in general have centralized routing something closer to BGP if you're familiar with how that works, but then also have the ability to route with metadata, for there to be a way for them to determine what routes meet their needs. So engineering and experimentation is required. But let's pretend this works. Let's move on, increasing the quotient a bit. Let's say this works. What can happen? Can we build a distributed mesh network? What I mean by mesh network, I mean literally not like this is the graph. The internet kind of looks like a mesh a little bit. This is a telecommunications network that we're all using. We're using it right now. It looks a little bit messy, but it's not like you can do a trace route if you want. It's built out of all these things called autonomous systems. The internet is in the name, between the networks. There are a sequence or a collection of largely independent networks that the internet routes traffic between. So it's kind of like a mesh in a way that there's no one network, but really it's not. It is extremely centralized again, using some appropriate measure of centralization. Think about things like undersea cables or traffic satellites which are owned by just a few companies. The fact that telecom, like to the last mile, is a monopoly in many areas, and that everyone's traffic is routed through the same switching centers locally, it's not as decentralized as it could be if we lived in a world where every single human being just had a router on that was providing a mesh network exchanging bits locally. What's frustrating about that is the technology to build such a thing has existed for a long time. We know about wireless networking. We know about distributed routing. We could have built a wireless mesh network that was global if we had the economic incentives to. And we didn't. The only way to build telecommunications networks has been to take massive amounts of capital, go out there, build the network with wires and satellites and stuff, and then charge rent in order to get it all back. So that's what's happened. That's how telecom works today. There is no economic layer in the OSI model for how to transmit data over computer networks. If such a layer existed, and I'm not the first on this idea, I'm just capitulating some of the information for you guys, it would potentially allow us to incentivize people to operate equipment in the real world to go build these networks. It might incentivize them to go purchase spectrum that they themselves could use to broadcast signals on a mesh network that they are compensated for packet per packet. There's a lot of... we can get into these. You can imagine how things like messaging and content delivery networks and social networks and knowledge, all these things have really interesting analogs in a world where it's not all being streamed out of some central server. I also think, by the way, that one of the common criticisms of mesh networking in general is how are we going to have mesh flicks? How are we going to stream all this content all the way across everybody? The millions of people that want to watch Game of Thrones, okay, now how are we going to do that on a mesh network that is inherently more inefficient? I think that's a little bit almost not believing in the distribution protocols enough. Why can't the mesh network also incentivize the mesh distribution and delivery of data that is encrypted and then unlocked at purchasing time, but streamed from someone nearby to you in the mesh? If we're all watching Game of Thrones tonight, that means that data only gets shipped to our neighborhood truly once, and then we can just exchange it with each other. So I don't really believe the idea that telecommunications and packet accounting has to be separate from data accounting. I think they're naturally the same thing, especially in a distributed network. They're ultimately just looking for a distributed way to route to a resource, which could be a server that gives you something over the net, or it could just be data. They're still just information and bits traveling over the same mesh. So I think integrating all that, having an economic incentive layer, is a really powerful way to build a next generation, more distributed, fairer, less... I keep saying centralized, but what do I mean? Why is that important? Let's go back to that comic from the beginning. Why do we actually want to build a mesh network? Why did we really, other than it's awesome? I think all the things that I like about sound money apply to communication. It must be robust. We do not want to be censored. We do not want to be silenced. We want to be able to attribute our statements. In a world of increasing duplicity, we may even require such a network to preserve the sound money itself. In the outline in the prior scheme, if energy is... if Bitcoin is going up by factors of ten and a hundred, it's changing the world, and there's going to be some violence as a result. We may need a network like this to preserve our ability to transact in Bitcoin or other networks in the first place. Let me also offer you a political snapshot in time of why I think right now is an interesting period in which mesh networking should really evolve and grow. Just a little framework: action, outcome. Pairing up react solution gold and money, divorce. These are executive orders from FDR and Nixon saying, 'We're off the gold standard, you citizens, and then following, nation states could no longer demand gold from you from holding US government demand.' So we get things like the ability to print lots of money, the 2008 bailouts, quantitative easing. Go look at the chart. The amount of money now is crazy compared to what it was before. I don't understand economics well enough to know what the impact of that is. It can't be good, guys. Can't be good. So what happens? Some nut comes up with Bitcoin. That's a thing now. He's like, 'Oh, I got to solve this.' Okay, great. There is a parallel story happening in telecommunications. Things like the Patriot Act, Citizens United, the ability for money politics, control of information, monitoring of the citizenry. That's all that. That's how it is now. We were okay with that. And so now what do we get? We get Snowden. We get the 2016 election. We have Cambridge Analytica. There's more coming. A lot of us have worked in this industry and our hands are not exactly clean. In part of this, I would probably even admit about myself, how much is this a political issue for people? I don't know. I wasn't someone for whom in 2008 Bitcoin was a significant political issue. I don't know. But there's some response to this. And the response is this notion of a mesh network that cannot have a Snowden revealing that the government has been spying on everybody because they don't have the capability to do so because of the ability for people to route effectively. Just like in the Lightning Network. I'm going to have to move faster. Alright, quick points. How do you actually build this thing? Let's say you wanted to and you decided it was politically the time to do it. How would you build it? Would you release your own blockchain with a consensus algorithm and tokens? I don't think so. I think you build it as a Lightning Network because what is the data that's being exchanged in the Lightning? It's the same idea of a perpetual transactional network that has no internal consensus model that settles into a lower layer for all of its accounting. And the information exchange in the Lightning transactions is just numerical differences. That data can be richer. It can be blocks of text. It can be packets of information. You can still use a Lightning-like channel based routing mechanism that uses an internal IOU mechanism that still settles out to an underlying chain in order to build these kinds of systems. Lightning, just like blockchains themselves, is a new kind of technology. You should not expect there to be one. If someone can show that it actually works, there will be many. And we will figure out different kinds of rules. There will be some new bloom and silly Lightning networks that are terrible and have bad acronyms. But there's a problem. It doesn't really work as simply as that because there's a time scale issue. Bitcoin's consensus model is going to get you consensus in minutes to hours, and that's why you go and build Lightning so that you can have certainty about your coffees that you're buying. Love's coffee so much in crypto. You get consensus about that in a few seconds. It could be even faster. But no matter how really fast it is, it's untenable to me to imagine a real mesh network that is using packet routing where every packet corresponds to a Lightning level transaction. I think the more reasonable picture is layers of such Lightning networks. What we see is there's an analogy between establishing a connection to a mesh node and finding a route through a Lightning network. And now in your sort of layer 3 mesh network, you can exchange packets and keep like an IOU sort of chain going, and then you can periodically settle at the close of connection or every N kilobytes of data transferred. That transaction is a Lightning transaction, and then eventually that even itself will settle into Bitcoin. So I really view this as a good idea. Let's reutilize this concept and build more and more layers on top of what ultimately is becoming a stronger base value. You can sort of see why I believe some of the things I do about the size and longevity of Bitcoin. Let's move faster. Let's go quickly through time. Bitcoin, that's what we start with. I think that's the base layer. I said this in my comment yesterday. Lightning Network, that's a transactional locking IOU layer built on top of it. I think on top of that, as I've tried to discuss now, you get a mesh network for data and telecommunications. What comes on top of that? Applications. Once you can actually incentivize people to store larger amounts of data and you can incentivize them to exchange bandwidth with each other, you have a substrate, you have a cloud. An actual cloud. The real cloud, by the way, should be called the mine because that's really what it is. This is actually a cloud. This is distributed. It is out there. This is on people's homes and computers and devices. If you have that, if you get monetized and you've built the economic incentives for that to exist, now your application layer has some legs to it. You can actually build really useful applications that store significant amounts of data because you've gone through this inductive process of building these layers. So it's a chain of settlement layers anchored in a proof of work blockchain. This is what I see it developing into. And of course it all depends on the internet. The mesh, like Bitcoin transactions on the internet, Lightning Network transactions on Bitcoin and the internet. The mesh itself, the likely first use cases are just going to be getting to the internet. So the internet is vital to this whole scheme. Yet I think it can be bootstrapped. I think the mesh itself can be self-hosted at some point. Now that starts to get really exciting because now there are no... if that is the world we live in, if Lightning Network has grown that large, Bitcoin has grown that large, many, many years from now, you now have a self-hosted, completely distributed network that settles into Lightning, which settles into Bitcoin, which is itself transacting on the mesh. Maybe it sounds crazy, but I think it's a cool way to build it. And I think anchoring all of this in the concept of value solves a lot of endemic problems that currently exist in the world of technology. You don't get personal data collection as easily because I'm having to pay you everything that you collect about me. And if I habitually visit sites that are sucking everything about me out, it's going to deplete my own mesh account, and that's something that maybe I'm in some device to prevent from happening. Software bloat? Can we just add more libraries to this? You have to pay for it. Maybe think about that more. Be a little bit more conservative about your underscore.js. DDoS has no cost today. It would have a tremendous cost in a network in which traffic is metered. Security? Right now if they steal your Twitter password, it's embarrassing. If that means they can steal your Bitcoin, maybe you are way more anal about the demands that you make on the services that you use in terms of security. I think Ethereum, by the way, is doing this backwards. I think they're starting at the top and they're saying, 'Let's just build the application layer. We're so desperately eager for Turing complete computation, and everything else will just materialize below us and we'll somehow scale everything up at the same time.' I'm a HODLer, I'm a Solidity programmer, I'm part of it, but I think it's backwards. But maybe let's find out. Keep going. Don't stop. Alright, so I have three more schemes. One is about identity. It's called Humanity. I may be lured. Play this game Dark Souls. No, I am a nerd amongst nerds, my man Toby. Alright, cool. Let's talk about... I have this one, and then I have one about space, and I have one about quantum mechanics. I don't know if I can get through all three. Dude, just go. Roll. Okay, here's what I like about Dark Souls. A silly video game for dorks, but it has this really cool concept. In Dark Souls, when you die, you become inhuman, you become Hollow. And the game is single-player, and you kind of poke around, and you have to kill the enemies, and you have to get some resource. The resource is called Humanity. It's a shard of some massive soul that is split out across the world. If you get humanity, you can eat it, and then you become human. And once you're human, the game is multiplayer, and now you can go battle humans, you can cooperate with them, PvE, PvP. If you kill them in PvP, you get humanity. So there's this token, humanity, that is exchanging in this network. What's cool about it is the multiplayer/single-player divide. It's like the identity and your relationship to all the other players is determined by this shared resource, this humanity that you're always exchanging in this semi-adversarial, semi-cooperative kind of fashion. So kind of a cool game that way. And it got me thinking a lot about how identity works specifically around blockchains. So this is, I'm going to go fast. That's me. That's really me. That's my identity. And then there's identifiers that I have, like my email address, a public key, that one, whatever, all sorts of stuff. Things like my data, data about me. What's my reputation? How cool am I? 10 out of 10, by the way, guys. That's actually 10 out of 10 factorial. My network, who knows me, who hangs out with me. Attestations about me. Where do I live? What do I own? Is it my relationship to physical things in the world? Things made of matter, like homes, cars. It seems to me that when people contemplate blockchain and identity, they do this. They go, 'You are made of meat. We can't deal with you. You're not even part of this. Your identifier is good. That's a solved problem. That is you. You are a private key. Let's just move on and start solving all those other problems because there's no money to be made in an identifier.' But I posit to you, I am not a private key. I am a man. And I feel it's reductive to turn me into one. And besides, public private keys, that's not a solved problem. Everyone thinks this. Why do you need a blockchain for authentication? We've had all sorts of schemes, all of them terrible for how to authenticate in networks and so on. There's a new one, Distributed ID. This is an example of a bare-bones DID document. What is it? It's public keys, and then various additional data about you. There's no need for a token here. I can understand why there's a need for a token in attestations about me, my reputation, all those applications of identity, but it doesn't seem like a token is actually needed for identity. But I actually disagree because none of those other systems... again, think about the idea of loss. They again assume that I am a private key. If I am lost, I'm out of the network. It's over. I can't sign anything more. I lose the ability to decrypt any of the data I've created with that key. And also interesting, I lose the ability to prove that I am that identifier because I don't have the key anymore. I think this is the problem of identity. I think there's no other problem that needs to be solved more than this problem, which is what happens when people lose their keys. I think everything else needs to sit on top of that. And I think if you really get into how to solve that problem in a distributed way, you get a solution that really does demand a blockchain. First problem is not fixable, by the way. If you encrypt something with a key and you lose it, I can't get it back. Done. No token, not a problem. This problem is fixable if you lose your identifiers. We can get you a new one. Often times, how do we do that in centralized systems? We issue you a new one. There's an identity provider. It's Google. It's the government. It's your company. It's something. You talk to them. You're a human. You leverage the social weight of identity, other identity things around you, and you convince them. And that's not how it works in blockchains and distributed systems. You lose your private key, the coins are gone. That's kind of maybe acceptable for fungible tokens because they're fungible. So if you lose these, the rest of them can pick up the slack, so to speak. For things that are fungible, this is a non-starter for CryptoKitties. If you lose them, they're gone. Encrypted cookies or whatever. CryptoKitties. If I lose my private keys, I lose my house keys. That's okay. I have a friend that has a copy. I can get back in. If I lose the private keys to my house, what does that mean? It's my house. Whose house is it? Is it gone? It's no longer ownable. This is a fundamental problem when you think about how to relate fungible, broken meatbag humans into the world of cathedral-like blockchain stuff. So can blockchain provide an economic context in which to solve this problem? I think the answer is yes. Presenting Humanity. It's a blockchain that I'm never going to build because I do want to talk about... let's say we have two identities, ABCDEF. These new identities. Let's go ahead and have a token. Token is called Humanity. And we have balances of it in these identities. They exist. We can send each other humanity in transactions. We can have an economic layer about it. Identities can spend humanity to update their public key. So just like that DID document, we're going to have the data of this blockchain. In addition to addresses, accounts, and balances, also have a public key just associated with each address. That public key is changeable. You can spend money to change it. I also assert that you can make links amongst identities by spending humanity. So they cost money to create links between you. And then why do you want to do that? Because there's a rekey primitive. I think the cool solution here is you can make a crime protection against Sybil attacks. You make it so that neighbors of an identity can update the public key associated with that identity by spending humanity. Now that's a really interesting way. You have a locally changeable but globally immutable ledger that associates two identities, A and B and C, the public key, the identifier, the public key. And I think this is extremely powerful because if done correctly, these links that we're creating, these aren't like Facebook friends. These are more like people you give your house keys a copy to. These are people that you extremely, extremely trust because they can unlock, should you be robbed or stolen from, they can unlock every aspect of your digital identity by rekeying you. And it has to be an economic transaction, otherwise the software should be able to be Sybil attacked. We can just DDoS and hackers can rekey each other. If people really believe Humanity is valuable, then this kind of behavior can be incentivized. Details, I don't know. Maybe if you have multiple connections, we do some average, or whatever, or there's some minimum cost. Maybe there's some unrecoverable aspect to this. If you're the kind of person that is every week losing your keys, eventually your friends should stop vouching for you. This Humanity is in some sense a little bit of a vouchsafing tool, trust domains. You can have your root identity that people are vouching for and that is recoverable, and you can have from it daughter identities that only your identity could ever rekey. So you're able to have private systems that only you can control, but if you get locked out, you can use this economically motivated rekeying mechanism to recover that root account and then go ahead and control your console the same way. I think it's the same idea. It's just when you anchor it in this idea of a system that solves this distributed rekeying problem, everything else makes so much more sense. Now if I lose the private key to my thing, I don't lose my house. I just call up my friends and say, 'Hey man, it's disaster. Can you guys do this?' And it might cost them real money. I might have to pay them back, but they know that I will because they trust me. We're using real-world social trust as a proxy for how to rekey each other. And I think this is the basis for future things like political representation and stuff, which I don't have time to get into. And I'm super late already. Do you guys want to see this one? Yeah, all right. I'll go fast. All right, what is there if there's money in space? What is it made of? It can't be stuff like the Ferengi. The Ferengi are the Jews of space, this horrible anti-Semitic stereotype. What do these guys actually need this stuff for? It's like certificates, metal. I don't know. That doesn't make sense to me because physical stuff is not something you can move among the stars cost-effectively. Money has got to be information in the future. And now that statement maybe didn't make sense very directly 10-15 years ago, but I think we all know what I mean by that. There's a blockchain context in which information can be money, and that is useful for space. But a problem: space is big. It takes a lot of time to move signals between the stars, and any system for value transfer has to contend with this problem. Let's look at two quick paradoxes. Finite speed of light. This is a spacetime diagram. If you're a physicist, you immediately can read this. If you're not, I'm not going to have time to get into it. But let's just say imagine you leave Earth, you go to Alpha Centauri. On the way there, I use a laser needle cast, and you send back a transaction paying off all your debts on Earth. But only Earth knows about that transaction. You get to Alpha Centauri. It's going to take four or five years for us to relay that fact back to Alpha Centauri. You can live like a king and double spend all those same credits while you're at Alpha Centauri at the bar. The challenge with that is that there's no centralized system which solves this. The bank can't prevent this from happening because it's an information problem. When ships travel that close to the speed of light and stars are so far apart, and we still experience time in the same way, we're going to have a transaction time which is infinitesimal compared to the communication time amongst ledgers. Another paradox: the relativity of simultaneity. Every cool paradox in relativity boils down to this concept: simultaneity is not an absolute idea. When things are in relative motion, they think things which are the same time are different times. It's very interesting. So if you have two ships that are making this journey in opposite directions, and both home planets send them a signal, they will receive the signals in the opposite order. How are they supposed to arrive at a consensus model of finance if time ordering is up for debate? These are extreme scales. How can a blockchain help? Blockchains can solve this problem by forcing essentially integration time. You wait for multi-year actions. You wait for decades for those transactions to fully confirm, and you have hash power on the Kardashev scale. This is incredible. That Gotti was so cool. Definitely going to happen. I'll just make the point that going back to energy for a second, if you all are bothered about how much energy Bitcoin is using right now, just wait a little bit. I'll also make the point that let's not confuse dirty energy with clean energy. Using a lot of dirty energy is messed up and sucks, but just using free energy that's otherwise just floating out into space to secure the money supply, that doesn't sound like a problem to me. And it is awesome to do it at scale. Another comment to make: you can't solve these fundamental problems of time delay by having bigger blocks or some dumbass thing like this. This is a fundamental limitation to the nature of space and time. You must build layers if you would like to transact more quickly than this. There's no other solution. So like, I know this is not space, but I go back to what it is like, like 10 minutes and so on. Think about the implications of an extreme model like this for the real model that we're trying to solve. That's a very physics thing to do. Take one variable, extremize it, see what the solution is, project it back onto reality. I also think long time scales is something we have to get used to in space. It's big. We're going to live for thousands of years. It's going to take tens to hundreds of years to move between the stars. That's a cathedral mindset. And I think if we really do have blockchain feudalism, this is an interesting model for how we can start building space-based cathedrals. Okay, that's okay. I can do this in like two minutes. That's my last one. Okay, so to the future. This is a cool question. I like this one. What if Satoshi was a time traveler and he came back in time and invented Bitcoin, and he's going to pop out in like 2150 when we have the 21 million bitcoins that get mined, and then he's going to implement basic income for everyone and save the world? And all the angels saying, 'So sayeth the great Satoshi.' I don't really believe that, guys. But let me get into some weird time stuff. How does all of physics work? Basically two laws. Noether's theorem says that to every symmetry of space and time and fields, there corresponds a conserved quantity. I won't go through all of these. The fact that the world is the same here as it is here, translation in space, means that linear momentum is conserved. The fact that it's the same here as it is here means that rotational momentum or angular momentum is conserved. There's a bunch of mathematical objects. If you remember your electromagnetics, the phase of a wave is sort of irrelevant. It turns out the conserved quantity associated with that, the circular symmetry of phase, is electric charge. Okay, cool. A bunch of stuff is conserved. It has to balance. Feynman sum over past reality is the mutual interference of every possible thing happening at once. This is a reasonable interpretation of how quantum mechanics actually works. How does a particle get from A to B? It does every possible path, even ones that are classically forbidden and ones that travel faster than the speed of light. And then how do we calculate the actual path? They all interfere with each other, and we have a rule and optimization function which helps us pick out the contributions of that interference, and we wind up with the classical path. So when you look at the way particle physicists do physics, what they're doing is they're saying, 'Okay, here are two electrons. They come in. They have a photon that goes between them and go off. This is one combination. They integrate over all possible such combinations and average them all together. And at each vertex in each path, everything has to balance.' So the way I think about that now is nature as an adversarial system that is adopting a consensus algorithm designed to prevent double spending. At each vertex here, momentum in equals momentum out, charge in equals charge out. Yet you have crazy stuff happening supposedly in which that doesn't happen. You violate those rules of double spending. You manufacture more charge than you came in with. Nature has to sort all this out. And I'm not saying it's a blockchain, guys, but it's an interesting parallel. One more idea: time entanglement. This is really cool. Conservation laws. Let's say they have a thing that has two states, it's like plus and minus. And conservation law tells you it always got to add up to zero. If you have a plus, you got to have a minus. You might have heard of entanglement in space, spooky action at a distance. So if you have something here, and let's say you entangle these two particles somehow, whatever that means, just pretend you know what that word means. And then you have these two things, you entangle them, and then if you measure one here, there's this rule that says over there it's going to be the opposite because of this conservation law. It has to balance in some weird way. And it could be really far away. It can be further away than light had time to travel. And what does that mean? Can we use this to send information? People would argue about this for long. I think this stuff is mysterious. I don't really know the answers to be honest with you. I do still find it spooky even having years of thought about it. It turns out we now have this new idea that you can entangle something now, and then later the thing that you measure has to be the thing that balances it, with the same degrees of certainty. Essentially, this is a new concept of entanglement through time, here and later, or now and later, as compared to entanglement in space, here and there. It's sort of being loose, but the future can be known in certain cases with specially prepared states with certainty, in the same way that the past can be known. So that's interesting. I'm not going to talk about this. It's good. It's going to take too long. Google 'Wheeler-Feynman absorber theory'. Maybe inertia is just signals from the future. That's what it means. But I'll just close because I've taken a whole hour. Today's blocks are secure because they are the tip of a long, heavy proof of work chain of blocks going into the past. That is why we believe in the security of Bitcoin. What if, guys, what if each block were constrained by all possible future blocks using quantum timing in such a way that things like 51% attacks are impossible to achieve because you can't change the current history because it's anchored for all time, out until we get to space, we're going to Alpha Centauri and whatever else? I don't know. I think that's an interesting idea. That's it. That's all I got, guys. I'm sorry. That's something to chew on.

Okay, so I had a lot of questions, but I'm just going to ask. Can you talk more about blockchain feudalism and space cathedrals? Because I think if the economy collapses and you have trillionaires that are built off of Bitcoin, that's not going to be a better and more fair world. It's going to be a less fair world. Maybe in the long run it gets more fair, but it feels like to me any universe in which Bitcoin really is worth millions of dollars a coin, high per Bitcoinization and things like that occurred, is a universe in which most people's savings were reduced to zero. That creates an unhappy world of powerless people and rich Titans. And I don't trust rich Titans. But I do think that rich Titans, like kings and priests of yesteryear, are motivated by ego and their place in history, and they're willing to pay and fund and manipulate people to large start projects which may take hundreds of years to complete. Witness things like medieval cathedrals. And so perhaps people's egos are the only kinds of things that have the longevity to spend that much money to get us into space and go places. So I'm sort of, I don't know, not looking forward to it. I've got some bitcoins. Maybe I'll be like a Duke. I think this whole scheme is somewhere... I don't know. It's a pessimistic outlook.

Questions? We should eat, guys. Yeah, I'm hungry. Question. More questions.

I missed the beginning of your talk, so maybe you answered this, but I'd just be curious to hear your thoughts on how the physics and proof of stake security is contained within the network. You don't have this tie to the real world, try connected to physics. It feels like proof of stake... I cannot like... I could be a bad physics analogy, so I want to... Well, I feel like proof of stake is a totally sensible system for when you don't need the extreme levels of robustness and the degree of history preservation that proof of work provides. One thing that you know, you can't just take Bitcoin's blockchain today, like 200 gigabytes on disk, you can't just go manufacture 200 gigabytes on disk that has the same difficulty numbers as Bitcoin's blockchain does. No one can create such a digital thing. It's not possible to do so. It is possible to trivially manufacture a proof of stake chain that is arbitrarily long that has an arbitrary history. The reason you know you're on the real chain is because you're in the network and you trust people. I don't think that's a killer failing. You have to trust people to get the right Bitcoin software, etc., etc. And it's a totally reasonable solution when you don't need sound money. So I think there's lots of applications. I noticed I didn't mention this Humanity thing. What actual chain does it run on? I don't know. I think proof of stake is a reasonable choice for something like that. But I personally prefer proof of work for Kardashev type 2 civilizations.