Beyond Cryptocurrency: Blockchain for the Real World


All right, re- really this- this talk is not actually intended to be about cryptocurrencies. It’s- it’s about beyond cryptocurrencies. You know, what if you wanna use Blockchain, uh, for real world assets. And, you know, uh, like Dan I have this problem, uh, uh, that, you know, most of what I work on at Stanford is my student’s very excellent ideas, including the- the next talk that you’ll see. So, it doesn’t really make sense for me to talk about, uh, uh, my student’s ideas but currently I’m on leave and so I’m going to tell you about something that I’m doing not at Stanford but then Stanford requires this obligatory disclaimer here. All right, so, uh, let’s say that you wanted to send a dollar from here in the United States to a friend of yours in India, all right. So, if we tried to do this with today’s banking system. You’d find that it probably takes like the better part of a week to do this. That to send your $1, you’d probably be spending $20 in fees. It’s likely that your friend in India would probably be paying more than a dollar in fees to receive this international wire. Um, and in fact, you- this scenario is preposterous, right. Like nobody would actually send a dollar to India because it’s- it’s- it’s just, like completely uh, clunky and ridiculous today. So, in an ideal world, this would be completely as simple as easy as sending, you know, right. You just stick a dollar into your computer and it would pop up in India or maybe your friend, you know, could get six, uh, you know, ₹60 or something out of, uh, out, out of some bank, uh, in India. And it should be as easy as sending email. So, the obvious question, you know, is this a job for Blockchain, right? One of these magical Satoshis has come down and given us this amazing gift of Blockchain, should we be able to use this to send a dollar to India? Well, uh, before we answer that question, there’s just so much hype around Blockchain, and I think we should cut through it and talk about what is, uh, Blockchain actually give us? And really it kind of boils down to three things. So, the first thing that we get out of Blockchain is coin distribution. How do you invent some new coin out of thin air like, you know, Bitcoin or, uh, er, you know, or any of these other coins, distribute it to people in such a way that there is a, a limited supply of Bitcoin but people still believe it has value? And that’s something that, you know, in 2008, ah, it seemed kind of impossible. Um, and yet we’ve solved that problem, and boy, does it generate a lot of headlines, right. Cause you have all these people who bought some assets for, like, pennies and then it, you know, became worth $20,000 or close to it, uh, a few months ago. Yeah, so that gets a lot of attention. The second thing, though, is I think the even more interesting aspect of Blockchain. And that is that what Blockchain gives you, is it gives you irreversible transactions, even with, uh, parties you don’t necessarily know or, or trust, right. So, that’s what allows me to walk up to you, like we’ve never met before, we don’t trust one another, I can give you some digital coin, ah, and you can give me some physical real world asset that I can walk away with. Because you’re confident that, that digital transaction is not going to be reversed. And then the kind of the third benefit that we get out of, ah, out of Blockchain, ah, are kind of a whole set of purely incidental benefits that didn’t really require any scientific breakthroughs, right. So, this will be for example like Blockchain works 24/7 unlike the stock market that only works during certain hours or you know credit card numbers are terrible, they’re just numbers that’d be better if they were public keys. But, you know, and Blockchain gives you like public keys naming your accounts which is kind of convenient. Probably the one that has also gotten the most attention is the regulatory differences that you get from Blockchain. Because, it kinda breaks all the rules. It creates a lot of regulatory uncertainty and, so, it’s kind of allowed people to innovate sort of, you know, Airbnb, or Uber style, in areas that were, you know, probably overregulated, but, ah, regulated for a reason. And, so what we really need in the long term is, of course, to harmonize regulation across, you know, a Blockchain and non-Blockchain kinda thing. So, I think, you know, the numbers one and three have really gotten all the attention. But the super interesting one is, ah, is number two. And I’ll get to why a little bit later on. So, okay, so why do we have Blockchain? Well, because of this, this paper on, on, on Bitcoin a, a, a few years ago that basically had this insight that we can solve problem number one and problem number two at the same time, in a mutually reinforcing way, using this technique called mining. Okay, so what is mining? Well, mining is basically obtaining new cryptocurrency as a reward for making other people’s digital transactions harder to reverse. And the mine technique that’s, you know, most popular these days is called proof-of-work. Um, and that’s obviously the one that, that Bitcoin uses and, and, ah, and Ethereum and the idea is, you solve this kind of hard search problem, that’s a function of the entire history of transactions, plus a new batch of transactions you wanna commit. And, ah, it’s very hard to solve that search problem, but once you’ve found the solution, it’s very easy for everyone to verify it. So, you sort of put a whole bunch of work into searching for a Blockchain that solves these constraints, um, and if you wanted to reverse transactions, you’d have to do as much work as a- all the good guys have done, ah, to create an alternate Blockchain. But, you know, those guys got rewarded, and maybe you won’t get rewarded, so you won’t be incentivized to do that. There’s other forms of mining, like instead of computation, there’s proof-of-storage or proof-of-memory. And then there’s also, uh, proof-of-stake, where, you know, in one way or another you kind of use, ah, your coins to, er, people, you know, your clout depends on how many coins you have and that, ah, you use that to, um, to influence what the next, ah, set of transactions will be. Um, but they all basically boil down to, to mining. Okay, so, as I’ve said, you know, I think the most interesting thing about Blockchain is these irreversible transactions. And where that gets even more interesting, is when you have multiple kinds of token on the same Blockchain and you can trade between them, right? So, let’s come back to this example, where I wanted to send a dollar to my friend in India. Maybe I’m at the U.S. in bank one, my friend is at Bank four, here in India. And maybe Bank four has a bank account at another bank, Bank three in Europe. These are called nostro accounts when a bank has an account in another Bank four might say, “Hey, if you add €0.93 cents to my account at Bank three, I’ll give whoever you want, you know, ₹60 at my branch in India,” You know, maybe there’s some other bank, Bank two that has accounts of both bank one and bank three. And bank two says, “Hey, I’ll I’ll trade you, you know, if you give me a dollar to my account at bank one, I’ll give, you know, €0.93 cents at bank three.” And so, what you could do is, with a Blockchain, you could pack this altogether, into an atomic transaction, and execute this transaction, um, and actually change your dollar into rupees without trusting any of the intermediary parties, like you’d like bank two, right. Um, and so in order to make this work, what you use for banks to issue tokens, that represent claims on deposits, like in, you know, dollars, euros, and rupees. Um, and then the ability to execute these irreversible atomic transactions. So, this is a very compelling example, but there’s something interesting here, which is that I haven’t used cryptocurrency at all. Notice that, that there’s no, no Bitcoin or Ethereum or anything in this picture, right. These are just currency-backed tokens that have been issued by these regulated counterparties, right? Because they’re banks. And, and, you know, we trust our banks to be solvent because we have a whole bunch of, of, ah, of infrastructure around making sure that our banks are, are solvent and, and, um, and responsible. Okay. So currency is, is one of many examples of useful tokens issued by, ah, known counterparties that might be useful to trade on a Blockchain. So, another would be, say, like Kickstarter-style rewards, right. Now, I go to Kickstarter, I spend $500 to fund some, you know, special foldable canoe, right? And that entitles me in a year to get one of these foldable canoes, if the project is successful. So, that’s effectively a future for this product that’s, that’s gonna come out. Wouldn’t it be great if there were a secondary market for the future and if in three months I decide that I don’t want the ca- canoe anymore, I could, you know, sell that feature to some- somebody else, right? Well, you can do that if the asset that you’ve got in exchange for investing in this canoe project was on a Blockchain that contained other assets like dollars and you could trade for those assets. Other examples would be like certificates of authenticity for art or safety-critical parts. Um, you know, any kind of securities like stocks, bonds, mutual funds, ah, you know, obviously we use tokens for transit rides, you know, cell phone minutes, um, there’s a famous kinda economics paper on babysitting scrip and what goes wrong if you don’t have, ah, ah, enough supply of it. But, you know, that obviously, you could do on this, ah, you know, prediction markets, ah, whatever you can think of. The fact is that, if you can issue these assets on a Blockchain, if you kind of tokenize them and issue them on a Blockchain and the Blockchain gives you the irreversible transactions, then kind of for free, you get a secondary market in any kind of asset that you issue. Okay, so what’s the problem? Well, the problem is the Blockchains can fork. So, in July of 2016, Ethereum which is like the, the second biggest Blockchain executed this weird, kind of irregular state change, that was just kind of hardcoded into the thing. Now, what happened? Well, basically, what happened was that there’d been this contract called the DAO and there was a bug in the DAO. And because of that DAO people lost $50 million. Which at the time was a lot in, in the Blockchain space, right? Um, and so basically, everybody said, “Oh, this is, this sucks. You know, we wish this bug hadn’t been there, you know, if only we could go back in time and, like, fix this bug.” And so, 85 percent of the miners said basically, “Okay, here’s what we’ll do, we’ll just kind of change the rules and make up this special one-time-rule to basically bail out this DAO contract and reverse the effects of, of this unfortunate bug. And so, 85 percent of the Ethereum miners decided to go ahead with this weird change and execute it. The other 15 percent said, “Hey, wait a second, you know, the rules are the rules, like, you know, it sucks that there was this bug but, like, you know, people aren’t gonna trust us if we’re going to make up the rules as we go along. So, we’re gonna keep going u- under the original set of rules and that, ah, branch of the, of the, Blockchain became Ethereum Classic. More recently, there was this dispute in the Bitcoin space, where there’s one set of people who wanted, wanted to basically do this thing called Segregated Witness, which is probably a good idea objectively, although the- there were some issues in how they did it. Um, there was another set of people who said, “You know, what we really want is more transactions per second so we need to make the block size bigger.” Also they were probably motivated by the fact that SegWit defeated some mining optimization that some people were using. So, what you had was Bitcoin on one hand and then this kind of fork that shared the same prefix of the history is Bitcoin called Bitcoin Cash. And, you know, if you think that’s kind of a dumb idea, and you don’t like Bitcoin Cash, or maybe you don’t like Bitcoin, you know, if you had a Bitcoin before and now you have one Bitcoin plus one Bitcoin Cash. So, you know, you could turn around and sell your Bitcoin cash for $600, say today, and you know, who are you to complain, you just got a free $600, right? It’s kind of like if you had a company and the company spun off a subsidiary and you got stock in the parent company and the subsidiary, you could sell either one of the two. But, ah, the problem is what would this mean for token issuers, right? Suppose that you’re a bank and you issued dollars onto the Ethereum network in, say, like June of 2016. Well, now you potentially face twice as many redemption requests, right? Because you’ve got the same number of dollars on the Ethereum and the Ethereum Classic Blockchain, right? And this is clearly, ah, very bad, right. You don’t want to face lia- liabilities for twice as many digital dollars as you’ve issued. So the moral of the story here is that mining is actually pretty scary for digital asset issuers, right? I mean, first of all mining is anonymous, right? All it means is that somebody with sufficient resources has, you know, executed the computation to extend or or fork the history. And, you know, so far, we’ve been kind of lucky with the high pr- profile. forks have been kind of announced in advance and each has been assigned a separate name like Ethereum and Ethereum Classic. But you can imagine, you know, a branch where it’s just like a fork with like two different, you know, competing pools of miners were kind of duking it out and saying, hey, everybody join our mining pool because then we’ll give you uh, more rewards. And if this goes on for a while and they’re both playing under the same rules and the mining is anonymous, you can’t even name the two different forks. So how are you going to say like, who’s allowed to redeem your tokens that you’ve issued and who isn’t. Another problem is that the mining rewards themselves might actually be- provide insufficient incentive for securing your transaction. So even on the biggest blockchains like Bitcoin and Ethereum, the mining rewards today are under $10 million a day, which is dwarfed by, you know the number of, you know, the amount of sort of dollars, like there’s trillions of dollars moving between banks on a, on a daily basis. Right. And what’s kind of worse is that now there’s like Bitcoin futures, you can trade. So If you’re a miner and you’re about to do something really weird, you can actually hedge yourself against the damage you’re going to do to Bitcoin by selling a bunch of Bitcoin futures, right? And then you can kind of wreak all kinds of havoc. And the other thing is there might be non-financial kind of extrinsic incentives to do this, right. Let’s say you’re a forward looking company and you say, hey, I’m going to issue my stock on Bitcoin by using, you know, colored coins to represent my shares and now my stock can be traded in a Blockchain. And then you do some high-profile deal with Taiwan and next thing you know like weird things are happening to your stock sales, they’re getting reversed and, and whatever. Uh, and well, you know, nobody’s come out and threatened you. But you know that there are a lot of miners in China and like you know, this is just could be a very uncomfortable situation to be in. Okay, so the moral of the story here is we have to protect asset issuers from forks in the Blockchain. And so the obvious idea here is why don’t you require whoever’s issuing an asset to endorse all the transactions in that asset, right? After all, you know, the issuer is a counterparty to these digital tokens, so they should, you know be allowed to know, you know, who’s transferring the the money and make sure that it’s not- that the the coins aren’t getting duplicated. So you could kind of call this the PayPal model, right. You’ve got this centrally managed ledger like PayPal keeping track of who has all these tokens and obviously, PayPal doesn’t face the threat of double claims on redemption because because everything happens through them. But if we went to a world with kind of a thousand PayPal’s, you know, you have a PayPal for your transit rides, a PayPal for your cell phone minutes, a PayPal, you know, for for each of these things. The problem is that you then can’t trade across these different PayPal’s, right? So this rules out even basic forex trading. And I would argue that most of the kind of, the most compelling applications of watching involve introducing new token types and being able to transact, have atomic transactions that involve multiple different token types. So, the idea that we we had, uh, with Stellar which is a nonprofit I’m involved with is, what if we have many issuers issuing tokens, but all on one ledger. However, each issuer is going to define tokens in such a way as to prevent double claims. And so what we’re going to do is we’re going to have the issuers, each issuer is going to de- designate a validator authority who is kind of authoritative for one portion of the ledger, namely the holdings in that issuer’s, uh, tokens, right. And then all the different validators are going to cooperate to maintain the single-global ledger, but if history did fork, there’d be no ambiguity as to, you know, which version of the ledger was authoritative for, uh, for which token type, because you kind of declared that in advance. And these validators, they’re known parties, they’re not anonymous at all. They’re things like banks or you know, big companies like IBM for example. So, the idea here is that, well, we already trust the token issuers uh, to remain solvent and to be willing to redeem tokens, because that’s what gives the tokens value. So, why don’t we also trust the token issuers to make the transactions irreversible. And uh, in order to do this, we need to make, uh, well, so so the idea here would be that, imagine that, you know, you’re only gonna agree that a transaction happens when all the sort of counterparties you care about do. So, if you’re interested in cashing out dollars at bank one and rupees at bank four then you’ll say like, hey, I’m only going to agree to a transaction when, you know, my bank in San Francisco and this other bank here in Mumbai, India have both agreed that this has happened. But these banks deal with other banks, right. And so, they’re not going to agree that a transaction’s happened until the banks that they depend on agreed it’s happened who also will depend on other banks, and so on, until you kind of take the transitive closure of these dependency relations. And eventually, the whole world has agreed that this transaction has happened. And so, at this point, once everybody’s agreed, it becomes very hard to roll this back, and even if, if you remember an example I had bank 2, we didn’t know who it was, but it was offering us a good rate getting from Dollars to Euros. And so, if bank 2 somehow decides the transaction didn’t happen, but the whole rest of the world does, well who cares, you know, because at the end of the day you don’t care about cashing anything in at bank two, and so bank two could just be in its own weird little fork of the world. Or if wants to exist, and continue to exist in the banking system it’ll have to agree that that this has happened. So, uh, so in order to make this happen, we need one assumption and this is something that I call, The Internet hypothesis. Right. So, the Internet is this globally connected network, but what’s interesting is that there’s no central authority that’s dictating the structure of the Internet, right. Instead basically, the Internet has arisen out of these pairwise relationships between networks that, uh, enter into peering and, and transit agreements, right. And so, of course, this decentralized and peering structure could have resulted in multiple different Internets, but it didn’t, right. It resulted in this one network that we call the Internet. And t he reason is that sort of transitively, everybody wants to talk to everyone, right? I may not know who you are, but you know, I want to talk to say, Google. Or you want to talk to Google or I want to talk to somebody who wants to talk to Google, and so on. And if you take sort of the, the transitive convergence of everybody, people want to talk to, uh, you’ll get kind of a lot of overlap. So, the hypothesis here is that counterparty relationships in, uh, in these tokens are actually, uh, you know, you have the same kind of a transitive convergence, right. So, you have banks dealing with correspondent banks, dealing with clearing houses and essentially, these clearing houses are going to be like the the tier one ISPs of payments. So, using this idea, we can define a, uh, a new kind of consensus, uh, protocol that I call Federated Byzantine Agreement. And the idea is based on Byzantine Agreement, which is a known technique for achieving consensus amongst a closed group of nodes. But the problem with classic Byzantine Agreement is that it bases security on this notion of of, uh, of majorities, right, or or supermajorities, right. So, uh, so, it’s vulnerable to this attack called the Sybil attack where you say, hey, any like two-thirds plus one of nodes constitute a quorum and so a bad guy comes in and creates, you know, a hundred fake nodes that it all controls, and so it overwhelms the number of good nodes. So, the idea is we’re going to defeat these Sybil attacks, uh, by choosing quorum composition in a decentralized way and this will be much less symmetric, but it will depend on the actual trust that people place in the entities in the system. So, if this works, each node V is going to pick one or more sets of nodes that, uh, we call quorum slices, and V basically picks as a quorum slice any set of nodes that it considers important enough to kind of speak for the network as a whole, right. So, the choice is going to be based on real-world identities, but in particular, if you care about redeeming tokens issued by by particular nodes, you probably want to put those nodes in all of your quorum slices. So, given this idea, we can define a Federated Byzantine Agreement System as a set of nodes V and this quorum function Q where Q of little v is a set of slices chosen by node V. And now, we can define a quorum to be a set of nodes, uh, U that is a, that contains at least one quorum slice of each of its members. So, here’s kind of a simple example, uh, I’ll have four nodes, each one has a single quorum slice, and I’ll visualize the quorum slices by drawing an arrow from a node to the other members of its quorum slice, right. And you can see that the V1 has quorum slice V1, V2, V3, the other nodes have quorum slice V2, V3, V4. So, if we consider V2, V3, V4, this is trivially a quorum because it says, it contains, it is every one of its members’ quorum slices. What about V1, V2, V3? That’s a quorum slice it’s not a quorum because V1 is saying I’ll agree to anything that V2 and V3 do, but V2 and V3 say we’ll only agree to something if V4 does. So, the smallest quorum including V1 is actually the set of all nodes in this example. So, here’s a little more, more realistic example, have a kind of tiered quorum slice structure reminiscent of the Internet where maybe there’s a top tier of quorum of, uh, of validators that depend on three out of four of each other and then maybe a middle tier that depend on two out of four the top tier, maybe there’s a leaf tier. But like the Internet, there’s going to be no central authority appointing the tier one validators, instead it’ll be left to market choices. And maybe a market decides, hey, you know, the top four retail banks, you know, these are, these are what should constitute the th- the top tier validators. But you don’t need complete consensus on this, so V7 and V8 might be paranoid, and they might say, well we realize we have to deal with the big banks, but we don’t trust them. We also want to depend on some nonprofits whose in centives are differently aligned. So they’ll say, we’ll wait for two out of four big banks to sign off on our transactions, but also one out of these three nonprofits, and the nonprofits depend on tw- two out of three of each other. Okay, so let’s say that this paranoia is well-founded and Citibank walks up to V7, pays V7 a billion dollars and says, hey, V7, I just gave you a billion dollars. We’ve got this Blockchain magic, so you know it’s irreversible. Give me, uh, a billion dollars worth of goods. I’m going to walk away with them and you should feel secure that you got your payment, right. And so V7 says, okay, I need two out of these four big banks. Got that. I need one of these nonprofits. Got that. This nonprofit, needs another nonprofit. Got that. These guys need three out of four. Got that. Good. Everything is good. Now, it turns out, a, the paranoia was correct, and these big banks are evil, so they collude to kind of reverse the effects of this transaction. And then they walk up to V8 and say, hey, V8 here’s a billion dollars that, you know, we shouldn’t have because we gave it to V8, but it sure looks like we still have it, so why don’t you give us a billion dollars for the goods will quickly run away with that, right? And V8 says, well, I need two out of four big banks. Check. I also need one of these nonprofits. Well, Stellar and EFF aren’t going to sign off on this, right. Because they know that the money now belongs to V7, but maybe ACLU is kind of lagging behind, so V8 talks to the ACLU, but the ACLU depends on two out of these three nonprofits, so the ACLU say, well, sorry I’m not going to believe this unless one of these other nonprofits signs off. And that’s not going to happen because the other nonprofits know this isn’t legitimate. So those nodes will basically protect V8 against this double-spend attack. Okay. So basically I- I designed this consensus protocol based on this idea. Uh, and it has some nice properties, uh, like, uh, basically guaranteed safety when you have quorum intersection after removing all the nodes that, that deviate from the protocol. And that’s a necessary property for safety in this kind of model. So what I’ve shown is also sufficient meaning that, you know, basically SCP kind of provides optimal safety. So, you might regret your choice of quorum slices but you won’t regret choosing SCP over some other, er, uh, consensus protocol. Um, it also guarantees that a well-behaved quorum won’t get stuck. The core idea in the protocol is this idea of federated voting, right, nodes exchange these vote messages but each vote specifies the, the senders- the voter’s quorum slices and so as you collect votes, you can kind of dynamically discover whether or not a set of nodes uh, constitutes uh, a quorum. So the way this works is there’s kind of two phases to the protocol. In the first phase, you have nodes nominating values until at least one value, uh, has, unless th- un- until they’ve agreed on, uh, nomination for at least one value. So here maybe you’ll see the, if everybody is, they’ve agreed that tx_3 is nominated and then they kind of continue to propagate values until they converge. Uh, and ultimately they will co- converge on some set of nominating values which they can, uh, combine in some deterministic way. Um, and you’re gonna actually show that, uh, it’s, they’re guaranteed that all nodes will eventually converge on the same set of nominated values. Now, there’s a catch here because of this impossibility result in, in distributed systems for consensus that they will never know. E- e- even though you’re guaranteed to converge you never know that you have converged. So you take your best guess that the systems converge then you move onto the second phase called balloting which is pretty similar to Byzantine Paxos, if, if you know Paxos, uh, but it uses federated voting, right? So basically, you prepare a ballot and you commit a ballot and if everybody’s converged in the first round, this will be fast and, and, and easy. And if not, uh, well, it will be a little bit slower but it’ll still be guaranteed to be safe and it won’t get stuck. So as long as the network behaves eventually you’ll, you’ll confirm some value. Okay, so, uh, so this algorithm is SCP is in production use by this Blockchain called Stellar it’s like the eighth largest Blockchain today. Um, and [NOISE] we’ve got actually, we have the, last night when I was preparing these slides we had 106 validators participating in consensus. This morning I happened to check again on the way and it’s actually up to 108 so I guess the, the system is, is growing and they’re achieving consensus every five seconds. And if there’s no, it’s not expensive to participate, you just need to compute a few digital signatures, uh, every five seconds. Um, so we can, uh, you know, have, it’s, it’s much cheaper to participate in the system. It’s an open network and anyone can join but of course merely joining doesn’t mean that you’re gonna end up in other people’s quorum slices, right? That happens when people know based on your reputation that, you know, they should, they should depend on you or that they wanna depend on you. So, we’ve seen all kinds of people participate like these big companies like IBM, we’ve telephone companies, uh, you know, money transmittal organizations, uh, universities, so there’s, there’s a bunch of different, uh, types of validators in them. And our primary use case today is for international payments to make this much cheaper and faster. But we’ve seen all kinds of other uses kind of emerge. So for example it’s being used by coffee futures, uh, by this, this company called BECS 360, in app currencies, geological data, cryptocurrency baskets, uh, you know, basically, you name it if you wanna issue a token and you want there instantly to be a secondary market for that token, Stellar is a good platform for that. So, compared to other approaches, you know, you could try to use Byzantine agreement of our closed server set and say, “Well, these are the big banks, these are the people who should participate.” If that actually happened, what you’d probably end up hav- having happened is that a bunch of kind of third parties would come and be like auditing the performance of these big banks to keep them honest. And so basically, you kind of get a poor man’s FBA anyway. So you might as well kind of do it right and get optimal safety from that thing. And of course, you might also be tempted to use, uh, to proof-of-work. But again, then your consensus is intricately tied up with coin distribution and these incentives and if, you know, you’re not interested in generating new coins but you want to just issue tokens back at their real world value that’s not necessarily the, the best thing, in fact the incentives might prove, uh, uh, insufficient or ill-suited, uh, for the assets that you’re issuing. All right. Thanks. [APPLAUSE]. We have time for, for questions? Actually, I’m really curious and that so you have 108 validators? Yeah. Um, so what do the quorum slices look like in the real world? Um. [OVERLAPPING]. What do people actually choose as validators? Yeah. I think what people choose is a much smaller subset of that. I think if there’s 108, there’s probably at most 30 that people actually depend on. Um, but, er? [OVERLAPPING]. Are those banks or nonprofits or? Uh, like I said it’s all over. So IBM has like, you know, five of them or something and all around the world. Um, Stellar obviously has, has, has like three of them because we’re the, the people putting this out. But yeah then there’s like a Telco. There’s, uh, I mean you can, you can go to, uh, uh, what is it, Stellar. I have a linked in the slides. So like, there’s some, there’s some various websites that, that, that, that list these, um, so like Telenda does one, that’s it, that’s a Telco, there’s like a Tempo, it like issues these Euro-backed tokens, uh, they’re a money transmittal organization, um, there’s yeah, uh, I’m not sure since I saw all kinds of people basically. Is there any other? Yeah. Do you have any model for what might be the effect of bribes? Of, the effect of? Bribing. One organization trying to bribe another to, uh, collude? Yeah. I mean, I think this is a, this is a big problem with, with, uh, the proof-of-work stuff. It’s not, there’s not, uh, it’s not an issue in Stellar because, you know, as an issue where you- you’re protected, I mean, you’re, you’re the one issuing the tokens so, you know, basically you could-. [OVERLAPPING]. Just stand over there. Well, I’m just gonna remove my, my laptop. So, so yeah, so if you’re concerned about bribes then you definitely would want to use something like Stellar and run a validator yourself such that you could, uh, you could guarantee that, you know, no funny transactions go through that you don’t approve and nothing gets rolled back. Hi. Any oth- any other? Hi. So you said 108 nodes, right? Right. So, the nodes getting any. [NOISE] Sorry what, oh. So nodes aren’t getting any benefits like token or anything by joining it? No. It’s more like why do you run a mail server instead of telling all of your employees to use Gmail or something. Because, you know, it’s not that expensive and you want some control. Oh, you know, it’s, it’s convenient to like have control over it. So, and so mining Bitcoin is hugely expensive, right. Running a Stellar validator, yeah, you need like a computer and, you know, you know, a few gigs of memory or whatever but it’s not, uh, it’s not particularly onerous, right? You do it for the control. Mainly proof-of-work as a node than anything else? No, no, no. There is no proof-of-work. You were digitally signing a message saying like, “Yes, I agree that this is the next batch of transactions.” So, it’s ve- it’s very cheap to run a validator. Um, and what you get is the ability to protect yourself against double spends because you, you tell people you’re dealing with, “Hey, the transaction hasn’t gone through until my validator has signed off on it.” And my validator won’t sign off until all people I consider important have signed off on it. [OVERLAPPING]. But let’s say uh. [OVERLAPPING]. You could charge a service fee if you wanted to, as a validator. Uh, there is, you do not, you do not make money as validator. There are, there’s a small transaction fee for just antispam reasons but that goes into like an inflation pool. We do not reward validators any more than we don’t reward mail servers, right? Like you run a mail server because it’s convenient for you to have email and because it’s not that expensive, you don’t run it because you get mining rewards or anything. So, that’s, that’s our model and, and that, that’s proving to work out, right, because p- a lot of people they want a copy of the ledger state anyway so they might as well run a validator or they’re issuing a token and they want to protect themselves against double spend attacks. Example, I build a data distributor application on top of Stellar. So, let’s my a lot of them say that I distribute 30 percent of the data across the validator on update nodes. They are storing my data but they are not gaining anything. So, you have any restrictions on Stellar like you can- [OVERLAPPING]. Yeah. So there’s, there’s a native currency which is sort of counterparty free. And there’s a minimum balance in an account in that currency and that minimum balance is proportional to how much ledger state you are taking up. Um, unlike a UTXO-based Blockchain like, uh, Bitcoin, we can truncate history. You don’t have to keep the whole history. So you can actually reduce the amount of state and get back to your mini- your minimum balance. Uh, but basically like it boils down to yes there’s, there’s anti spam protections to make it expensive to consume a huge amount of state in the ledger. Thank you. Cool. Thanks dudes. Yeah. [APPLAUSE]

Leave a Reply

Your email address will not be published. Required fields are marked *