0:01 Hello everyone, it's time for BlockTalk. 0:03 I've got a couple sponsors to mention first. 0:06 First off, we have CoinAgy, a professional tool for cryptocurrency traders. 0:10 You can link all your exchanges to one place and do some trading. 0:14 These guys also do a really good web show. 0:16 I was on their show last Wednesday. 0:18 Search it up on CoinAgy on YouTube and sign up today at coinagy.com slash blocktalk. 0:24 The other sponsor is Jumbucks. 0:27 They're a pretty cool altcoin that does lots of free tools for traders to use. 0:32 This is one of their sites here, getjumbucks.com slash wisdom. 0:37 You get a nice 4x4 little trading tool. 0:40 We appreciate their help. 0:42 Now without further ado, hello. 0:46 We've got everyone here in the basement. 0:49 We've got a special guest calling in, Paul. 0:53 Hi, how's it going? 0:56 Hello, how's it going? 0:58 Oh, fantastic. Thanks so much for joining us. 1:00 Thanks for having me. 1:02 So I guess if you could explain as simply as possible, 1:06 what is Bitcoin Hivemine and prediction markets? 1:10 Okay, well, so what prediction markets are, 1:14 I think it's easier to describe them. 1:16 They're known by an equivalent phrase which is called event derivatives. 1:20 So you can have a derivative that pays you a dollar 1:23 if the stock price of Microsoft is above some value, 1:28 like $43 or something per share. 1:31 And you can just have a derivative of that share that just pays you a dollar 1:36 if something happens to trigger that payment, 1:40 and otherwise it just doesn't pay you anything. 1:42 And so those are normal derivatives that have always existed, 1:45 but these prediction markets are event derivatives. 1:48 So they'll pay you a dollar if something happens in the real world. 1:52 So usually they're most popular with things like elections and sports, 1:58 but there are also other things like Academy Award winners or something. 2:02 But the cool thing is that these things are a lot of fun to trade, 2:07 and they have a lot of sort of financial value to people, 2:12 but they also inform third parties on the likelihood, 2:16 like what's actually driving events' likelihood. 2:20 So if something has a high price, 2:23 this thing is going to pay a dollar if Hillary Clinton is elected. 2:26 And if it's trading for like 95, 99 cents or something, 2:30 you know that it's basically inevitable, 2:32 or that a lot, a lot of people think it's basically inevitable 2:35 that she will be elected. 2:37 And whereas if it were this thing that's worth a dollar 2:40 if Hillary Clinton is elected, 2:42 if people are buying and selling this thing for like 10 cents, 2:45 you know that it's basically got something like a 10% probability of happening. 2:50 So there's this really cool stuff about like informing 2:54 what factors are actually likely to be connected to other factors 2:59 in a very – something that's very kind of subjective, 3:02 but can be measured in a very objective scientific way. 3:05 So it's very neat. 3:07 The term that I've described a lot is the wisdom of the crowd. 3:11 Yeah, sure. 3:13 So I guess when there's monetary incentive, 3:15 people are more likely to predict the right answer across a large group of people? 3:20 Yes, the problem, the wisdom of the crowd, the crowd is very wise. 3:23 I mean basically it's kind of a – it's a bit of a tautology 3:27 because of course if the crowd isn't wise, then kind of like what is, right? 3:32 So but the trick is you have to be able to aggregate the opinions of the crowd 3:37 in the best way. 3:40 So you don't want to spend a lot of money aggregating these opinions 3:43 and you don't want to have two opinions that, you know, 3:46 we have that right now in Bitcoin. 3:49 You want there to be like – you want it to successfully aggregate 3:52 and you want everything to be incentive compatible 3:54 and you don't want to have – there's all these frictions in aggregating information. 4:00 So like I always bring up that when you have a conversation, 4:03 everyone has got to basically talk to every other person 4:06 to kind of like figure out where they're coming from or who they are. 4:10 And this is a big, big like N squared kind of problem. 4:14 So that it's easy for people to be friends when there's like a small group 4:17 and everyone knows everyone else's name. 4:19 But then as things get bigger, you get these other problems. 4:22 But then things get too big, you get other problems go away, 4:25 but new problems appear. 4:27 And so the crowd is very wise, 4:29 but you need a way to aggregate the opinion of the crowd. 4:33 And that's what prediction markets do very well at large scale. 4:37 They're very, very well at large scale, which is what makes them really unique. 4:42 They're the only thing that I am aware of that can actually aggregate 4:48 the opinion of a large group of people in an efficient – 4:52 anything resembling an efficient way. 4:56 Can I ask an early question about prediction markets? 5:00 Sure. 5:01 What was it about Bitcoin or cryptocurrency that enables this system? 5:06 Obviously prediction markets have been around for a while. 5:10 What is it we gain from the new sort of crypto that makes it unique? 5:16 Well, the history of prediction markets is very bleak 5:21 because for years and years some of the smartest people around 5:27 have made like written pledges and written kind of like endorsements 5:32 that these things should be more encouraged, they should be more allowed, 5:37 they shouldn't be harassed or closed down. 5:40 But despite sort of the best of the best kind of tireless advocacy for this institution, 5:48 it really has not gotten – it really hasn't gone anywhere. 5:53 So there was something, the Iowa electronic markets, 5:55 but that was sort of scaled back where you can only have a few choices 5:58 and you're limited to like $500. 6:01 And the Intrade, the most successful kind of prediction market, 6:09 I guess marketplace for predictions, that was closed down. 6:13 It was constantly harassed and then it was ultimately forced to close. 6:18 So the story for me, I mean I was interested in prediction markets 6:22 long before Bitcoin existed. 6:25 But I just drew a lot of parallels between things like eGold 6:30 and Liberty Reserve, Liberty Dollar, even DigiCash, PayPal, 6:37 where they're sort of like you're on this track and you either kind of get neutered 6:41 or you just get destroyed and you can't really scale up. 6:46 And every time someone scaled up to get big, 6:49 to the point where they're actually changing the conversation, 6:53 improving the conversation, they got kind of eliminated. 6:56 So I think that – I just see a really similar story 7:00 and I was really inspired by the way that the blockchain had kind of fixed this problem 7:07 and I thought that this was basically almost the same problem. 7:11 But there's another layer to it, which is that Bitcoin currently is kind of missing – 7:16 if you want to have markets in Bitcoin, you have these – 7:21 you have to kind of – it's easy to set up bets with another person, a second person. 7:29 And you can even do this multi-sig bet where you have a third party be the – 7:34 sort of the arbiter or something if something goes wrong. 7:37 But the main benefits are really lost. 7:39 It's kind of – it's really similar to Bitcoin itself 7:43 where you had all this stuff that you could do with signatures, 7:47 but if you had – you needed a third party to prevent double spends. 7:51 And with this, it's really a very similar thing 7:53 where you could have all these synthetic derivatives and assets and all this betting 7:57 and you could do all this really cool stuff. 7:59 But the main benefits would be lost if you had this trusted third party 8:02 who had to be the oracle. 8:04 And then that guy could just take the other side of every trade 8:08 and he would control two out of the three votes 8:11 and then he could just steal everyone's money. 8:13 So I just thought it was – I just thought it was tremendously similar 8:15 and I was just blown away by this sort of shadow of a problem 8:19 kind of I thought repeated itself in Bitcoin, 8:22 prevents Bitcoin from doing things like having synthetic – 8:27 I'm not a big fan of like the BitUSD, but stuff like – 8:30 if you wanted to have like BitUSD or BitOil or something 8:35 or like BitCorporations and other things like that, 8:39 or any kind of bet really about anything, not just in finance, 8:44 but just about you could want to bet on anything. 8:47 Like a long time ago, there was a Bitcoin Savings and Trust Ponzi scheme 8:52 and someone on the forum bet that it wasn't a Ponzi scheme, 8:56 which effectively like there was like this arbitrage opportunity 8:59 if you could bet with him and bet in the Ponzi scheme 9:03 and like extract all this money. 9:05 But the problem with that was that the guy who was betting 9:07 that it wasn't a Ponzi scheme was sort of like didn't have enough money 9:10 or wasn't serious about meeting this obligation. 9:13 So there's all these kind of examples where all the stuff is there for the bet, 9:19 even two or three multi-sig or something like that with this arbiter concept, 9:23 but the arbiter can still steal everyone's money. 9:25 I just thought it was like it's just the same problem kind of circled around. 9:30 So you mentioned the oracles briefly there, 9:32 and that's kind of I guess the metric for analyzing the bet, 9:37 like figuring out which way the if statement goes. 9:40 So could you give a little more detail of how oracles work in your system 9:45 and just in general I guess? 9:48 So yeah, so that's the major problem, right? 9:51 And that's like I'm saying all the main benefits are lost 9:56 if you need to have this trusted third-party oracle. 9:59 So what I do is it's kind of a complicated thing, 10:02 but there's a white paper on this, 10:04 and what happens is that I create a very strong incentive for people to vote. 10:15 People have to vote on the outcome, of course, 10:17 but I create a very strong incentive for them to vote near other people. 10:21 And so there's kind of a lot to explain at once, 10:24 but for starters, there's like a vote corporation that earns money. 10:29 It's got this forced digital scarcity, and it earns money as people bet. 10:34 And this money, as a function of sort of the usefulness of the network, 10:39 this money kind of flows into them. 10:41 And so if you want to earn this money, 10:45 you can buy in to this digitally scarce token corporation. 10:51 And so you've bought in, and if you're in, you can vote, 10:56 and if you don't want to be in anymore, you can sell, 10:59 and you can get out that way. 11:00 So you kind of avoid an exit scam kind of thing there. 11:03 But it's kind of hard to explain because there's a lot of things going on at once, 11:09 but the econ theory behind it is that you want to establish this concept of reputation 11:14 and then have people's incentives aligned so that whenever they do anything you don't like, 11:20 they pay for it. 11:23 So the first way is to abstract this concept of reputation by creating this corporation 11:28 with sort of shares. 11:30 And then to go into another layer of it, 11:34 these people vote not on each individual thing as each event happens. 11:40 They don't vote one at a time. 11:41 They vote for many things all at once. 11:44 And then I do a big statistical trick called singular value decomposition 11:49 to see whose voting is the most deviant. 11:55 And then I punish this person the most, 11:57 and then I punish everyone else kind of proportional to how close they are to that guy. 12:02 So you can be punished for being ñ if you vote against the majority on something 12:09 where everyone else was voting with the majority, 12:12 you're punished more than if you vote against the majority on something 12:15 where people were confused. 12:17 And there's like a number of graphs and stuff that try to explain. 12:20 So, again, it's kind of not worth even really explaining the details 12:23 because just the kind of the crux of it, the impetus, 12:29 is that you want ñ I wanted something that mirrored real life. 12:34 So in real life, how do you know what's true? 12:38 If one person says X and another person says Y, 12:42 well, then what you start to do is you start to look at the other things that X and Y say. 12:47 So if Y is like a gibbering madman who says all kinds of random stuff, 12:53 then you start to get the idea that there's something wrong with Y. 12:57 It's Y. It's not this particular report doesn't mean anything. 13:02 So, boy, as if X is constantly right about everything 13:06 or at least normal about everything, 13:09 you know that there's something about X that's more reliable. 13:12 So after this kind of big matrix of votes is collected and this process is done, 13:21 I redistribute shares in the corporation if anyone disagreed. 13:26 So unless the consensus is totally unanimous, some people will lose. 13:32 They will have bought shares, but they will lose them to other people 13:36 who bought, say, 10 shares, but they'll end up with 12, the winners. 13:41 So, again, that's to avoid the exit scam, 13:45 which is the whole point, to have this economic kind of abstraction of reputation. 13:51 But there are so many other things that are kind of hard to explain. 13:55 I mean, another thing is that when the questions are listed, 14:00 there's an incentive there, which we can talk about if you want, 14:03 but there's an incentive to only create questions that are very, very easy to answer. 14:08 And there's even a fail-safe if you screw that up. 14:11 So if you create something that's too difficult to answer, 14:14 they can all coordinate on the middle answer, which is in between true and false, 14:18 which is like 0.5. 14:20 And so the questions don't measure anything other than the person's willingness to lie. 14:28 Everyone should be able to easily just answer the questions. 14:32 And so if they're not, it's not because they didn't know the answer. 14:34 It's because they're up to something dishonest. 14:38 And so that's another layer on top of that. 14:40 And then, finally, I just draw the process out. 14:43 Time is a net advantage to the honest person 14:45 because you can freeze money using the blockchain and lock it in time, 14:50 but over time information will become easier and easier to learn about. 14:56 So today we don't know who will be elected president, 15:00 but it will become clearer to us as time goes on. 15:03 And then in December of 2016 in the United States, 15:06 the United States president, sorry about that U.S. bias, 15:09 but in the United States we'll know in November, in December, 15:14 it will be very clear to us who is elected president. 15:17 So the betting can take place now when it's very unclear, 15:21 but the resolution can take place in the future when it's very, very clear 15:25 and it's fresh in everyone's mind. 15:27 So that's just more – there's kind of a lot of layers to it. 15:31 Ultimately, the process is so slow, 15:33 and the process of having multiple people vote on multiple things in a big time period, 15:39 large rare votes instead of frequent small votes, 15:43 the large rare vote, it actually collapses up into like one kind of easy-to-measure thing 15:50 because of the way singular value decomposition works, the technique I used. 15:54 It's very, very easy to see once every three months or so 15:58 which people are kind of like the honest people because you get – 16:02 which people are honest and which people are like totally on a different planet 16:05 because if you want to lie, you really want to lie about a lot of things, 16:09 and so their report will look very, very different. 16:12 So even if you only know about one or two of the things that are voted on, 16:15 they should like all be backwards or something. 16:17 So it should be very, very easy to spot on a large infrequent scale. 16:22 Then the miners even have options like veto things and just force them to happen again. 16:28 So it's just very, very slow. 16:30 So I just ultimately try to slow the process down to a degree where the attacker 16:36 just wouldn't even want to bother trying to go through with this elaborate – 16:42 So anyone who I guess really wants to figure out the exact fundamentals 16:46 or how the whole network functions, they can read the white paper. 16:50 The white paper is very long, and I released it when I came up with – 16:54 when I published the white paper a long time ago. 16:57 I also released it with code that had tests to kind of help describe sort of what was going on. 17:04 So there's all kinds of stuff like that, and I'm happy to answer any specific questions, 17:10 but they are kind of like specific. 17:13 They're not – I've tried this a lot of times, and every time I've tried to explain – 17:17 it sounds – it makes it kind of sound bad, but people came up to me like I think at – 17:27 like a couple months ago, but people in Bitcoin came up to me, 17:30 and they said we didn't know how to explain. 17:32 It's okay. You shouldn't feel bad that you really can't explain this 17:35 because no one really knew how to explain Bitcoin for the longest time. 17:38 So I guess – I don't know if people are just trying to make me feel better or whatever, 17:41 but it's kind of complicated to explain. 17:44 All the pieces there are just me trying to make this look – 17:49 this resemble how it is that you learn about whether or not Hillary Clinton was elected 17:56 in the most – like assuming that other people are trying to lie to you, 18:00 and so it is kind of very rounded off, but it does kind of sound weird 18:06 when you try to explain it for the first time. 18:08 Yeah, your white paper – like the table of contents is almost longer than Satoshi's white paper. 18:13 Yeah, that's bad for me. 18:15 But part of the problem is that Satoshi is a lot better at programming than I am, 18:20 so I kind of had no choice but to stuff the thing full of explanations 18:25 and hope that it would make sense. 18:27 Like he could just do the right thing, which was like release the software, 18:31 but I was like I'm not good enough with – nowhere near good enough with C++ 18:37 to sort of do that by myself. 18:41 Well, you explained the game theory pretty well with it, 18:44 and we can see like the Truthcoin white paper has been circulated a lot. 18:48 We've got some other questions here. Dave, do you have a question? 18:52 Yeah, thank you. Paul, first of all, thank you. 18:57 My pleasure. 18:59 I'm looking forward to reading your white paper. 19:01 I haven't yet, so excuse the question if it's covered by this. 19:08 I love that our community is making things like this possible. 19:16 Prediction markets are really a fantastic thing, and I agree with you 100%. 19:21 I haven't been in the real world enough. 19:25 It's true what you were saying about prediction markets being a tautology 19:36 in a sense. 19:38 In some cases, they're a self-fulfilling prophecy. 19:40 I mean if the same people that are betting on an election 19:43 are also the ones that are voting on it, then it becomes true. 19:48 But also just Adam Smith's invisible hand, 19:52 economic incentive to put your money where your mouth is 19:57 tends to make the truth rise to the top there. 20:01 I think it's great if we can put this in a blockchain-based 20:07 or some kind of provably fair system. 20:11 But I always try and challenge people that are building these kinds of things 20:17 to not forget that the real world has evolved quite efficiently as well. 20:28 In the real world, as far as trusted information, 20:33 reputation is the thing that tends to create things that are honest over time. 20:42 Bloomberg wouldn't sell any terminals if they got the stock prices wrong, 20:46 and ESPN would be soon out of business if they were reporting incorrectly 20:50 on the Super Bowl and so on. 20:54 Given that these organizations have advanced access to the data 21:01 and have incentives to report it correctly, 21:08 first of all, we trust them, but second of all, 21:12 with a system such as what I imagine that you have written out in your white paper, 21:19 would there not be economic incentive for these types of organizations 21:23 to buy the votes and profit from the information dissemination 21:33 and acting as an oracle? 21:35 Front-running. 21:36 Yeah, front-running, so to speak, or otherwise, 21:42 just kind of establishing and continuing their dominance in those areas. 21:51 And why would there be any economic incentive for someone not so connected 21:59 to buying vote coins? 22:03 Okay, so I think I'm not sure, so I might need some help answering your question 22:08 because you touched on a lot of different issues, of course. 22:10 So I think you mean that some people have – first you said that some people had – 22:16 the first part of your question part was that some people have access 22:20 to more information, and then you said why would they buy the vote. 22:24 The voting is just reporting on the event after it's happening, so it's not trading. 22:29 Right, why would anyone who did not have an informational advantage 22:37 choose to buy a vote coin, or I guess flip the other way, 22:42 is wouldn't all the vote coins already have been purchased by someone 22:46 who is ahead of that second person in the information stream? 22:52 Okay, so again, I think you mean traders again because the voters, 22:59 there's no inside information. 23:01 The voting happens way long after the event is over. 23:05 So you'd be betting on what's the stock price of whatever American Airlines group 23:13 on October 5th, but you wouldn't vote on that until like December 1st or something. 23:17 So it would just be some number from October 5th to December 1st. 23:21 That wouldn't change, and that's by design. 23:24 By the time that people vote, they all have common knowledge, 23:28 and there's plenty of time for people could have Googled that number at any time 23:33 in any of the intermediate kind of time. 23:36 So no one would want – there shouldn't be, although there could of course possibly be, 23:42 but there really shouldn't be any significant difference in information by the time 23:47 for anyone who owns any vote coins. 23:49 They should all have the same information hopefully. 23:51 Okay, if everyone has the same information, 23:54 why is there any economic incentive for anyone to acquire a vote coin? 23:58 So you don't want – yeah, okay, I think I understand what your question is now. 24:01 The vote coin – the people who own the vote coins are the employees. 24:05 So you can trade – the New York Stock Exchange is run by a company that is itself listed 24:13 on the U.S. Stock Exchange. 24:15 I don't remember. I think it's like IE or something, but whatever it is, 24:19 the vote coins are the employees. 24:21 So if you have inside information or – that was part of the joke for Intrade, Intrade.com. 24:28 Part of the joke was that that was insider trading because that was kind of like a funny – 24:33 so the people with information should be trading, 24:36 and they should be buying and selling shares of markets, 24:38 but they should not necessarily be buying and selling vote coins at all. 24:43 In fact, I would imagine that this would be a little bit like mining 24:46 where it would be something like drudge work that people would just do to get paid, 24:52 which they wouldn't really care very much about. 24:55 The outcomes themselves at all. Does that make sense? 24:59 Well, yeah, but you don't have to do data entry if you own the data. So if you're the entity that is the first publisher or the publisher of record of a particular entity – or sorry, of a particular bit of information, then you could automate the voting of this. 25:19 So again, I ask, if you were running a racetrack, why wouldn't you be the one that's reporting on which horses are winning? 25:33 Okay, I understand. So ultimately this will be something like, what was the stock price of whatever as reported by Google Finance or something. So you're asking, why doesn't Google just do it? And they might do it. 25:47 The answer would be, why haven't they done it already? And they may do it. And actually, on the website I have an FAQ, like a weaknesses section, like a threats weaknesses section. 25:59 And I think the second one that I list, which is that this project might inspire widespread use of legal prediction markets, which is a failure case that I accept. I'm totally fine with that. 26:15 I would be thrilled if Google would start offering this. And as you say, you wouldn't need anything other than multi-sig payment channel, right? Because someone could just open the channel with Google to bet on whatever this is. 26:30 And then Google could just sign it or commit to signing it later as a third party. And then you wouldn't need anything. You would literally need nothing at all except for that. You would only need Bitcoin and one lightning network spoke or something. 26:45 So the entire thing could be replaced by whoever's doing the data. But that would be centralized. And I think most people are doing decentralized stuff kind of for no reason. But I don't think that this project falls into that category because there's a long history of trying to make these things work when they haven't been allowed to work. 27:11 And you'd think – this is one thing that I thought when I first got into Bitcoin. I thought, okay, now someone will do this. As we've seen, there are many websites, Farallay. There used to be bets of Bitcoin. There's bidbet.com. 27:26 So I thought, okay, someone will just do Intrade and they'll just do it with Bitcoin as this rail to get in and out. You just deposit and maybe they'll just convert it to dollars for you or just keep it in Bitcoin or something. 27:38 But the exit scam is too – what I thought I think is too insurmountable a problem for it to scale. Because once it scales, once it gets big and you have like a tiny mini Wall Street at some company where there's some value there that is in the millions or billions. 27:59 There's enough value on one website for one person to live comfortably for like 100 years. Once you've reached that, you're in this threat zone. 28:09 And so I think the exit scam – but even more than that though is that it seems that many people don't trust. They're not willing to – they have a lot of – they have a great desire to participate in this process where they want to move their Bitcoins around and change them into assets and wagers and other things. 28:29 But they just don't – they don't want to go to some website and just send their Bitcoins away and get a SQL row in return or something, an HTML number. 28:41 They don't want to send their Bitcoin away. So that's kind of why I don't – but you're completely right. 28:48 If someone would – the entire project is really the oracle. So if someone would be the oracle, that would obviate the need for the project. And I'm totally okay with that and I would prefer that to happen. 29:03 But I think we can all say a similar thing about – because this is stored value. It's not – this isn't really even like Silk Road or something where if it gets shut down, you can just start another one up or something. 29:14 You only lost the value of whatever transaction you were planning on making. 29:19 With these, the value is stored. So you could store up a lot of money and you could be trading and earning money in there and they can all just be wiped out and you're back to zero or something. 29:28 So it is very sort of high stakes and you could say that – you could make a similar argument basically that why don't we just go have Bank of America run Bank of America coin and it could be exactly like Bitcoin, except it could just be signed by them instead of by miners. 29:47 That certainly wouldn't be my argument, but I appreciate that we want to strive for decentralization. 29:57 What I'm concerned is that this quickly, for economic reasons, devolves into essentially a centralized solution because if everybody's getting information from the same place, 30:16 if there's economic incentive to regurgitate the information from that same place, then that economic incentive could also be inured by the original publisher of that information. 30:34 This segues into my question. They just have a head start in the automation of the Oracle process. So what are your thoughts on automating some of that Oracle? 30:43 Are you going to be providing sort of machine readable bets where you're kind of encouraging people to run automatic Oracle? 30:53 It's one of the first questions that I thought about when I was designing this. 30:57 And I think the conclusion that I reached was it doesn't actually harm the decentralization if you – and I call this kind of re-centralization. 31:08 So if you have the decentralized layer to recover to, it doesn't matter. You really only gain, I think, convenience and efficiency by re-centralizing kind of on top of it. 31:20 So everyone might go to Google to get this thing, and that would kind of make it seem like Google was providing the sole source of the data or something. 31:33 But if some censorship force told Google that they couldn't – you can't provide this data anymore, then people would be able to recover from that. 31:44 It wouldn't destroy the thing. They would just say, okay, we can't get it from Google, but we'll just get it from anywhere else. 31:52 We'll get it from Yahoo, or we'll go to like New York Stock Exchange.com or something. 31:57 And so I think it makes a really big difference. Let me see how I want to explain this. 32:07 I'm totally fine with the entire thing being computerized except for the fact that it populates a big list and one human being – not one, one per user or something, 32:20 but it's just kind of manually checking it through to make sure that they're all right. 32:25 And the way that I achieve that is that I give, as you may have noticed, I give people an incentive to throw other people off. 32:32 As long as you don't throw more than half the people off, you want a few of them to actually be off because you benefit if they screw it up. 32:40 So you actually have a strong incentive to like double check the results. 32:46 And that's all I really want because the way I see it, the decentralization is just a means to an end, which is surviving the censor. 32:55 And so I say if you've got something that scrapes Twitter and scrapes Google and scrapes whatever and just like helps you fill in the ballot, 33:07 I say I think that's great, and I think that could be encouraged to a very great – a very strong extent 33:17 because if something goes wrong and people say – people can claim, well, the software is accidentally saying that it's three when it should say 30 or something, 33:29 but we're going to leave it as three because we think other people are going to do that. 33:33 We think other people are going to do three. 33:36 There's actually an incentive to say that but then secretly go behind the scenes and change yours to 30. 33:41 So no one will ever really know if everyone else is actually leaving it as three or they're changing it to 30. 33:47 And so I don't see it as a very strong – it's not as dooming. 33:53 I know what you're sort of saying is that this is kind of – isn't it centralized? 33:57 But I think it's – that's something that I thought about a lot and you – 34:01 Well, it sounds like the centralized systems are flawed in their scalability where once they get big enough, economically they fail, 34:08 whereas this system can scale economically much better. 34:11 That's the idea is that you've really – you really shift the blame around a lot 34:17 and you try to get the – you try to get the miners to sort of – the same way the Bitcoin works where it's sort of like – 34:25 it's a very slow kind of expensive gradual accumulation of this proof of work. 34:31 And so it's a lot easier to kind of see what's going on and you don't need to – it's not like – 34:37 it's not like the servers are raided and everything is shut down or something. 34:41 It's just like very organic and kind of liquid. 34:45 And so you need to go back 10 blocks or something. 34:49 You can kind of do that and if you – 34:51 The whole network won't fail. 34:53 It's just departmentalized so only one part might go down. 34:56 We have a question from Taylor actually. 35:00 Can you explain the payout mechanism? 35:02 Like how does the system guarantee its solvency? 35:06 Do you mean for – 35:08 Like paying off the bet. 35:10 Where is the money? 35:12 So there's this thing called a market scoring rule which happens to be very convenient. 35:18 Although since the invention of the Lightning Network, I'm going to add a little bit onto this, 35:23 but it's basically the same thing. 35:26 There's this thing called a market scoring rule where all you have to do is edit. 35:31 The market exists in some state. 35:34 So if it's a simple binary market, yes or no, Hillary Clinton will be elected. 35:41 There will be two states, one for yes and one for no. 35:45 And the market starts with some account. 35:48 So you have zero yes shares, zero no shares, some account that has money. 35:52 And if you want to edit the state, like let's say you wanted to buy two yes shares, 35:59 you can just say, okay, what should the market account be at zero, zero, 36:04 which is what it is today, and what should it be at zero, two? 36:08 And the formula will just give you what the market account should be, 36:12 and then you just pay the difference, and you just update the market state. 36:16 So it's just this totally atomic process that's actually very convenient 36:20 because it's very, very similar to what each Bitcoin transaction looks like. 36:26 They're just like one atomic thing. 36:31 And so this lets everyone just – if you want to sell the shares, 36:38 all you have to do is edit whatever it says, 40 shares to like 300 shares, 36:43 and there's this much money in the market. 36:45 You just say, okay, instead of 300, I want it to go down to 298, 36:51 and then you pay your two shares, and then you get the cash back. 36:55 So this concept of a market scoring rule is actually immensely convenient. 36:59 But the other cool thing it does is passively it issues the shares. 37:03 So what I'm planning on doing is setting it up so that ultimately this can evolve 37:09 into something where big players just buy up a lot of all the states 37:14 so they'd pay like $10,000 to just get $10,000 of state one 37:19 and $10,000 of state two. 37:21 That's $10,000 or whatever, $10,000 Bitcoin or what have you. 37:27 It doesn't really matter, of course. 37:29 But then they'd just have $10,000 of each, 37:31 and then they would be able to open payment channels with people 37:34 because they'd have $10,000. 37:36 That's how you get from zero issued to $10,000 issued without things like spam 37:40 and kind of confusing. 37:43 Zero to $10,000 or something. 37:45 Yeah, and then someone could just open a channel with them, 37:48 and they say, okay, I preemptively bought like 20 of each or something, 37:53 but then just keep editing that back and forth. 37:55 So that's kind of – 37:57 Just the house. 37:59 So those people would be like – so someone would create this synthetic thing 38:04 like Bitgold or whatever, and then these other places would be like different exchanges. 38:08 There's like the BATS exchange, and there's like IEX, 38:12 and there's all these different exchanges. 38:15 So Wall Street is not just one exchange. 38:18 It's many exchanges, and they all have different rules, 38:21 and some of them hate each other and stuff. 38:23 So these would just be like kind of – that's kind of the new vision. 38:26 But to answer the question, which I didn't really get around to answering, 38:30 it's just this scoring rule, it just prevents – what it does is it has a bounded loss, 38:34 and I force the guy who creates the market to pay all that up front. 38:38 So he gets money if the market is popular. 38:41 He gets a tiny, tiny cut, but he's got to pay up front to create it. 38:45 So it's an anti-spam, and it solves the bounded law. 38:49 And you know, a lot of times it doesn't even hit the bounded loss, 38:52 so the person will get a refund. 38:55 But it's just a formula, and it can't lose money. 38:59 It's this formula based on like diffusion of – it's from like physics. 39:03 So it's not that – you know, it's not like – 39:07 so each state is updated with this rule. 39:11 And so there is no like – there's no real storage of money literally, 39:17 but in an abstract sense you have bitcoins in this sidechain system, 39:23 and then you update, update, update, 39:26 and you just end up with more or less cash based on the updates. 39:31 And I don't know if this – I didn't make this clear. 39:33 I have a huge Excel spreadsheet. 39:35 I don't know if anyone even in the bitcoin world enjoys Microsoft Excel at all 39:40 to the extent that I did when I was working in consulting. 39:43 But the partial derivative of this account function 39:51 that relates the total outstanding shares to the amount of money, 39:54 that is just – that determines the prices. 39:57 So as you buy more of one than the other, 40:00 you are knocking the price around at the very same time. 40:04 So if you bought like a million of – 40:07 let's say there's like 13 of state one outstanding, 40:11 and there's like 20 of state two outstanding. 40:14 State two would like probably cost – it would cost more than 50 cents on the dollar. 40:18 It would cost like probably 54% or something or whatever. 40:23 But if you then bought like 10 million shares of state two, 40:29 it would cost you about $10 million or 10 million units 40:33 because you get the first few slightly for cheap. 40:37 You get the first few at like 60 cents or 70 cents, 80 cents, 40:41 the first two or three based on the trades – 40:45 the money that had already been put in by people who had made trades. 40:48 But then as you had knocked the price in this – within this one trade, 40:53 you would knock this price to the value of one. 40:56 And so you're just buying shares from yourself. 40:59 So you just sink like a million dollars. 41:01 It's going to cost you like a million dollars to buy a million shares. 41:04 And you didn't really do anything with a lot of – 41:08 with about 9,900,000 of those. 41:11 You just – the price was at one, and you just – 41:14 what you really did is you pointlessly put a huge amount of money at risk 41:17 because on the off chance that it doesn't happen and state one wins, 41:21 you lost all of your money for really no reason. 41:24 And that would be a very – and then people would be buying in, 41:27 and they would be buying in at around zero. 41:29 So that would be a very interesting situation to land yourself in. 41:32 But the formula is documented, and I didn't come up with it. 41:37 Other people, actual guy who has a – a brilliant guy who has a PhD in physics 41:43 and in economics came up with it like 15 years ago and lots of peer review, 41:49 and it's not that complicated. 41:51 It's very clear that it has a bounded loss, 41:53 and you can check in Excel by just typing in values all day, 41:57 and you'll never get it below. 41:59 So the bounded loss is very small. 42:01 And you can change the value as much as you want. 42:03 You can make the bounded loss like 10 cents. 42:06 It's directly related to the amount of liquidity you want in the market. 42:10 So I would imagine that most people would set it very low 42:13 and then just force other people to deal with it. 42:16 But if there's a need to make the market more liquid, 42:20 you can actually just pay for that. 42:22 So that was a big – that was a lot of luck involved with how conveniently that worked out. 42:27 So that is a lot more convenient. 42:30 It depends on how much liquidity you want to offer in your own bet. 42:33 It does. 42:34 If you want to run a bet for, say, your own thing, 42:36 then, yeah, you have to offer up some liquidity to it. 42:38 Absolutely. 42:39 You're absolutely right. 42:40 That's something that a lot of people often overlook 42:42 is that these markets are a little different from – 42:46 they're a little different from normal markets where there's a use case for the actual thing, 42:51 like something like oil where you need it to heat your house or move your car or something. 42:57 You can just kind of sit back and wait for the market to just form. 43:00 And the market will just form, 43:01 and then everyone will be able to use the price to make their capital investments, 43:06 and everyone will be very happy because we'll have a price system and everything will be great. 43:10 And so that's great. 43:13 But it's something like where there is no real use for the thing. 43:17 I mean oftentimes Intrade had success with the political futures 43:24 because they were very entertaining, 43:26 and I suspect because a lot of people wanted to kind of cheer on someone they liked 43:31 or maybe even hedge against someone they didn't like. 43:34 I've always thought – I had like a crazy uncle who I thought was doing that where he would just – 43:38 he hated whatever candidate, and so at least he could make a lot of money if the candidate won. 43:43 I thought that was kind of what he was doing for a while. 43:45 He was kind of a nut. 43:48 But basically there's like a lot of entertainment kind of value. 43:51 But if you don't have this value somewhere, there won't be really a big need to trade. 43:58 It depends on all kinds of factors, right, 44:01 and how people's preference for like risk and stuff, 44:03 and how many people have the information that you want. 44:07 But yeah, you're right. 44:09 A lot of times these will need someone to front. 44:12 Sometimes in one extreme, a very common extreme, some organization will want to buy. 44:21 This is what I really wanted to do with Bitcoin's block size. 44:24 I wanted to have just people put up a lot of money so that someone could just buy the information they wanted, 44:30 which is like the answer to some question, like should we fire the CEO, 44:34 or should we change the block size, or should we elect a different president or something. 44:40 So I wanted someone to just put up this money so that there would be this big lure to like really start everything going. 44:47 So if you want to buy information, you can do that here. 44:50 And if you don't pay enough, you run the risk of the market not being successful, which is unfortunate. 44:56 So just as a follow-up, that sounds like it guarantees the solvency of the bet, but how do you guarantee the pair? 45:03 So I just – I jam the final prices. 45:07 So the – sorry, there's this thing on my computer that's making a lot of noise. 45:12 But the solvency is guaranteed. 45:19 The loss is bounded. 45:21 If you hit an extreme, you've lost the most you can lose. 45:27 So obviously, if you don't hit an extreme, you hit somewhere in the middle, and you're sort of happier. 45:33 So after the oracle, the voting procedure is done, those people have determined what has actually happened. 45:43 From there, you get the final – you get like whether or not Hillary Clinton actually did win. 45:48 And there I just jam the prices. 45:51 I just say, okay, once we've reached a certain – once the market has reached a certain time, a certain block number or whatever, it's matured. 45:59 And you can now – instead of selling, you just withdraw. 46:04 And you withdraw – the prices no longer vary as people bet. 46:09 They're just jammed at the correct value, and then everyone can just withdraw. 46:16 Does that make sense? 46:18 Like it's just – the loss is bounded. 46:21 So no matter what happens, you can't lose more than – you can't lose any money. 46:25 And you – whether or not – if Hillary Clinton is elected, all the yes shares will be redeemable for one. 46:32 And if not, all the no shares will be redeemable for one. 46:35 And so you ultimately – instead of having prices that are a function of this rule, they're just a function of – they're instead a function of that voting process. 46:46 That happens on a different track. 46:48 So all the market stuff is – yeah. 46:50 They're redeemable for one of the coins that are part of your separate blockchain? 46:56 Yes. 46:58 Yeah, so this is a sidechain, and there's a lot to say about that. 47:03 But yeah, it's just a bunch of different rules for manipulating Bitcoin. 47:09 So you would end up with – and then you would – so once you had these Bitcoin, you could theoretically withdraw them to the Bitcoin main chain, not a sidechain. 47:22 You could have it back in the Bitcoin world if you wanted to do that in theory. 47:27 And I don't know if you are aware, but I did something else called Drivechain, which is about that. 47:33 And there's a lot to be said about sidechains and other stuff about that. 47:40 It's very interesting stuff. 47:41 But yeah, but this is – the idea is just that this is – everything that Bitcoin is, but it just adds more message types and obviously more – a more complicated database. 47:55 But yeah, that's basically it. 47:57 Some people might argue that I guess the project that you originally helped out with, Augur on the Ethereum network. 48:04 Ethereum is far better suited for a prediction market. 48:08 I guess could you give us your experience of working with the Augur guys, kind of how that whole thing played out? 48:15 So I published this idea a while ago, and a lot of people were interested in it. 48:21 And Augur, what eventually became known as Augur, they were some of those people who were interested. 48:30 And they were kind of – long story short, they were about – at one point, there were about four different teams. 48:38 And I didn't think the Augur team was sort of the best team. 48:44 I thought that a different team was better. 48:47 And I was just working at Yale as a researcher, and I was not really working on this. 48:53 It was just kind of like a spare time thing. 48:56 And so Augur then did a lot of very aggressive kind of things. 49:03 Like they partnered with Vitalik, and they switched to Ethereum. 49:07 And then they did like a big – they planned to do like a big crowdfund. 49:12 And then they changed their name to Augur, and they like made – they hired like a marketing guy. 49:17 And they did like a lot of posters and stuff. 49:21 So yeah, they did a lot of stuff. 49:25 I'm familiar with Jeremy from – they're a former marketing guy. 49:28 He was an interesting guest on the show for sure. 49:31 Yeah, they're very interesting. 49:35 Yeah, the real problem is not – so yeah, they're really nice. 49:40 And they're smart guys. 49:43 But there are kind of a lot of problems with – one of the big problems with Augur is that the people working there didn't totally – 49:53 I mean it's really complicated, right? 49:55 Like imagine how long it took for people to understand. 49:58 I hate to force the parallelism, which is really unfair. But it took people a really long time to kind of understand Bitcoin. 50:04 And there were all kinds of people many months into the creation of Bitcoin who were saying like – who are today's experts. 50:13 But at the time, they didn't really know kind of what was going on. And the Augur team was just kind of learning it a little too slowly. 50:21 And they really wanted to make lots of changes. You know how this is very common in Bitcoin as well? 50:27 Again, where people hear about it for the first time and then they say, okay, that would be great. 50:32 But let's just change the deflationary thing or let's just – why do we have it? 50:37 Like we should add know your customer to this or something. 50:42 So they hear about it and they kind of want to make this – people get inspired. 50:44 They really want to make – add their own thing to it even if their own thing is like something that you thought of like two years ago and this is not a good idea. 50:52 So they were kind of like – but they kind of did all this stuff very quickly at the beginning of 2015 or whatever, 2014. 51:02 I don't even – I lose track of the years now. But they started to do this stuff very quickly. 51:08 And I had – the other thing, the weird thing was that the person who was the most productive volunteer, he didn't want me to reveal he or she. 51:19 I have no idea. They didn't want me to reveal their – I assume it was a he. 51:23 Didn't want me to reveal their identity or that they were working on the project. 51:28 So they kind of did a lot of – they did like all of the groundwork, C++ groundwork. 51:33 They were kind of doing it secretly for a while. 51:37 And Roger Vier contacted me. He heard about the project in late 2014 and we met. 51:45 And he was really interested in it and he started paying this developer, but it was sort of like put me in like a weird position where I didn't really understand. 51:56 The other thing, I didn't really care because – but things kind of got a little out of hand with Augur and they – I don't know. 52:05 Ethereum is really new and a lot of people I know are convinced it won't work for a variety of completely different reasons. 52:13 So I find it very unlikely that Ethereum will still be with us in two years. 52:20 But I don't know. Maybe it lasted this long. 52:23 But a lot of things lasted a while, right? 52:26 MasterCoin. There's like a head MasterCoin. 52:30 I mean Litecoin. Who would have thought that Litecoin would still be here? 52:33 Litecoin and MasterCoin. 52:35 What? I didn't catch? 52:37 A lot of self-deprived Litecoins around, yeah. 52:39 Yeah. So we still have Litecoin. We still have Ripple. 52:44 They said they were closing it down I thought, but it's still there, right? 52:47 It's still on CoinMarketCap. 52:48 So, I mean, it's like weird stuff. BitShares lasted a very long time. 52:52 I don't – you know, it's sort of still there. 52:56 So it's like – these are kind of weird projects, but none have been as big as Ethereum. 53:02 But I just – it seems so – things really kind of – I don't know. 53:10 It's hard to know what to say about that, but I don't think it's – you know, lots of people just openly laugh at Ethereum 53:15 and they stopped trying to correct the mistakes like three years ago, 53:21 and now it's just kind of – we have the talk of Microsoft hosting Ethereum, 53:27 which is like a total contradiction. 53:30 They host Ethereum after a while. 53:32 Like their blockchains are only like 10 gigabytes and it's only been like four months. 53:35 Very, very fast, right? There's that paper. 53:38 See, this is the thing. You don't really know. 53:40 There's a huge chicken and egg problem where you don't know if you're any good 53:45 until you really start to get the actual momentum, right? 53:48 Because – so you don't know if you're any good. 53:53 So once you get big enough, then someone will say, 53:56 okay, I'm going to write a little academic paper about Ethereum, 53:59 because now Ethereum's big, right? 54:01 So then someone wrote that ring of guages or something paper where they're like, 54:07 by the way, you can just do these little computations that like multiply these matrices or something 54:12 and it will just be a huge amount of work for Ethereum to do, 54:17 but it now costs you very much. 54:19 And it was like, how is that? 54:21 But more fundamental to me is that the whole operating premise – 54:25 so this is the thing. You talk to a lot of Bitcoin experts, 54:28 and it seems like many of them have completely different objections to Ethereum. 54:34 So some people will talk about the organization and how there's lots to talk about there, 54:39 and other people will talk about the size of the blockchain, the blockchain growth. 54:47 But I mean, for me, it's just like, what is the point? 54:50 Before Bitcoin, someone would try to prevent you from sending money in this fashion. 54:59 So you needed the money to be sent in a reliable, kind of redundant way like this, 55:07 where you could recover. 55:10 But with Ethereum, I'm like, who is stopping you from doing computations on your own computer? 55:16 No one. So the only thing that – what it must be offering is the opportunity 55:22 to use those computations to move money around. 55:25 But again, you can kind of already do that to a large extent with this – 55:31 you're going down this oracle road again. 55:32 So either you're going to do the entire oracle thing, or you're going to just use multisig. 55:38 But there's a bigger problem, which is that Ethereum can't actually work very well with oracles, 55:43 because you need to – if someone has an incentive today to be dishonest, 55:48 which they directly do, just by being one of two counterparties to any bet, 55:55 they have – you know, if you're on the losing side of the bet, 55:59 or you don't even care if you're on any side of the bet, 56:02 you've got to be able to control the outcome of the bet that's directly worth money to you. 56:07 So the only way you can extract this long-term surplus 56:12 is to have some kind of reputational – some kind of abstract reputation. 56:17 So either you literally have reputation, as in the case of someone like, you know, 56:23 Dread Pirate Roberts, Silk Road, you have some kind of abstraction, 56:27 or you have some kind of feedback or something like that. 56:30 Or you have kind of like what I did. 56:32 You can make like a really, really, really abstract reputation 56:36 that exists like in the form of this corporation. 56:38 But whatever it is, you need to have some way for people to compete on reliability 56:42 in order for the oracle path to succeed. 56:47 But Ethereum is too general. 56:50 You can copy someone else's – you can free ride off of someone else's reputation 56:55 by just making a little thing that just watches what they do. 56:57 So as a consequence, no one can compete on reputation, 57:01 and it's going to be basically impossible. 57:03 I mean, the security margin in this – 57:07 just like the miners determine the security margin in Bitcoin, 57:11 with this, it's the market capitalization of those vote coins 57:14 that determines how much vote coin do you have to buy a lot 57:18 to shake all the outcomes. 57:22 And assuming that nothing else works, you know, none of the miner stuff, 57:24 which could catch that downstream. 57:28 But basically, the security margin – 57:30 so I'm imagining these vote coins would be worth a good chunk of change eventually, 57:35 especially eventually once they've grown. 57:38 But how can they be worth if their value is that they earn money? 57:43 The vote coins are kind of a share in an abstract corporation, right? 57:48 So they're earning – they're paying out shares of revenue. 57:52 So how can they – how can it be worth, like, I say $1 million 57:58 if someone else can just copy that, you know? 58:01 They just write a little piece of code that just copies that. 58:04 That should have – that should be able to capture – 58:07 it should be able to undercut this thing, 58:09 and it should – it can guarantee that it's going to perform just as well. 58:12 And so, long story short, I mean, this should be worth – 58:17 this tiny piece of code should be worth $1 million, 58:19 and anyone can write this tiny piece of code. 58:20 So it – because it will be getting these dividend streams 58:26 just the same as the main vote coin kind of set would. 58:32 So it's kind of – it gets a little technical, but I don't think – 58:36 but I don't – so I don't think Ethereum has a use case other than – 58:40 it can't go down the oracle path. 58:43 So I don't know, you know, what the use case is. 58:47 I've asked people for years to tell me what they would use Ethereum for, 58:51 and for a while I got back nothing, 58:53 and then people would tell me they use it for Augur, 58:56 which is my own project, securitously. 58:59 So – but people, you know, have really never given – 59:02 the use cases that are featured on the Ethereum website are, like, not – 59:08 you know, they're not serious. 59:10 These things about little locks and stuff that you are not – 59:12 I mean, it's not – these are not, like, you know, 59:18 this would be much better done. 59:20 Almost everything that Ethereum does would be much better done not on Ethereum. 59:26 Just, like, there's no reason to decentralize almost everything that is there. 59:31 And in another sense, none of it is decentralized, 59:35 because if law enforcement disagrees, 59:38 they can just, you know, bring bolt cutters 59:41 and just cut the lock or something, 59:44 and it's, like, there's no – you know, it's very – 59:47 I just don't get it, personally. 59:49 But you talk to a lot of different people, 59:52 they'll give you a lot of weird problems, 59:54 and there's a lot of burnout in the, you know, 59:57 the Bitcoin wizards or whatever you want to call them, 59:59 the people who know a lot about – 1:00:01 they're just, like, tired of explaining this stuff. 1:00:04 They've been explaining it since 2009 or something, 1:00:05 so they just don't even bother talking to anyone anymore. 1:00:08 And so now you kind of – it's kind of just like – 1:00:12 so I don't – you know, there's a lot to say about Ethereum. 1:00:15 We could talk about it all night if you want. 1:00:17 One thing that Ethereum can't do is buy beer from our vending machines. 1:00:21 I don't know who here wants a beer, guys. 1:00:23 I still got some money from our guests last time. 1:00:26 Maybe I'll get a round. 1:00:28 All right. I'll send it off here. 1:00:30 And yeah, you should check out our machine here, Paul. 1:00:34 You have to turn me around. 1:00:37 You can have the light go off at least to show that it's working. 1:00:42 So we have an old vintage testing machine here. 1:00:46 Oh, cool. 1:00:48 And we've packed it with a Raspberry Pi to basically accept – 1:00:54 why is it not doing anything now – 1:00:56 to accept Bitcoin payments, 1:00:58 and yeah, it's been working splendidly for us. 1:01:01 I've got to thank the guys in the room here for setting it up. 1:01:06 Pi is still on. 1:01:09 No, Pi is still on. 1:01:11 It should go. 1:01:13 Yeah. 1:01:15 We'll see how fast blockchain is. 1:01:17 And Paul, while he's doing that, I've got a quick question for you. 1:01:21 You mentioned Augur and whether or not you believe in Ethereum and Augur and so on. 1:01:27 You have to agree that if there's any value to the world in prediction markets, 1:01:32 and I certainly believe there is, 1:01:34 that there will be multiple prediction markets out there. 1:01:40 Has there been any talk about standardizing any kind of formats 1:01:46 for interacting with them as far as oracles are concerned? 1:01:52 It would seem that there might be some advantage 1:01:54 to being able to upload information to both Augur and Hivemind 1:02:05 and also to whatever other sites are out there 1:02:11 so that when an election is won, 1:02:13 you can put your reputation on the line on all the different prediction markets. 1:02:21 And I imagine that it might be possible to make bets 1:02:30 between the different prediction markets 1:02:33 in order to kind of tighten and solidify that common information thing 1:02:40 so that if you wanted to make a bet on Hivemind 1:02:44 that a particular bet on Augur 1:02:48 or a particular oracle reporting to Augur was accurate. 1:02:53 By interlinking these various things, 1:02:56 you could get arbitrage happening 1:02:58 that would tend to very quickly reveal 1:03:04 when there's a systematic problem with one of the underlying networks. 1:03:08 Yeah, I don't know about that. 1:03:11 I mean, it's an interesting idea. 1:03:13 So first of all, as I tried to explain earlier, 1:03:17 the security margin is the market capitalization of the vote coins. 1:03:21 So if you copy that 10 times, 1:03:24 all you've done is decreased security by a factor of 10, 1:03:28 but you haven't like – you don't make up for it 1:03:33 by the fact that there's 10 different systems. 1:03:35 It's just 10 systems that are all one-tenth as secure. 1:03:40 So I'm not sure. 1:03:43 It's hard to say. 1:03:45 I think there definitely will be a lot of – 1:03:47 there will be more than one service like this, 1:03:51 but I'm hoping that it plays out a slightly different way. 1:03:53 And so I'm hoping that – I'm not hoping, 1:03:56 but I kind of expect that Hivemind would be used 1:04:00 for things that are very sensor-attracting 1:04:05 or something kind of like these things that I think are – 1:04:08 I think will be kind of really important. 1:04:11 This sort of like president versus GDP, 1:04:15 CEO versus stock price kind of really kind of a lot – 1:04:20 that potentially embarrass a lot of people, markets. 1:04:23 And I think that a lot of other stuff will be done. 1:04:26 Like I think sort of like short-term sports betting will be done. 1:04:30 There will be like prediction market version of that 1:04:33 and prediction market version of like other more entertaining things 1:04:35 I think will eventually be kind of on a website 1:04:38 or on like a smaller system of some kind 1:04:42 where it's just – it's not as serious. 1:04:45 But I think for this, like if you want – 1:04:47 like I'm imagining this could contain 1:04:50 like some kind of synthetic Dow Jones Industrial Average 1:04:53 in the U.S., like U.S. Stock Market Index. 1:04:57 And I think that that – I'm not sure. 1:05:00 I mean something that serious is going to – 1:05:02 I think it's going to want to converge to one thing 1:05:06 that is the thing. 1:05:09 I think not necessarily network effect of money as with Bitcoin, 1:05:12 but there are still other smaller network effects 1:05:16 with like respect to developers and what you might call 1:05:25 like human capital knowledge and stuff like that. 1:05:28 Like much easier to just learn one. 1:05:31 So I think there might be – it's hard to say completely, 1:05:37 but I don't think it's – I don't think that you'd want to spread. 1:05:45 See the thing is what you can already do on Hivemind 1:05:48 is you can already create lots of different markets 1:05:51 in different ways. 1:05:53 And you can – you create these different decisions 1:05:55 and then from there the decisions are the bottlenecks. 1:05:57 You can create markets that are very liquid from there, 1:05:58 markets that use a lot of decisions. 1:06:02 You can do lots of different things kind of within one already. 1:06:05 So you would get a lot of differentiation 1:06:10 even within one system. 1:06:12 So I'm not sure given the risk that the system would collapse, 1:06:15 which I think is high even for one or two 1:06:21 because it's very experimental. 1:06:24 But I don't necessarily think it would play out totally like that. 1:06:26 I think it would be slightly different. 1:06:28 I think there would be more of this kind of log distribution 1:06:31 where there would be like one big one 1:06:33 and then maybe sort of like a sub one. 1:06:35 And then like a lot of that are just basically centralized websites 1:06:39 that are just sort of for fun or shorter term type of thing. 1:06:43 What's your opinion on the I guess upcoming election 1:06:47 that's going on in Bitcoin in regards to the expanding, 1:06:52 doing a segregated witness 1:06:54 or just going straight to bigger blocks 1:06:56 like hard fork versus soft fork? 1:06:58 Well, I waited a very long time. 1:07:00 But I did sign. 1:07:02 I just signed like four days ago, 1:07:05 Greg Maxwell's scaling roadmap. 1:07:08 So I guess I endorsed. 1:07:10 I waited a really long time. 1:07:12 I read it very carefully because it's become quite the issue, right? 1:07:15 It's the talk of the town. 1:07:17 But I guess so I couldn't really find anything I really disagreed 1:07:21 with seriously in his thing. 1:07:23 So I signed it. 1:07:24 But I think it's very clear that the debate is not really about 1:07:29 anything other than control over Bitcoin. 1:07:34 I don't think it's really about anything 1:07:37 because I don't know if you read my blog or something. 1:07:42 I've already proposed like three different ways of solving this 1:07:46 like in a kind of a Pareto improvement 1:07:48 like such that everyone can kind of get what they want 1:07:51 or no one is worse off, 1:07:52 but other people can still do very well. 1:07:55 So I've published a lot of things 1:07:58 and tried to advocate privately and publicly 1:08:03 for these like win-win solutions, 1:08:06 but they never really catch on. 1:08:08 And I think that's because no one really wants to make everyone happy. 1:08:12 I think I'm the only maybe the only one who wants 1:08:14 like the only one in the conversation who wants that sometimes. 1:08:17 And other people I think they just want to win 1:08:18 and they want to kind of assert dominance 1:08:23 or protect their turf or do some other thing 1:08:26 that is not related to. 1:08:28 Because I mean I've done this like so many times for months 1:08:31 since like May of 2015. 1:08:35 And since it like became an issue 1:08:38 and no one seems to be interested in doing those things. 1:08:45 And the sidechains I think are the weirdest example of that 1:08:50 because you could have these things as a sidechain 1:08:54 and everyone could get what they want, 1:08:56 but no one wants to do that at all. 1:09:00 It's really hard to say that it's I no longer believe 1:09:08 that the debate is really about anything other than. 1:09:11 And it's ultimately of a very low consequence, right? 1:09:14 Because even if you know other than the principle 1:09:17 which is of tremendous consequence. 1:09:19 So some people say on principle Bitcoin can never hard fork 1:09:22 or it can never do anything. 1:09:24 You know the rules are set and they're not being changed. 1:09:31 So some people say that and so obviously in that light 1:09:34 it makes a huge deal. 1:09:36 But I think ultimately I mean think about it 1:09:39 like the miners can just set the soft limit back down to one megabyte. 1:09:42 They can soft fork back down. 1:09:44 So I mean and then if there's a hard fork all the users, 1:09:48 all the investors, they'll have two copies of each coin. 1:09:53 So ultimately it's not, it's really not. 1:09:57 It's really I think it's blown up 1:10:00 and I wrote about this very recently. 1:10:03 It's you know it's blown up and it's made a lot of people angry 1:10:07 and there's people have become very political 1:10:10 and there's been no recovery from that. 1:10:14 There usually isn't in Homo sapiens 1:10:17 and there hasn't been here. 1:10:20 So that's kind of, it's very hard for me to believe 1:10:23 the debate is actually about anything anymore than just. 1:10:26 Well it's true Paul and that I think leads to a good question. 1:10:32 As you intimated it's not about the specific issue 1:10:36 so much about you know people wanting to win 1:10:40 but maybe even more importantly 1:10:43 it's a fundamental question about the governance 1:10:47 of these distributed systems 1:10:50 and how to manage the governance in a decentralized way. 1:10:55 So I'm going to ask you, 1:10:57 what is your vision for Hivemind's ultimate governance? 1:11:02 If it takes off and becomes the de facto prediction market, 1:11:07 decentralized prediction market, 1:11:08 who decides which versions of the code, 1:11:14 you know 2.0 and 3.0 run down the road? 1:11:19 Are you looking at building anything into the actual system 1:11:24 to have people vote in a decentralized manner 1:11:28 for the type of code that makes these future decisions? 1:11:34 Yeah well it's interesting, 1:11:37 I actually, there is an idea for that. 1:11:40 So the first thing I published was a blog post 1:11:45 that I called the win-win block size solution 1:11:48 which was about using prediction markets, 1:11:51 a very clever one in fact, 1:11:53 where anyone who wanted a hard fork, 1:11:56 they could buy these coins that would give them, 1:11:59 you know it would give them today's exchange rate, 1:12:01 and you could basically get a refund 1:12:03 if it didn't hard fork, if Bitcoin didn't hard fork, 1:12:06 you get today's, excuse me, today's price. 1:12:09 You do the same thing if you didn't want the hard fork, 1:12:14 you get a refund if it did hard fork, 1:12:16 but you could keep your Bitcoins if it did not 1:12:22 or whatever it was you wanted. 1:12:24 And there was like a lot of clever math there 1:12:27 and it was a really neat idea. 1:12:29 The weird thing is, 1:12:31 that would have used prediction markets to, 1:12:34 so as a side effect I didn't even bring up, 1:12:37 as a side effect you'd end up with two prices 1:12:40 and I would imagine that one of those two prices 1:12:42 would be slightly higher than the current exchange rate. 1:12:47 See what I'm saying is you'd have, 1:12:49 you'd have basically what I call fork futures. 1:12:51 You could buy one type or the other type of Bitcoin, 1:12:55 but you wouldn't need the, you wouldn't, 1:12:57 you'd be able to recover completely 1:12:59 if you bought the right type. 1:13:01 So if you bet on hard fork 1:13:03 and the hard fork didn't happen, 1:13:05 you'd win money and you'd win enough 1:13:07 such that you could cash out as the, 1:13:10 at the today's exchange rate, 1:13:12 no matter what happened with the exchange rate in the meantime. 1:13:14 So you have these, basically these futures on the forks 1:13:19 and that was like, that was like a really cool thing 1:13:22 because everyone could just buy 1:13:24 and sell exactly what they wanted 1:13:26 and you'd, as a side effect of this, 1:13:27 you'd get two prices. 1:13:29 You get the price of, you know, Bitcoin original 1:13:32 and the price of Bitcoin hard forked. 1:13:34 And my, my total, my 100% expectation 1:13:37 would be that one of those would be 1:13:39 slightly higher than today's exchange rate 1:13:41 and the other one would be like near zero or something 1:13:44 because I would imagine that 1:13:46 there would be some expectation 1:13:49 on which of the two would survive 1:13:51 if the fork actually happened. 1:13:54 And so, long story short, you could, 1:13:58 you could like, 1:14:00 within this prediction market environment, 1:14:02 everyone could kind of get exactly what they wanted 1:14:04 or they could trade until they were, 1:14:06 had reached that point and, 1:14:08 and that would enable prediction markets 1:14:10 to help Bitcoin decide what to do. 1:14:12 But it would not, I don't think you could totally use, 1:14:15 you know, it would be like operating 1:14:17 on your own brain or something. 1:14:19 I don't think you would be able to, 1:14:21 I thought about it, but I don't think Hivemind 1:14:23 would be able to predict markets on itself 1:14:25 because it would have no way of knowing the outcome 1:14:27 after the hard fork because the hard fork would, 1:14:29 you know, potentially, you know, 1:14:31 it was like kind of, 1:14:33 you reach like an unmeasurable zone. 1:14:35 But I thought about having like a tiny, tiny, 1:14:38 this is the original proposal 1:14:40 was just to pick multi-sig people, just pick, 1:14:42 you just elect seven people 1:14:44 and their job is only, 1:14:46 only to report on the exchange rate 1:14:50 and whether or not you forked. 1:14:51 Thank you very much. 1:14:52 The exchange rate and whether or not you forked, or, and, or whether or not you did other things, but you just, they just have to report like two numbers, basically, a one or a zero, and then the exchange rate in dollars or whatever. 1:15:03 And if you did that, then you would have, you could have these markets, and I think markets are much better than I, you mentioned voting. I'm not a huge fan of voting, 1:15:10 even with money, even when it's weighed by money because there's nothing at stake for a bad vote. So you, it's a plutocracy I've described in the past as, so having money bet is the, having money vote is a plutocracy, right, rule of the wealthy. 1:15:28 And it puts the same wealthy people in charge no matter what the decision is about, whereas this prediction market thing, it allows people to really lever up and bet if they happen to know a lot, and it allows them to stay out of it if they don't know a lot. 1:15:43 You could have a really rich guy who just doesn't know anything about manufacturing two-by-fours, and so he doesn't buy a, he doesn't invest in a two-by-four factory or something because he doesn't know anything about that. 1:15:55 But you can have a sort of a poorer person borrow money or go to a bank or something and say, I know everything about this. So markets I think are much better. 1:16:08 That makes sense. But when you're talking about an underlying system, the whole proof-of-stake concept essentially is a plutocracy, but that's arguably a good thing because the people who have the most money at stake in a particular system have the most incentive to have that system have integrity. 1:16:35 So if you're talking about using voting for governance, I think that's a little bit different than using for governance of how a system runs. I think that's a little bit different than using voting for Oracle purposes because there's obviously a market incentive for somebody to cheat in the case of an Oracle situation. 1:17:06 Reputational issues aside, if there's a wager that they can bet, if the cost of dissenting is smaller than the gain from that cheating, then there's an economic incentive to do that. 1:17:21 But if you're talking about the fundamental trustworthiness of an underlying system, take Bitcoin as a simple example, there's clearly a disincentive for the people who have the largest stake in that system to undermine the trustworthiness of the system. 1:17:42 You're completely right. This is very similar to that. The only difference is that – so if you're going to allow people to – so like I'm saying, allowing people to – capitalism is similar to a plutocracy because the rich people have more votes that they could deploy. 1:18:00 But what you allow is you allow – when you have this capitalist kind of prediction markets thing, you allow people to abstain and they can – in plutocracy, you should really never abstain because you risk nothing by voting. 1:18:14 So you could be bribed or you could just have a huge ego and then you would want to vote. 1:18:20 But this is really everything that you want, I think, plus the option to abstain. 1:18:26 And I think something you mentioned earlier about the wisdom of the crowds, and you brought it up and I wanted to bring it up because I had thought about it for a while. 1:18:35 And I think the real reason that prediction markets work is not so much that the truth – you mentioned that the truth rises to the top. 1:18:44 You said that when people have their own money at stake, and that's certainly true. 1:18:49 But what I think really happens is that there's a lot of noise. 1:18:52 And when you ask people to put their own money up, it just filters a lot of the noise. 1:18:57 I think the real thing is that the market ends up at the right place and then it reaches this point where everyone is just too afraid to put their money up. 1:19:07 And so I think that's the very, very subtle difference. 1:19:11 But I think it's more about getting people to shut up more than it really is about getting – 1:19:16 sort of like it just says the conversation has ended because no one is willing to disturb this kind of equilibrium that we're in, which I think is very interesting. 1:19:27 And then the person shuts up. Yeah, it makes a big deal. 1:19:30 Yeah, I write a lot about this in – I have like a whole sequence of papers. 1:19:34 And the first one I kind of try to describe why prediction markets are important and why I'm doing this. 1:19:40 And I write a little bit about – I don't know if you've read it. You might enjoy it. 1:19:43 I write a little bit about how – why a bet is so much better than talking, 1:19:48 and why prediction markets are so much better than a bet because they kind of scale and there's no inherent – 1:19:53 the bet is very confrontational and personal. 1:19:57 There's this example where Mitt Romney offered this guy a $10,000 – Newt Gingrich – Rick Perry a $10,000 bet. 1:20:04 And it ended up like backfiring horribly because he was like perceived as an out-of-touch rich guy last election cycle. 1:20:12 And so all this very interesting stuff happens with the bet even though the bet is immeasurably better than the conversation I think. 1:20:20 But the bet can be very – but yeah, you're right. It's about the bets that are refused. 1:20:26 That's the major thing that the prediction market will show that the bet does not. 1:20:30 The bet – you never get to observe the bets that are refused as some guy. 1:20:34 But as some guy, you can look at the – you can look at in trade or election betting odds or something. 1:20:39 You can say, okay, I know what bets have been accepted out of everyone on the planet, 1:20:44 and therefore I also know what bets are not being accepted. 1:20:49 And that is a big – that's a big difference I think. 1:20:52 So that's a really interesting point. 1:20:56 Prediction markets are not immune to the same kind of sociological phenomenon that you touched on and other ones. 1:21:04 The most obvious one it would seem would be people manipulating prediction markets in order to affect the real world result. 1:21:18 Like if it's just a question of who wins the Super Bowl, well, then that comes down to the players on the field. 1:21:26 But if it's a question about who wins the election, if everyone starts looking at prediction markets 1:21:32 in order to see who the frontrunners are, and if the frontrunners tend to gain more momentum 1:21:40 because other people start wanting to vote for the winner, 1:21:45 then there's an economic incentive for people to manipulate those prediction markets in order to say, 1:21:53 hey, look, see, the wisdom of the crowd says I'm the next president. 1:21:57 You're 100% right that there's this – it's a very strong and very obvious psychological bias. 1:22:07 You don't want to back a loser because if you back a loser, 1:22:12 and there was evolutionary times, you could literally be killed or something. 1:22:15 There's this fear of backing someone who's not a winner that is overpowering. 1:22:22 But the effect that you described is not specific to prediction markets, of course. 1:22:26 So it's obviously – in the United States, Fox News is a pro at this. 1:22:31 They put the guy they want and they say, clearly he won the debate. 1:22:34 Clearly we're hearing great things about him, blah, blah, blah. 1:22:39 So it's not at all specific to prediction markets. 1:22:42 Prediction markets do better as far as the – 1:22:45 but yeah, you're entirely correct that people feel a strong urge, 1:22:48 particularly when voting, to vote for someone that they already – they assumed was going to win, 1:22:53 which is bizarre because you think that their incentive would be to just stay home and let the person win. 1:22:58 But they really want to – you want to join the winning guy 1:23:03 because we evolved our brains long before secret ballots were invented. 1:23:08 So we have no concept of this kind of safety behind the secret ballot. 1:23:13 Assuming the ballot is actually secret, which is like – 1:23:16 it's unclear how long that will remain the case, if it's ever been the case or anything like that. 1:23:21 But you're right about – there's a strong – but there's a more interesting – 1:23:27 so yeah, these prediction markets are usually relabeled as decision markets 1:23:33 that are used specifically to inform a decision. 1:23:35 As I argue in the case of the block size markets, 1:23:40 I say we should pick the one that has the higher market price or that has the one that's not zero 1:23:45 because that's the one that the money says is the best, that the opinion has been aggregated. 1:23:51 The consensus, to use the loaded word, is to go with this fork or something 1:24:00 And so there, there's an interesting question of will people have an incentive 1:24:08 to bet money to screw with the decision? 1:24:12 And to what extent is that sort of type of thing likely? 1:24:17 And the usual answer that's given is that while some people have an incentive 1:24:21 to manipulate the outcome, everyone prefers having more money to having less money. 1:24:27 And another trap door, but yeah, to flesh out that first point, 1:24:31 is that normally the market is going to settle on – 1:24:35 so you mentioned that the elections case, right? 1:24:38 And how the cause can become the effect. 1:24:41 But sometimes two things are just caused at the same time, right? 1:24:45 So sometimes one candidate really is better 1:24:48 and it had nothing to do with the markets or the media. 1:24:52 And then one person, they were just more likely to win. 1:24:55 And given that we live in a world that has media and has people with brains, 1:24:59 you know, why does the person with more name recognition do so much better always, right? 1:25:05 So it's basically the same thing. 1:25:08 There's this fear of backing the wrong person. 1:25:12 And so sometimes prediction markets are just accurately saying 1:25:16 that this is going to be magnified to the extent that this person will win. 1:25:20 And there's no causality at all and it's entirely in effect. 1:25:25 So markets will always be in effect. 1:25:27 Sometimes they'll be a cause, but they'll always be in effect 1:25:29 because they'll always be measuring – as long as this knowledge exists 1:25:33 and the potential to make money exists, they'll always be there to catch that thing. 1:25:39 So no matter what – yeah. 1:25:41 It's like string theory. 1:25:43 Like the act of observation will always change the result. 1:25:47 That's – yeah. 1:25:49 Well, I mean, it's one of the ideas of string theory. 1:25:53 I'm paraphrasing it a bit. 1:25:55 You have a question as well. 1:25:57 Yeah, just back on the implementation side, 1:25:59 where are you storing like the state of the market? 1:26:02 Is this a native app, so P2P that does a discovery via DHT? 1:26:09 Yeah, this is a sidechain. 1:26:11 So it's everything Bitcoin is just with different messages, more messages to be specific. 1:26:19 If so, is there a lot of sort of overlap with OpenBazaar and Obi-Wan, 1:26:24 like the content and the messaging code? 1:26:26 Yeah, that's a good question. 1:26:28 I don't think so, but I don't really know enough 1:26:30 about how OpenBazaar is doing what they're doing. 1:26:34 But I don't know to what extent they – 1:26:37 so this – the goal for me was to change as little as – 1:26:42 this is basic software modularity, right? 1:26:45 I didn't want to change anything about what's now known as lib consensus, 1:26:49 and I didn't want to change anything about Bitcoin 1:26:51 or any of the proof of work stuff. 1:26:53 I just wanted this to be – 1:26:55 it's basically Namecoin's merged mind thing, 1:26:59 is what this will eventually be. 1:27:01 Of course, it's not merged mind, so we can test it. 1:27:04 But basically, it's not going to – 1:27:07 it's going to change as little as humanly possible. 1:27:10 So that's a basic modularity concept. 1:27:13 I don't know enough about if OpenBazaar is going to do – 1:27:18 I'm really – I have a lot of interesting – 1:27:20 OpenBazaar, there's a lot of interesting questions 1:27:22 about, like, storage for them that – 1:27:26 I mean, I don't know – are they – 1:27:28 I don't really know enough about – 1:27:30 sorry, I don't really know enough about how they plan on doing it. 1:27:32 If they – but all I can say is that 1:27:34 I plan on making it very, very, very similar to Bitcoin, 1:27:36 where there are headers and there are blocks, 1:27:39 and that's – the blocks have messages. 1:27:42 The messages get passed around, 1:27:44 and then it's just about getting the messages around, 1:27:48 and then the software will just build a database 1:27:51 in LevelDB or whatever 1:27:53 that will actually be used for everything. 1:27:55 It just won't – 1:27:57 you just need the blockchain to get that 1:27:59 score of the heaviest chain. 1:28:01 And that's – I don't even want to touch that, 1:28:03 because it's a lot – 1:28:05 it's above my pay grade, so to speak. 1:28:10 It's a lot of work goes into that, 1:28:12 and I don't even want to think about it. 1:28:14 If you want to make an offer, 1:28:16 you have to kind of have a server up that's available. 1:28:18 Yeah, no, there won't be anything like that. 1:28:20 No, yeah. 1:28:22 So that's why I wanted to bring up the scoring rule 1:28:24 that that gentleman behind you asked about, 1:28:26 which is that it's just an atomic state update. 1:28:28 So as long as you can reach the network at all, 1:28:32 you just broadcast your message 1:28:34 and the trade goes through. 1:28:36 And then I wanted to bring up that part about Lightning, 1:28:38 because, of course, 1:28:40 again, I copied what Satoshi Nakamoto did, 1:28:42 but it's not – 1:28:44 it doesn't play very well 1:28:46 to have to do that, 1:28:48 to have everyone store everything. 1:28:50 So try to do the same thing. 1:28:52 I just tweaked – 1:28:54 Lightning is very different from this, 1:28:56 but there are a lot of disadvantages, 1:28:58 but one or two critical advantages 1:29:00 to mine with this issuance. 1:29:02 And so I kind of got my 1:29:04 little Lightning-inspired thing 1:29:07 that I'll write about, 1:29:09 and hopefully it's a lot of work, 1:29:11 this kind of stuff. 1:29:13 But I have a lot of advantages. 1:29:15 I wouldn't need to route. 1:29:17 These people would connect to – 1:29:19 directly connect to – 1:29:21 mine would be much more hub-and-spoke, 1:29:23 so they just connect to the hub. 1:29:25 And then if the hub goes down, 1:29:27 they just can't trade 1:29:29 for whatever it is, 1:29:31 a month or whatever it is they did, 1:29:33 whatever the custodial period was. 1:29:36 I guess it won't be. 1:29:38 Yeah, please. 1:29:40 Do spreads converge as markets 1:29:42 get more liquid? 1:29:44 Is it expensive to change a mine 1:29:46 in niche kind of bets? 1:29:48 Oh, well, that depends. 1:29:50 So it kind of depends. 1:29:52 If you set the market-scoring rule 1:29:54 up the way – 1:29:56 the default way, 1:29:58 the prices sum to one perfectly. 1:30:00 So they sum to one unit. 1:30:02 It would be one bitcoin, 1:30:05 but it wouldn't really matter 1:30:07 because you could rescale. 1:30:09 Everything is just passively rescaled. 1:30:11 In finance, only the return matters. 1:30:15 So that doesn't matter at all. 1:30:17 But the prices would sum to one 1:30:19 so there would be no spread at all. 1:30:21 But I know what you mean 1:30:23 because there still is thematically 1:30:25 a spread because 1:30:27 every single trade 1:30:29 is changing the price 1:30:31 in this system. 1:30:33 You are trading against yourself 1:30:35 some for X percent 1:30:37 where X is the liquidity 1:30:39 of the market currently. 1:30:41 So X is a function 1:30:43 of the liquidity of the market. 1:30:45 So if the market is very liquid, 1:30:47 you won't move the price at all. 1:30:49 You'll be able to buy 1:30:51 all the shares you wanted 1:30:53 at 40 cents and there will still 1:30:55 be 40 cents and you'll get 1:30:57 whatever you paid divided by 0.4 shares. 1:30:59 But if it's not liquid, 1:31:02 you'll be drifting against the price 1:31:04 which is unfortunate. 1:31:06 So I guess 1:31:08 the answer to your question is 1:31:10 still yes. 1:31:12 It's always the spread 1:31:14 or what the trader 1:31:16 experiences as the spread 1:31:18 is always going to be a function of 1:31:20 liquidity. 1:31:22 That's what's interesting about these markets 1:31:24 is that markets 1:31:26 gain. They do better with scale. 1:31:28 They're really remarkable. 1:31:31 Everything else degrades with scale 1:31:33 but markets, they love scale. 1:31:35 They want more people. 1:31:37 They want everyone to be in one spot. 1:31:39 So they're very, very interesting 1:31:41 and they're very cool. 1:31:45 It's a sad thing to see 1:31:47 when I go on. 1:31:49 A lot of people contacted me 1:31:51 when they were starting their 1:31:53 prediction market website. 1:31:57 A lot of people, 1:31:59 and 1:32:01 there's a lot of trouble getting 1:32:03 the critical mass of users 1:32:05 because 1:32:07 no one wants to be the first guy to trade. 1:32:09 So the scoring rule and 1:32:11 forcing the market 1:32:13 guy, the creator, to put up 1:32:15 the bounded loss 1:32:17 cash, even if it's a very small 1:32:19 amount, mitigates this 1:32:21 to a great extent 1:32:23 but I don't know 1:32:25 if that extent will be enough. 1:32:28 The critical mass is very important 1:32:30 so that's a sad thing. 1:32:32 It's really sad to see a market that 1:32:34 no one is trading in because you know that 1:32:36 no one will want to jump in and be the first guy 1:32:38 so it will just kind of, even if a lot of people 1:32:40 are actually interested, 1:32:42 if the engine stalls, 1:32:44 it's stalled out and that's 1:32:46 too bad. 1:32:48 I think I have one final question for you. 1:32:50 I asked Augur the same thing 1:32:52 where prediction markets are 1:32:54 technically illegal, I guess, 1:32:57 in the United States but 1:32:59 their answer for what they were doing 1:33:01 was they were being mindful of the law 1:33:03 and working with legislators 1:33:05 to, I guess, build 1:33:07 a tool that could be used 1:33:09 in a legal way but still 1:33:11 have a legal potential but they weren't 1:33:13 at fault, I guess, for any 1:33:15 misuse of their platform. 1:33:19 It still, I guess, goes against 1:33:21 the idea of prediction markets 1:33:23 are illegal in the US so could you, I guess, 1:33:25 tell us more about what your stance is 1:33:27 on the legality of it? 1:33:29 Yeah, well, if they were illegal 1:33:31 there would be no need to decentralize, right? 1:33:33 I could just start 1:33:35 the website. 1:33:37 But I'm just a guy who 1:33:39 publishes papers 1:33:41 and some software code 1:33:43 and Roger Ver hired me away 1:33:45 from Yale and in our contract 1:33:47 it says that he has to pay for my legal fees 1:33:49 if I get sued so 1:33:51 I don't take any... 1:33:55 In the United States 1:33:57 your liability really begins 1:33:59 when you take money from people 1:34:01 which is something that I haven't done 1:34:03 on purpose so 1:34:05 but Augur did do that 1:34:07 and I advised them not to do that 1:34:09 but they did it anyway because 1:34:11 Vitalik and all these other people 1:34:15 Vitalik and all these other people 1:34:17 are like pro-crowd sale 1:34:19 and I don't know about that 1:34:21 so I'm very anti-crowd sale 1:34:24 I think the crowd sale does a lot of bad things 1:34:26 at once and it's wrapped in this 1:34:28 kind of free market 1:34:30 let people do whatever they want with their money 1:34:32 thing but it doesn't quite 1:34:34 when it comes to open source software I think it actually 1:34:36 does a lot of 1:34:38 it's the 1:34:40 wrapping of the free market 1:34:42 stuff is kind of 1:34:44 it hides a sinister 1:34:46 core 1:34:48 It's not illegal 1:34:50 because you're essentially just creating a means for people 1:34:53 So first of all 1:34:55 the platform doesn't even have any prediction 1:34:57 markets on them at all and you can just 1:34:59 create 1:35:01 even if we were calling these 1:35:03 even if we were trying to label these 1:35:05 things as markets 1:35:07 because again it's just a set of rules 1:35:09 it's just a protocol 1:35:11 but even within the protocol assuming 1:35:13 the protocol were a thing 1:35:15 a user can create the decision and then 1:35:17 from there a different person 1:35:19 can create a market using 1:35:22 and then from there different people can trade 1:35:24 but 1:35:26 right so I 1:35:28 wouldn't even if I created the system which 1:35:30 I'm really not doing I'm just 1:35:32 writing about it a lot and it's slowly 1:35:34 being built by volunteers 1:35:36 which are like sort of managed by me 1:35:38 but not really right and nothing stops 1:35:40 you know 1:35:42 but it is like 1:35:44 a lot of other people are kind of helping to do 1:35:46 the actual development work 1:35:48 but the other thing is 1:35:50 the core of the question is that 1:35:52 there is no 1:35:54 I mean this is just a totally different 1:35:56 thing but to really answer 1:35:58 the real question is that I do 1:36:00 understand that 1:36:04 prediction markets are not 1:36:06 you know it's really not clear that they 1:36:08 are actually totally illegal 1:36:10 in the United States but obviously yeah 1:36:12 there would be a lot of places you'd have to register 1:36:14 with and things like that but 1:36:16 it's not clear that this thing 1:36:19 meets the definition of a prediction market 1:36:21 A, it's not clear that it meets the definition 1:36:23 of anything or any market 1:36:25 even though that's how I 1:36:27 try to describe it but 1:36:29 B, there's 1:36:31 people have wanted 1:36:33 to make the case for a very long time 1:36:35 I mean 1:36:37 ultimately 1:36:39 ultimately 1:36:41 there is 1:36:43 I guess it's hard to explain but 1:36:45 there is like 1:36:48 a lot of people 1:36:50 I think want this to happen 1:36:52 but they just 1:36:54 kind of feel like they have to 1:36:56 say they don't want it to happen so 1:36:58 it would be very cool to have 1:37:00 it just sort of happen slowly and experimentally 1:37:02 and not 1:37:04 I would want it to happen 1:37:06 kind of very 1:37:10 I don't know very 1:37:12 what's the word 1:37:14 gradually I guess 1:37:16 organically right and so 1:37:18 it's just this is an idea 1:37:20 and if 1:37:22 it inspires 1:37:24 things to happen then that's what it did 1:37:26 but I really only see it as an idea 1:37:28 at this point still 1:37:30 and you know it's 1:37:32 a lot of stuff is written down 1:37:34 on paper and that's free speech 1:37:36 and then a lot of stuff is software 1:37:38 but I don't know what the 1:37:40 most of the software isn't written by me 1:37:42 and I don't know what 1:37:45 it's MIT open source 1:37:47 well I guess there's still a concern though 1:37:49 like what happened with Silk Road 1:37:51 and Ross Ulbrich 1:37:53 like he didn't list anything on the site 1:37:55 but he was the one who created it 1:37:57 and made an example out of it 1:37:59 yeah well I think that would be very interesting 1:38:01 I mean I think I would do better than he did 1:38:03 he's in court right 1:38:05 he's in court and six people did die 1:38:07 and it was unfortunate 1:38:09 it would have been a very interesting experiment 1:38:11 if he only allowed like marijuana sales or something 1:38:14 because six people did die 1:38:16 and I think five of them 1:38:20 were heroin if I remember correctly 1:38:22 and one of them 1:38:24 the only one that wasn't heroin 1:38:26 was like a 1:38:28 very young teenager 1:38:30 who was like left alone 1:38:32 by his parents and he did 1:38:34 a bunch of kind of like 1:38:36 hallucinogenic drugs and like jumped off a balcony 1:38:38 or something and I think that that would be 1:38:40 you know that's a much easier 1:38:42 you can blame the parents 1:38:44 for that or something right 1:38:46 you can go in as a defense attorney 1:38:48 you can just say look we had millions of customers 1:38:50 but where are the parents 1:38:52 you know and then you can kind of like 1:38:54 you know I mean like it's of course 1:38:56 but I think you know 1:38:58 you have to really imagine it in court 1:39:00 I mean it's clearly he was running 1:39:02 I obviously think 1:39:04 the sentence was 1:39:06 was 1:39:08 very extreme and I'm not 1:39:11 a very anti-drug war 1:39:15 I have a lot of nuanced views about it 1:39:17 but certainly all kinds of 1:39:19 all kinds of 1:39:21 club drugs and marijuana and stuff should be 1:39:23 100% legal 1:39:25 in my view but 1:39:27 as should most 1:39:29 almost everything and the enforcement 1:39:31 is a huge problem 1:39:33 but I think there's a huge difference 1:39:35 between that and kind of what I'm doing 1:39:37 a guy 1:39:40 who wrote about 1:39:42 creating a better world through 1:39:44 information but you know maybe I could be 1:39:46 totally wrong about this maybe they'll be playing this video 1:39:48 in court and I'll be thinking oh wow 1:39:50 it was a big mistake to go onto that 1:39:51 well Paul if they do this video in court you're gonna hate this next question from me 1:39:56 have you have you had any talk about how prediction markets can and will naturally 1:40:06 usher in assassination politics yeah I know I wrote a giant essay about how they won't so I'm 1:40:14 actually I'm thankful to get that question so that's number six in the sequence you can go 1:40:19 on to the website Bitcoinhivemind.com and it's on there somewhere and yeah I know people are 1:40:24 obsessed with that question and I actually worked through that I'm actually happy to get the 1:40:28 question and I hope they play it in the hypothetical courtroom because I did actually before I I wrote 1:40:35 the whole thing and I had finished it in November of 2014 I think oh no I think maybe 2013 I can't 1:40:44 even remember now but I waited until December 3rd I think December 3rd 2013 I didn't publish it at 1:40:53 all and I spent all of November working through the assassination markets thing and the only reason 1:40:58 I didn't publish it earlier because I wanted to see kind of what other people that had to say on 1:41:02 the subject I thought it would bias people's views but the short answer is that it the market 1:41:09 will you know the markets will is okay so it's kind of hard to explain but the market is always 1:41:17 going to be pricing the person's health accurately so it will kind of converge the actuarial tables 1:41:24 likelihood of of them dying so if you put a million dollars on it won't die which is would 1:41:33 appear to give someone an incentive to put you know their own million dollars on no they will 1:41:40 die and then shoot them and then you think something a little bit more sinister what if 1:41:46 somebody who wants to be a particular politician die makes a wager that they are going to die on 1:41:53 a particular day by a gunshot against that on a market with a billion dollars worth of Bitcoin 1:42:03 that you know so-and-so will not die on this day by a gunshot then somebody who is able to actually 1:42:13 make that gunshot yeah I got on a consensus so there's everything right so um the so the thing 1:42:23 is I've thought about it and it is possible to use Bitcoin to like do stuff like this but I it is a 1:42:30 little complicated and I haven't written down how and I don't plan on writing that down because I'm 1:42:35 not interested in that but these are these prediction markets are not assassination markets 1:42:40 so Jim Bell's essay about assassination markets is is has what you suggest where the person can 1:42:46 place a secret wager that's very specific and and then the guy the assassin can claim it and so 1:42:55 these are different so one difference Jim Bell's assassination markets is more of a market in the 1:43:03 sense of a like a supermarket where you would go and buy things at a fixed price but friction 1:43:10 market has the price react to different to the bets as their place so if you if you bet a lot 1:43:16 of money on the person not dying what someone who thinks that they will die even if they're 1:43:24 even if I'm not planning on assassinating this person but they say wow a million dollars is a 1:43:30 lot of money someone's gonna take them up on that there's now an incentive for these totally 1:43:35 passive innocent third parties to just put a little bit of money on die right because someone 1:43:41 would probably take them up on this on this offer right they'll think like oh here's a million 1:43:46 dollars you know I think the person will probably die but by doing that you actually drain the 1:43:51 battery you realize you're draining the million-dollar bounty so there's this there's this 1:43:55 kind of this kind of negative feedback loop that is inescapable and so the question is does the 1:44:03 negative feedback loop is it strong enough to withstand you know the context into which it 1:44:11 exists but you know this particular setup where it's like public blockchain is very is very not 1:44:21 amenable to this this type of thing and in fact there are superior alternatives which I write 1:44:26 about in the in the essay so if you think the person the politician is really incompetent you 1:44:32 can actually you can set the market up so that you make money if they if they fail to you know 1:44:38 decrease unemployment or something and and that's you know a much you can set it up so you make 1:44:43 money instead of having to pay a million dollars but I from being used for assassination politics 1:44:51 it would seem would be to have a a bet cancellation feature that some committee if you will could 1:45:02 decide so I mean let's say that you and seven trusted friends it's in the protocol and it's 1:45:09 in the essay that very and I magnify everything and I say you know in the protocol you can say 1:45:15 that you can have like I I tried to mention a long time ago when this this conversation began 1:45:22 about if the question is the question is supposed to be very easy to answer and so if it's not easy 1:45:28 to answer they can they can select this 0.5 value of like we're all agreeing that this question is 1:45:35 not answerable or something but there are rules there would be guide publicly commonly known 1:45:40 guidelines in advance obviously so that people know it's something you know like two minutes of 1:45:45 googling you know and then give up and then you could just say if one of the rules is no no 1:45:58 questions about that and then you and then you there's no since everyone knows in advance it'll 1:46:04 be voted 0.5 the market will never actually be able to control any money in it but there are 1:46:12 other things you know in the there is like different like where it regresses to private 1:46:16 individuals hiring other private individuals and so so I encourage you to read the whole essay and 1:46:22 and then let me know what you think but I have never thought that I do believe that someone will 1:46:29 figure out how to do I mean I've already figured it out how to use I think Bitcoin to create 1:46:35 assassination markets but I'm not going to tell anyone about how that how I would do it because 1:46:41 I'm not interested in doing that and I don't really agree I think prediction markets are better 1:46:46 version actually the more effective change vehicle than assassination markets anyway because you can 1:46:52 you can you can describe exactly what it is you don't like and you can profit if the politician 1:46:58 screws it up while promoting the guy who's actually and prediction markets are just better 1:47:03 because they're they're really nice you know it's hard to understate a lot of people have no idea 1:47:08 what they are how they can be used but what really bothers people often is how helpless and confused 1:47:15 that they feel and so when people feel like there's a process that actually works or there's 1:47:22 some phenomena that actually reliably gives them information I think they would they would be 1:47:28 happier about that so I think there's lots to say about that in fact that essay is something like 1:47:33 26 pages long on itself and it's like all text so there's no pictures so but yeah there's a that's 1:47:42 a I used to get that question all the time and I used to kind of sit back and say I don't know 1:47:47 what do you think about that and I don't know if but yeah there's a there's lots to say about that 1:47:54 but I don't think it I don't think it's likely at all I would be a little disappointed but certainly 1:47:59 that would be a case for censorship resistance though you would need certainly would not allow 1:48:06 those things to exist in in any organizational and decentralized business in the United States 1:48:15 or elsewhere but yeah that's a neat that's a neat thing to talk about in bars final questions I 1:48:21 guess that anyone watching the show I guess has that they can go to Bitcoinhivemind.com and get 1:48:28 in contact with you to explore this topic a lot further we're coming kind of here on two hours 1:48:35 like people I know you had a question is it really small okay we'd like to end the show with our our 1:48:42 pseudo prediction market that we do where we predict the price of Bitcoin and I got to share 1:48:47 this one from last week because it's actually pretty phenomenal the price was crashing while 1:48:51 we were doing it and we predict a week in advance what the price of Bitcoin will be and I think 1:48:57 Jeremy actually got this one he guessed the opening candle price perfectly Yuri I'm sorry it went up 1:49:05 a little bit back down but whatever you're doing on the chorus exchange so we'll go around the room 1:49:17 here and we'll get everyone's I guess prediction and the show with what they think the price will 1:49:22 be so we'll start off with our guest of honor here all what do you think the price of Bitcoin 1:49:28 is gonna be next Thursday at this time next Thursday at this time what is it right now I 1:49:32 got on my phone and 86 I have 386 75 I don't know if that so okay well I mean the logical thing 1:49:43 would be to just say that it will only go up a little bit so I'll just say it'll be 390 in one 1:49:48 week I mean so I'll be a really boring guess and I'll just say I'll just say that it's it's they're 1:49:53 all net present value zero projects and it's by the grouping of lines a lot of people guess the 1:50:00 same price would be a week later and yeah sure enough we rebounded with another crowd at work 1:50:06 there cam what do you think I'm gonna say I'm so conflicted I think by next week we'll probably 1:50:30 be down to 350 I'm gonna be a little bearish probably the ball we'll go for 15 or 15 you're 1:50:48 planning on buying Peter Peter and Dave David I was hoping to go last I was gonna take one extreme 1:50:59 or another he said for 15 I'd go for 16 and and then get everything high or or you said 350 I'd 1:51:07 go 349 and get everything low watch too much price I actually have a funny joke about that 1:51:14 after you're done with that okay but but but why don't why don't I say why don't I say for 16 I'll 1:51:49 know but that you joke about that strategy but I know when I was working at Yale I was working for 1:51:55 a very important person who knew lots of other important people who are in the like Society of 1:52:00 forecasters or something and they there was a guy who was doing very well and the guy I was working 1:52:06 for dr. Nordhaus is like super smart super important guy he's like asking him he wasn't a 1:52:11 forecaster himself or something but he asked this guy oh you know what's the secret and he's just 1:52:15 like I don't know he's like I just you know I just I try to be on the high edge or the low edge 1:52:20 because there'll probably be a shock and then you know and if it's in my direction I I win or 1:52:25 something so that's you joke about that but actual like you know this is not like any serious 1:52:30 forecasting but there was like a like an informal thing just like this and the guy who was winning 1:52:36 was doing exactly what you just joked about so so it's there's a lot of merit to it obviously and 1:52:44 and it highlights perfectly forecasting when there's no skin in the game right there's no 1:52:50 yeah we're almost yeah we have a nice kind of a testnet that's sort of bubbling up so when it's 1:53:17 maybe like in a few days you know like if it'll be so maybe I'll send you out I'll ping you back 1:53:24 and then we'll kind of see kind of where it's going and maybe you can tweet it out and then 1:53:29 maybe a couple months later yeah more interesting conversation yeah all right fantastic all right 1:53:35 cool from us here at a block talk we thank you so much for watching our one viewer and anyone 1:53:40 so we'll see you in the next thousand blocks