Epicenter - Learn about Crypto, Blockchain, Ethereum, Bitcoin and Distributed Technologies - Joey Krug: Augur – A Decentralized Crowdsourced Prediction Market Built on Ethereum

Episode Date: August 9, 2016

Prediction markets are fascinating financial instruments which have proven to be accurate at making predictions on things like the outcome of elections, geopolitical events and sporting events. But in our heavily regulated financial world, they are the subject of much controversy. In 2013 for instance, a well-known US prediction market, Intrade, was forced to shut down following a civil suit filed by the CFTC. While some people consider prediction markets to be a useful tool for society, others consider them to be a form of gambling, which is one of the reasons why they have been met with such resistance in certain countries. A purely decentralized prediction market would operate outside the scope of the regulated financial world and be resistant to censorship and outside intervention. We’re joined by Joey Krug, Co-Founder and Core Developer at Augur, a decentralized prediction market built on Ethreum. The project, which was initially meant to be a Bitcoin sidechain, has ported to Ethereum and is currently in beta on the Testnet. Users can forecast real-world events, such as the outcome of the US election, and earn profits if they are accurate in their predictions. Reporting on events is crowdsourced using a consensus-based system similar to proof-of-stake. Topics covered in this episode: How prediction markets work The controversy around prediction markets The Augur project and what it is trying to achieve The evolution of Augur The different components of Augur Reporting on events and the reputation token Security, attack vectors and how they can be mitigated Augur’s business model The ethical aspects of prediction markets Augur’s structure and governance Episode links: Augur Augur Git Repo Augur Docs EB139 – Martin Köppelmann: Gnosis – The Ethereum Prediction Market EB97 – Paul Sztorc: Truthcoin & Prediction Markets, From Information-Overload To Crowd Intelligence EB98 – Robin Hanson: Futarchy, Prediction Markets And The Challenge Of Disruptive Technology EB141 – Ralph Merkle: Revolutionizing Democracy Using DAOs This episode is hosted by Brian Fabian Crain and Sébastien Couture. Show notes and listening options: epicenter.tv/143

This is Epicenter Bitcoin episode 143 with guest Joey Krug. This episode of Epicenter Bitcoin is brought to by Jax. Jacks is the user-friendly wallet that works across all your devices and handles both Bitcoin and ether. Go to JAAWX.io and embrace the future of cryptocurrency wallets. And by Hi. DotMe. Protect yourself against hackers and safeguard your identity online with a first class VPN. Go to hi.combe slash Epicenter and sign up for a free account today. Hi, welcome to Epicenter Bitcoin. The show Edge talks about the technologies, projects, and startups driving decentralization

and the global cryptocurrency revolution. My name is Sebastankojuio. And my name is Brian Fabian Crane. We're here today with Joey Krug. Joey Krug is the co-founder of Auger. Many of you have probably heard of Auger. Auger is a prediction market that raised about $5 million US dollars in a crowd sale quite a while ago. And he was previously, he was a computer science student.

Pomona College and it's been very young but diving fully into into the crypto world and into the world of prediction markets. So thanks so much for coming on Joey. Thanks for having me. So this is a topic that we somehow don't seem to be able to get away from. So we've done I don't know how many episodes already on prediction market This is certainly the fifth or sixth episode or even more that we've done here probably even more and all I think is probably the best known project in the cryptocurrency space, the best known prediction market project, maybe along with Gnosis now, Auger probably still a bit better known. And yeah, they did a big crowd sale quite a while ago, which got a lot of attention, and now it's still sort of, you know, they're building it, and a lot of people are waiting to see the launch.

so I look forward to diving into this today. So maybe to get started, Joey, can you tell us a little bit about how did Auger come about? Like, what's the origin story? Yeah, so the idea for Auger is kind of, it's really kind of a combination of like three, three different ideas proposed by other people. We just kind of combine them in an interesting way. So there was a paper in 2014 called Decentralizing Predictual Markets and Limit Order Books. and that was paper written by some guys at Princeton.

And it was basically the idea to make a decentralized prediction market. And so, like our order book mechanism, the mechanism for issuing shares, things like that, all came from there. And there was a blog post by Vitalik called Shelling Coin, which was this idea for getting real-world information into the blockchain, using basically a set of mechanism. And the idea is that you'd put up a bond with ether. and if you were wrong, you'd lose the bond. So we used that. And the third thing we use is there's this project called Truthcoin,

which proposed using similar system to what shelling coin did, except instead of using something like Ether as the bond, you'd create a new token that was directly tied to the prediction market, which makes it easier to reason about the security. It also kind of prevents, like, you know, if you use Ether, maybe 2% of Ether would participate. If a whale comes in, they can mess up the entire system. But if you have a token where everyone's participating in it,

and if they're not, they're losing it, then it's a bit easier to reason about the security. So we basically combine those three things to kind of create auger. We started working on it in 2014, and we were initially building on Bitcoin Core, and ended up switching to Ethereum. And what made you excited about this idea of prediction market and particularly decentralized prediction markets?

So the most interesting thing to me is really, that it kind of allows for the first time of a few things. It allows for there to be a global kind of financial market where anyone can go to it, you know, without capital controls. A person in China can participate on it, someone in the U.S. can. Today, that's not really the case for most financial markets. It's also relative to certain things. It can be cheaper.

And then the other thing that's kind of interesting to me is just the idea of using prediction markets to actually, like, predict things. You know, currently today we don't have very good information. about the future. What we do have is mostly from people like pundits who, you know, are kind of like broken clocks to write, you know, a couple times a day and that's about it. And so with prediction markets, we can aim to get better information about the future. Yeah, and of course we have, yeah, we have plenty of episodes that explore this topic in depth, and we can get back into that a little bit as well. And of course, we will have links to that. So you mentioned that originally

auger was based on bitcoin and the idea was to build it as a bitcoin side chain what happened there why did you end up switching to ethereum yeah so originally the idea was like to add some op codes to bitcoin actually quite a few op codes um and then once side chains came out we'd make a two-a peg between bitcoin and you know our side chain of bitcoin um the way we kind of got building on Ethereum was I'd been working on it on Bitcoin about a month and I'd added a few off codes for like trading stuff basically. And one weekend I was sitting around and I was like, I thought about using Ethereum. I was like, maybe I should just check it out and see what it's like.

So I decided to try to build what I'd built, you know, over the past month on Ethereum. And I was able to do it in one day, the same work that I'd spent, you know, 30 days on on Bitcoin. So then that kind of made me think, well, if we're making all these modifications to Bitcoin, are we really getting a whole lot of Bitcoin security model by doing that? And the answer I kind of concluded was no, not really. And it made more sense for us to just kind of take the gamble with Ethereum as opposed to building on Bitcoin. And the ability to make stuff a lot faster on Ethereum is also pretty useful. Because with Bitcoin, you have these really, really low level op codes on Ethereum, you have relatively high-level programming languages

that you can build applications with. Of course, if you compare Bitcoin to Ethereum, you know, Ethereum would be the logical the logical blockchain on which you would build something like this because you have that much more flexibility. Do you think that something like Auger could potentially live and become successful on a side chain? Or have you given up that idea completely? I mean, like it could, but I basically kind of ditched that idea. You know, I think I think what would be most interesting is if there were a good side chain between Ethereum and Bitcoin. So there's like VTC relay right now.

It'd be great if there were like a two-way peg side chain. The only kind of problem with that is the people of Blockstream aren't really willing to add support for verification of Shaw 3 proofs to Bitcoin, which is a problem since the Ethereum uses Shaw 3. So you'd probably have to have some like intermediate. very side chain, which would get kind of complicated. So let's dive into prediction markets. As Brian mentioned, this is a topic that we've covered extensively on the show. So we're not going to go in too deep into prediction markets, but I'd like to just reiterate

some of the, or for you to give us your perspective on some of the best use cases for prediction markets today. Again, I think the most useful use case today, like if Agarra were live right now, I think would be allowing people to trade things that they normally, you know, wouldn't be able to trade for whatever reason. So my favorite example is, you know, if you're Chinese national and you want to buy something like Apple, but you can't because there's too many capital controls, if you're a huge whale and you have, you know, big institutional connections, you can do it. But as an average Chinese middle class citizen, you're kind of stuck with the current Chinese

markets, which most of the stocks aren't that good. And you also have lots of problems with people like average Chinese middle class citizens are trading commodities at 20 to one leverage because they want to speculate on stuff, but they can't. And so the stuff they're speculating on is just kind of crazy. And so I think if you use prediction markets on something like Ethereum, you can actually get the same profit and loss of something like Apple. Of course, you don't have any dividends. But it allows people to trade things that they never were able to trade before, which I think is pretty useful. And it's interesting how you use this word speculation, especially with sort of the Chinese

market. There's this interesting sort of thing that happens with prediction markets where some of the people that talk about is definitely look at it as as a speculative tool and something that you could use for essentially gambling. And others would look at it as something that can be very useful for society. You know, I would encourage our listeners to go back to the Robert, the Ralph Merkel interview. What is your position on that? Where are you stand? with regards to prediction markets and there are usefulness for society or do you see them more as really a speculative tool? So I would say, I mean, I see them as both. I think speculative tools are inherently useful for society. You know, of course, like Occupy Wall Street people would disagree with me, but there's a great paper by the economist Frederick Hayek.

It's called the use of knowledge in society. And it's basically about how speculative markets or financial markets have prices and prices are equivalent to information. So better prices means better information in society. So I think that they're kind of intertwined. You can't really separate them from each other. So by virtue of having better financial markets, you have better prices. So you have better information that society can use to make decisions. So I kind of view it as like if you want to make better decisions, there's a couple ways to do it.

I mean, one is to just ask an expert and hopefully they'll give you a good answer. The other is to use something like a prediction market and try to make the market as, you know, liquid. and healthy as possible to give you the best answer that you can get. So I think the speculation kind of ties in very nicely there. You can't have prediction markets without speculation, and you can't really have the information without it either. They're both, it's like a yen and yang, you need both of them. So there's been a lot of kind of repression of prediction markets,

but the concept's been around for a long time. At the same time, they're not really used very much, although of course you can think of some of the existing markets as prediction markets, but kind of like pure prediction markets are a very marginalized thing. The biggest one in trade was shut down or forced to shut down. Why do you think there is so much difficulty around creating and operating prediction markets when at the same time a lot of financial markets are, you know, government supported, regulated, etc.

Sure. Right. So I think if you look at the history of financial markets, they've always kind of been like this. The stock market was initially considered just gambling, and it was basically something that shouldn't exist, and governments didn't like the idea of having stock markets. And you see the same thing for almost every other type of financial market in history. Same thing for options markets, same things for futures. And so it's kind of natural that prediction markets as a sort of different category of financial market, which follow the same progression, where they're kind of ignored and people don't really get them or they kind of don't think it's a good idea at first.

So it makes sense to me that we've kind of, you know, that that's happening. And so I think it's just like a natural thing that eventually they'll become, you know, a more widely recognized type of financial market that people will participate in. You know, like a good example is if you're a global macro investor and you want to make a trade on something where you have some macro information. A prediction market's almost always

given that it has liquidity, a way better way to do that than an existing financial market because you cannot get the exact contract that you want. Whereas with the prediction market, you can have a contract on anything, any future event. So I think eventually they'll kind of be recognized as normal financial markets, but it just takes time.

So you think this is really because prediction markets are fairly young and because they're not too well understood, but once that changes, they will be accepted. There's nothing deeper that's creating a bit of aversion and resistance to prediction markets. So I agree with that for the most part. There's one kind of small bit of a version that you can have with prediction markets, which is that since they're a very broad financial tool, they're basically a financial market where it says you can make a financial market on any future event.

And that's kind of what a prediction market is. Since they're so broad, you can use them for things that are like kind of, you know, things that people want to speculate on, which makes people kind of wary of them. So if you look at like the Future Map project in the mid-2000s, which is a project like with the CIA and DARPA, they were going to have prediction markets on predicting basically geopolitical events like terrorist attacks or, you know, whether a certain dictator would be disposed, things like that. And so there is a version for those reasons. But unlike the financial side, like, you know, predicting some economic data, like, say, GDP or something directly, I don't think there's much a version there. Prediction markets have been shown to work pretty well for, you know, big popular culture phenomena. Like, they're very much used in political, to make political predictions, sporting. Also, I was sort of surprised to learn that there's a lot of prediction markets around Hollywood movies and which ones are going to go.

growth the most, et cetera. So, you know, these categories seem to, seem to work very well with prediction markets. Do you see other categories in which they're successful or other categories which may emerge in the future where prediction markets could be a good tool to predict outcomes? Yeah, I think, and I think one of the most interesting, you know, things you can predict with them is, like, data surrounding companies. So, like, you know, how many sales a certain product will do whether Apple's going to release X or Y first. Things like that would be very interesting as opposed to just hearing them from like the rumor mill.

So things like that. And there's also the really wild out ones, you know, like Robin Hansen talks about where you have like a market on like, should we fire the CEO of ex publicly traded company and what would the stock price do if we did that? I think somebody will probably make a market like that on Auger. I think it would be very interesting.

Microsoft in the mid-2000s, could have avoided, you know, a long slow decline if they had a market like that, I think. Because after the last CEO left and then, you know, sat down a dollar or whatever his name is, became CEO, the stock price jumped quite a bit. They've been doing pretty well since then. So if you had markets like to kind of, you know, catch stuff like that early, I think it'd be a good thing for society. I was watching the, there was a panel at DefCon. with, I believe, your co-founder on prediction markets.

And the moderator gave this example of how could you establish a prediction market to find out whether or not Walmart was doing well as a publicly traded company. And one way you could do that is have local prediction markets on how well the stores are doing. Or if the construction of new stores is going on schedule and then do this sort of microeconomic indicator at a very local level would allow you to establish predictions at a more macro level on how is this company doing? Yeah, I like that idea of using it. One thing that they're also really good for is predicting deadlines for things.

Prediction markets are pretty accurate on that. Like just recently, I started a prediction market internally in Auger for a deadline for like the next version of the beta. So stuff like that is kind of cool because you can see. what people actually think, you know, from like a level where they're willing to lose money if they're wrong, which is a lot different than what you get if you just ask someone a question. Let's take a short break to talk about Jax. Jax is a cryptocurrency wallet created by the people at the Central.

Now, there are two cryptocurrencies that matter at the moment. One is Bitcoin and one is Ether. But using them can be tricky, what wallet use. secured him. Where did I leave my umbrella? It's all a big mess. And that's where Jacks comes in. Jacks is a unified wallet. It works across all your devices. It works for their Android phone, Apple, iPhone. It works for your desktop computer. And they have browser extensions for Chrome and Firefox. And it works for both currencies at the same time. It works for Bitcoin and it works for Ether. One of the things that makes Jacks as delightful as walking through the Fifth,

that Hongji Small v. Paris on a Sunday morning and getting a whiff of fresh pastries is how they leverage HD wallets. So they use a 12-word single backup seed for all three currencies and make it super easy to sync your wallets across all your devices. So if you're using the Chrome extension or the desktop app, you just can whip out your phone, scan the QR code, and boom, your wallets are synced. And plus, the people at Jacks take your security very seriously. It's open source, so anybody can look at the code. And plus, they never hold any kind of. customer funds. All the keys are stored locally on the client side. So go to jacks.io, that's j-a-double-x.io, download the jack's wallet right now and understand what it's like to use a next generation wallet.

We'd like to thank jacks for those supportive epicenter. You mentioned that prediction markets are, well, good or have a good track record. Now, one of the things you said or that was written on the website is that prediction markets kind of leverage this idea of the wisdom of the crowds. Now, I was just kind of reading a little bit about that book. And in that book, he says that one of the assumptions to have this wisdom of the crowd's effect is that you have independence. So, you know, different people make predictions.

And let's say now if we talk about GDP, right, I'm going to say, okay, GDP next year in the U.S. is going to grow by 1%. and then sort of all of those are taken together, and then you take the average, you have this kind of wisdom to crowds effect. But when you don't have independence, of course, is when I see what everybody else is doing like you have in kind of regular markets.

So what is your fault on that? How do you think about independence and its effect on the usefulness and accuracy of prediction markets? Yeah, so, you know, The wisdom of the crowd, like if you read that book, I think it's by James Sirowakey. It talks about, like, most of the markets, sorry, not markets, most of the examples they use are not actually, like, financial markets. So I think that independence matters a lot more in things that aren't financial markets because you don't have money on the line.

So when you have money on the line, you're thinking it becomes a bit clearer because, you know, you're going to lose money if you make the wrong decision. And he does touch on financial markets. He talks about how I don't remember if it was Challenger or which NASA mess up with the O-rings, but the stock market basically pointed out which company made the O-rings based off of a huge decline in their stock a few days after the crash. So I think it works in financial markets too. It's just that independence is a bit different because you don't really have as much independence. You just have a price and people are saying whether they think the price is wrong or not. The nice thing about financial markets is though, or any market in general, is everyone thinks they know more than everyone else.

Or at least the most participants do, which is why you see like, you know, the vast majority of hedge funds lose money compared to just buying the S&P. And so I think that that kind of counteracts the drawbacks of having lack of independence. the real money factor basically. Yeah, no, I think that's a valid point. And I guess prediction markets are going to be as efficient in that regard as other financial markets, right? So, of course, they will not be perfect. And there will be strange phenomenon up, but that's just the same thing we have with stock market prices or exchange rates and stuff. So moving on to go a little bit more in depth with Augur.

what does Auger's architecture look like? What are the different components in there? Yeah, so there's really kind of like three main roles in Auger. So there's the market creation side of things, which is, you know, traditionally with the prefecture markets, you're just kind of stuck with whatever markets, the place you're going has. And so you can't create another one if you want to.

With Auger, anyone can kind of create their own prediction market contract in a few minutes on Ethereum. And so there's that, role, there's also the trader role, which is, you know, the person who obviously is buying and selling shares in the markets. And then the third role is the reporting role, which is used to kind of resolve the markets in Auger. So, if I understand correctly, the first thing that you said is that anybody can create predictions. So what you're saying is that with Auger, anybody can create their own predictions, uh, independent of, you know, the community's will or some central authorities, uh, desire to host those predictions is, Is that sort of the general value proposition or something like this? Right. Yep.

Okay, so let's talk about reporting. This is an interesting sort of concept where in a typical prediction market, you would have some sort of a central authority or a central body that would be bringing in data feeds and reporting on the outcome of a prediction or an event. How does that occur in Auger? How does Auger get its data on the outcome of an event? Right. So the way Auger kind of does it is it has this reporting process where this basically, you know, basically we sold this thing called reputation in the crowd sale, which is a very crypto-economic reputation.

It's basically saying that I'm willing to have this thing that's worth value and I'm willing to lose it if I report inaccurately or dishonestly. And so the reporting process kind of has three different. phases to it and everything goes well you only get to the first one and the way it works is initially people are randomly selected to report on events so there may be a market about presidential election there may be one about the share price of Apple and you get selected to report on one of those I get selected to report on the other something like that you submit your report and then what happens is it basically

takes basically a simple average of our reports and then someone can kind of challenge it if they think it's wrong. And to do this, they post a bond or they put money up, and then everyone reports on it. And at that point, if their answer is the same thing we agreed on, the person loses the bond because they basically just wasted the system's time. If it's different, then they actually get money back because they, you know, challenge the system and we're actually right. Isn't the final part of reporting this? You can actually fork the network if this phase where everyone reported and everything ended up getting it wrong as well. And so the way reporting works is kind of every two months you're selected to report on a random subset of events. And the idea here,

so with reputation, which is the thing that you sold during the crowd sale, then the expectation two is now if I hold some of this reputation tokens, it gives me two things. One is kind of it gives me like a job, right? So I have to report on events. And then the other thing is it gives me some kind of a chance for a return, which could be through either the appreciation of the token, right? It gains value. It's tradable. Or through some fees that get paid to me. Is that correct? Yeah, that's right. So the way the fees work is when someone creates a market, half the fees go to the reporters. The other half are split up amongst the person who created the market and people who are market making on the market or providing liquidity.

I'm curious here. Why did you call this reputation? Because reputation, as I looked up the definition before, and it said reputation is the beliefs and opinions that are held about someone. So when we think about reputation in our life, then this person has a reputation to be reliable to be a fraudster, something like that. And that's a person. very much kind of attached to that person and their actions. Now, you're making this reputation tradable, which seems to be completely contrary to how we commonly understand reputations. How does that make sense? So we thought about tons of different names, and this is kind of the one we stuck on. I kind of thought it was kind of a bad name for those reasons as well, but we couldn't think of a better one.

It's a very crypto economic reputation. So yeah, if you think about reputation, you traditionally think of something like, oh, that guy's a reputable person, you know, he'll pay you money back if you load them $50. Or that person has, you know, 200-something stars on eBay. It's probably fine to buy from them. Or this person has a 4.5 star rating on Amazon. They seem like a good seller. With Auger, though, yeah, it's quite a bit different.

It's a very economic reputation. It's basically saying that because in crypto systems, it was, least as of 2016, there's not a real good way to kind of have that sort of ephemeral reputation that you have with Amazon or the fact that you just know someone. And so basically you do things like Sybil attacks and you do the fact that people are pseudonymous. And so to get something that's kind of more secure, you really need an economic one where it's basically saying, I'm reputable and the way I'm going to prove to you that I'm reputable is that I'm willing to lose a bunch of if I try to screw you over.

It's kind of what it is. That's kind of the way it works. Okay, very interesting. So does that mean, let's say now the situation have been different, and there were good reputation systems, decentralized, blockchain-based reputation system when you guys started Auger. Do you think that would have been better to rely on a more sort of real sense of reputation the way we know it from life versus this?

tradable reputation that you currently have? That's an interesting question. I think I mean so from a purely economics game theory perspective, the answer would be no

because you want it to be such that if someone, you want someone to have no incentive to do something wrong. So you want them to lose so much money that anybody they can make from doing something wrong is far less than what they would actually lose from just doing the wrong behavior. So from an economic standpoint, the answer is no. From a more practical standpoint,

I mean, yeah, if you had something that were like, like, you could have, like, you know, some way to prove that someone was, like, 100% perfectly reputable, then, yeah, it'd probably be simpler. I don't think, I don't think, I don't think you can even do that in real life, you know. There's no way to prove that someone's, like, always reputable and never going to make a bad decision. Like, if you look at even in-trade, end-trade resolved a few markets wrong. And they're like the most popular prediction market that's existed so far. So I think the, you know, the economist and in me would say that I like the incentive mechanisms better of having an economic reputation as opposed to a more feel-good style one.

So let's come back on how this mechanism works. Exactly. So as, as predictions come to term, you have these reputation, the people that hold Recapitation tokens make, will vote on the outcome of that prediction. And one point that Brian made earlier when we were talking about this, which I thought was interesting, is that as this, as, you know, auger becomes larger, this may become hard to scale. And so you may start seeing similar things that happened in Bitcoin, for instance, which is, you

know, the pooling of miners. Well, perhaps you would see just pooling of those reputation tokens into one central authority which would be able to take in perhaps data feeds externally and sort of manage the voting on those predictions in a much more efficient way. Do you, is this something? How do you plan on keeping it so that it's, so that the reputations? tokens, holders, and the voting on the outcome of predictions stays decentralized when, you know, I guess the desired outcome is that Auger becomes some huge prediction market.

Right. So one thing that really helps with that is the randomized reporting. So the fact that you're only having to report on the subset of events. I mean, I've played around with the math a bit. You can get to a few hundred thousand offense a year and as an average reporter, even if someone with a good amount of rep like me, you're not going to have to work quite a whole lot. I think with the amount of rep I have, I'd have to work maybe a few hours every two months reporting on Auger if there were a couple hundred thousand events a year. Now if it scales even farther than that, so there's two things. With a couple hundred thousand events, you can get the vast majority of volume on the vast majority of things people want

to speculate in the world today. If you look at the Chicago Mercantile Exchange, they only have maybe 40 or 50 things you can trade on. The total amount of stocks in the U.S. is I think if you count the main publicly traded exchanges. I think it's maybe like around 10,000, 15,000. And so you can get

quite far with a not large amount of finance. It's still even farther beyond that. So people start using Oger for really micro things like the Walmart examples. Then, yeah, does become more concerned about pooling and things like that.

The nice thing about about auger is though since it's a very economic reputation the incentives still hold regardless of whether it's pooled or not and so basically that kind of idea is that if markets have these things called outstanding shares and so at the end of a market most people have probably sold their shares and exchanged them and got out of the market along before it's reported on but there may be a few percent of people who still have their outstanding shares held so the idea is that if at any point the value of the outstanding shares is worth more than the market

cap of reputation, then it becomes economically viable to do an attack on the network. The nice property, though, is that the network can fork, which means that it can split in two sets of reputation, and the idea is that the market, you know, outside of Augur, so ignoring argument markets would value the set of reputation that actually reflects the reality. So if one set says Obama won in 2012 and one said says McCain won, you only require a tiny bit of the efficient market hypothesis for the market to value the set that says Obama won higher. And so I think that like while pooling is a concern, it's more about the economic incentives rather than the exact like number of reporters. So let's say now I,

you know, I'm speculating different assets, right, different cryptocurrencies, et cetera. A lot of people do that, you know, some may buy, some maybe bought rep, right? And they, I mean, you know, you were talking about a few hours every, every few months. A few hours is a few hours, right? So I think most people would not want to do that, right? So, and if you look at something like Bitcoin, a lot of people hold their coins on exchanges and in hosted wallets, etc. And the reason there is really kind of a very minor gaining convenience and to take into account, you know, significant loss in security. And here it seems like the incentive is much bigger because a few hours of work, you know, let's say someone's time is worth $30.

right? That might be three hours of work even every two months, right? That might be $100, $50 a month, right? It might be $600 a year. That one would essentially be spending in time. Whereas if I can put it in some hosted service, that just does it for me. Don't you think it's inevitable that the vast majority of people are going to do that? Yeah, I think long-term people probably will do that. But it's like not something that, it's not something that I'm super worried about because the incentive still hold regardless of whether there's pooling. And there's always the final backstop of, you know, forking the network. So it's not something that I'm like super worried about, I guess is what I'd say. Okay. Because then the other thing that maybe is worth diving into a little bit about it.

So if you don't vote, you're going to get some punishment. Is that correct? Right. So the way it works is there's like, there's two types of rep. There's dormant rep, which is where you don't get any trading fees and you basically can't use it to report. It's basically meant so if you wanted to trade it, you switch it to dormant and then trade it and then switch it back to active. The other type of rep is active where you get trading fees, you have to report. So if you have your rep active and you're, you know, claiming fees and you don't report, you're going to lose part of your reputation.

And then you also have, if I remember correctly, you know, with the shelling coin idea, the point is also that if you're too far out from, if you're an outlier, right, if you don't kind of aggregate around the consensus, you also get punished, correct? That's right. So if there's a market on, you know, how many inches of rain will fall somewhere, and the median answer is 12.5 and you say 23, you're going to lose rep for doing that. Today's magic word is reputation.

R-E-P-U-T-A-T-I-O-N. Head over to Let's StockBitcoin.com to sign in, enter the magic word, and claim your part of the listener award. So you mentioned with the rain thing, you're an outlier, the consensus of 23, you wrote it 12, and now you get some punishment.

Now, my concern here too is, so let's say you have some pooling going on, some people put the funds in there. Now, and all of a sudden, 30, 40% of the reps are held by that, that, I mean, then their predictions really start to set that shelling point and they start

become that shelling point. So, I mean, I think number one is, right, if I want to vote on my own, it becomes more risky, the more people put their money into the service that kind of votes together. And so I think that's one thing that actually makes it less attractive to do that. And then the other concern, I guess, would be once they have 40%, 50% or something, couldn't they just dictate the events? And then what makes it that those reports would actually be the true value and not some other thing? So there's a few things.

So you don't actually know how the pool is going to report ahead of time because it's a commit and reveal. But there's still the valid point of, should I join the pool just so that, you know, I'll always be reporting with them anyway. And the answer to that is I think, you know, it's one of those things, 99% of the time it will be less risky. But the issue is, what if the pool messes up? So it's kind of like, it's kind of like, you know, playing rushing the rat with a lot of empty barrels in the chamber.

And once in a while, the pool may screw up. Someone may attack the pool and reports wrongly, or the pool operator may just do something malicious. And the problem is, once that happens, then the backstops kick in, and those would kind of screwing people in the pool over. And the reason is the first backstop is everyone reports on everything. The penalties go up for that. And say the pool has 60%, so even that doesn't work. And the network forks and two sets of reputation. And in this one, it's really not the reporters deciding things so much as much as the traders who are trading your reputation.

And the reason is because the traders should value the set that represents reality higher. And it's something that's like it does rely on efficient markets. But it's like if markets are so inefficient they can't choose something that clearly reflects reality, then maybe prediction markets aren't such a good idea in the first place. So it's something more comfortable relying on. But so in this case, let's say now 60% are in this pool. and they vote the wrong result. I mean, a fork, because 60% of the funds are in there,

it also means that the private keys for 60% of the tokens are in there. So how that fork, how would that actually happen? You wouldn't actually be able to fork then, no? So the way a fork works on augurs actually, you just basically post a bond of like 1% of rep. So you do need like 1% of people to basically either pull their money together or, you know, if there's some whale that has a bunch, then the whale could do it. And it can only be done after the second backstop of reporting on everything,

or everyone reporting on something. And so if someone posts the bond to fork, what happens is the people who, you know, the pool of 60% or whatever is going to lose, like, around 40% of their rep in the fork. The person who posts the bond is going to lose their rep in the original branch, and they're going to get double back in the fork. So the incentive there is for someone to do the fork so that they basically make money. And so it doesn't matter if they own 60% of rep because they're going to lose a lot of it in the fork. And the other thing is you don't need, it's not like, you know, Bitcoin or Ethereum where you need a, actually, I guess even in Bitcoin and Ethereum, you don't need a majority of a fork. The thing is people will just ignore you if you fork and it's a bad fork.

It's like say I wanted to fork Ethereum, and it was just over some random thing, say I lost my private key and I wanted to get my money back and I fork it, people would just ignore it. But with something like this, you know, if I forked the Augre Network, and it's because the network said that Trump won, even though he didn't, then there's quite a bit incentive for people to actually follow my fork as opposed to just ignoring it. If I have that choice, right?

But if my money is lying in the pool, I might actually not be able to access it at that point. Yeah, so that's the big danger of using a pool. That's what I meant when I said. It's less risky 99.99% of the time, right? But it's the one bullet in the chamber that you might accidentally get shot with if the pool messes up. So, yeah, the way of fork works is you have two sets of reputation, so you have one set in the pool. And your other set, you'd get it on the fork, except you would have less rep because you were. on the side that, you know, committed to the pre-fork report.

But yeah, I mean, you're certainly right that there would be, I mean, in a way, the risk of putting your rep into a pool is kind of comparable to the risk of leaving your coins in an exchange, right? Like, goes well, 99%. And yeah, sure, you may be a risk. But yeah, I guess my concern is just that in this case, it seems like the incentives are so strong to do that much stronger than with Bitcoin. At least that concerns me a little. But anyway, Sebastian, you had a question. Yeah, so let's talk about the user experience side. So can you tell us how, you know, what, what is a typical guy like me who wants to use a prediction market like Auger have to do to use it, basically? So the way it kind of works is we're making the

UI basically a web interface. So you'd go to a website and download it's all the code, basically to your browser, and it's all around client side from there. If you had an Ethereum node or in the background, it would use that. Otherwise, it would ping, I think it uses EtherCamps, posted nodes. So what it does is it, all your transactions

are signed in the browser and they're submitted to the network. So if you wanted to trade on something, you would, you know, there's two things you could do. One is you could search for a market. So if you have a real specific idea of you want to trade on, say you're looking to trade on a presidential election market, you might search presidential elections

or 2016 US elections, something like that. If you're just kind of browsing around, there's a category system where you can basically kind of filter by different tags that people can choose for their markets. And there's also filters by volume and fees. So as a trader, you're going to want to search for like lowest fees, highest volume markets, because that's where you're going to be able to trade the most to get the best prices. And what you would do is you click on the market you like and it'll be multiple outcomes. If it's a binary market, just two outcomes. And you can, you know, buy yes that it will happen or buy that you think it's not going to happen and there's an order book and you can input how many shares you want to buy the max amount you're willing to spend. So basically it's limited orders. Market orders for something like Auger aren't really a super good idea because like while you can pick up orders on the book, it can be very dangerous in like early days when there's not a lot of liquidity. You know, if I place a market order on the New York Stock Exchange, that's fine. I place it on Auger. I could be buying. I could be buying. I could be buying. You know, you know, if I place it on Auger, I could be buying. You know, you know, I could be buying. You know, you know, I could be

at 99 cents a share if I'm not careful. So it uses limit orders by default. You'll click buy. It'll commit your trade and then it will reveal it a block later. It does that to prevent front running. So a miner can't just sit there and front run all the trades. And then you'll have your shares. And then so you can do two things. You can hold them or you can sell them to another person. Or the third thing you can do, which is kind of weird that I don't think most people will do. maybe like the professionals will, which is you can buy the other outcome and then exchange it for a complete set.

And the reason is in a binary market, you have two outcomes, yes and no. They combine, they're worth one. So if you own the yes and you buy the no, then you can exchange them for a complete set and get one back and exit immediately. Or you can just sell your no or sell your yes as someone else. Okay, that's interesting.

So either you hold it. I mean, so there's three outcomes. So either you hold your coins, either you can sell, or shares, you can sell those shares to someone else. Or you can just, you could buy the whole, the whole outcome and have that as the basis for, to basically bet on another prediction. Yeah, it's kind of like that. So it's like, say we had, keep it simple, say we have presidential election market and the two outcomes are Hillary and Trump. and say you're long Hillary. So you bought Hillary at 60 cents a share

and she goes up to 80. Say Trump's been saying some really stupid stuff. She goes up to 80% and you want to sell. So you can just sell it to someone else. The other thing you can do is you can buy Trump for 20 and then sell the complete set. And the reason is because the reason the complete sets exist at all is

if you think about traditional financial markets like the New York Stock Exchange, Apple shares already exist. they may issue more at some point in the future. They may buy some back. They're probably not going to go bankrupt to get acquired. So the shares are already there. But with prediction markets, the shares don't exist.

If you make a new market on Auger on the presidential election, the shares kind of have to be created. And then where they're created is with this complete sets thing. So if I'm a market trader or market maker and I want to provide some liquidity to a market, I may spend $500 by $500 complete sets and then offer them on the order book for people to buy. Let's take a short break and talk about hi.m. Me. Hi, dot me is a VPN provider. And if you don't know yet why you should need a VPN provider, let us help you. I'm sure you were like me and when all the crazy revelations came out during the Snowden time of all the spying that is being done by the NSA and other government agencies, you were shocked and you said, not with me, not with my own rights. Now, the way government agencies can spy on you, there's many of them, but the most, the most easiest way is by simply going to your ISP and getting all your traffic capturing all your traffic and the VPN can protect you from that it can give you a secure tunnel from your computer

to any of the exit nodes all over the world so that all your traffic goes to this secure pipe that's encrypted and cannot be intruded on and with hiked.m. you can choose any of their 30 exit nodes all over the world so you can enter the internet in a secure location The best thing about Hyde.me is that they have a free plan, which includes two gigabytes of unthrottled bandwidth per month. So you can go to Hyde.combe slash Epicenter to create your free account. And when you use that URL, you'll automatically get 35% off if ever you decide to go premium. Now the premium plans are really great. They include unlimited bandwidth, access to all of the 30 exit nodes that Hyde.Me provides,

and you can install it on up to five devices at a time so you can have this running on your phone, your tablet, your computer at work, your personal computer, and just be completely protected all the time. And of course, hi.me accepts Bitcoin. So we'd like to thank high. dot me for the support of Epicenter Bitcoin. I'd like to come back to the software. So you mentioned that it's a web app.

This means that I can run the Auger client locally and I can connect it to my own ether node and I can connect it to the contract and make those commits myself. and fund the contract myself basically from my own machine, right? Right. Okay. So then this brings me to the next question.

Then what is the business model for Auger? Where do you guys fit in all this? So, you know, like basically we're just, we're reporters in the system. So when we did the rep distribution, 16% of it's distributed amongst like around 20 people, founders and advisors and employees. And 4% is held by the foundation, which is used for things like we've given some to people in the community who have done cool stuff. Like someone made it like a bunch of auger shirts. We gave him a little bit of rep.

Things like that. We'll probably give away some of that and like bug bounties. And so that's kind of, you know, right now we're mostly focused on building the protocol and the platform and getting it out there. And then later on we'd be focused on building more like possibly like for-profit ventures around it. There's some really cool things you can do with that. If you think of the UI we have right now, it's not super easy to use because you have to do lots of stuff to make it be 100% decentralized. So we can't score the user's private keys. That means you can't have something like blockchain.info.

Or we can't store it really a cache of the data so we have to rely on the community to do it. So that makes it kind of flow. So there's lots of things like that that for profit could potentially build like a UI around it that's easier to use. around it that's easier to use. Maybe you could do something really cool. Like one idea I've been playing around with this, the idea of having like ads in the UI, but giving part of the revenue from them

to the people who trade in the UI to subsidize their trades. So like if you use that UI, you basically have lower trading fees. That's kind of cool stuff that you can do with crypto that isn't really feasible with like, you know, say PayPal or something. Okay. I understand that. Now, I wanted to talk also about security.

So this is something that has sort of come up since the DAO attack, especially around Ethereum, and, you know, the ability of these contracts written in solidity. Have you taken anything away from that? What have you learned from the DAO hack? And what are the some of the things that Auger is doing differently from that project to ensure that this type of thing wouldn't happen in the DIA? in the auger smart contract. Right. Yeah, so, yeah, I paid attention pretty closely to the Dow. Once they got attacked, I didn't really follow it too closely until then.

And so there's quite a few things. So if you look at, you know, the types of vulnerabilities that occurred in the Dow, yeah, like the call dot value issues where you're just kind of sending gas to a contract that may or may not be trusted. in the haste of the attacker's contract, it was definitely not to be trusted. And so there's lots of things we're doing to kind of mitigate that. We try to use send wherever possible as opposed to call that value to send only sends a minimal amount of gas. And then all the auger contracts themselves,

some of the stuff has already always been planned. Like we want to have a white list on them so that we can basically trust the auger contracts to communicate with each other, but it doesn't mean we can trust others. Just things like Mutexas, which is the idea where you make it so your contract can't be recursively called. So if you look at the DAO, what happened was basically they just kept recalling the split DAV function over and over again. So you can prevent that with a thing called Mutexes where you basically just check if the function's been called more than once. And if it has, you don't let it proceed. And then some of the other stuff we've been doing is I've been reading all the NASA and JPL programming guidelines.

the MISR ones from the automobile industry. And then I've been implementing them all in our codebase. So there's lots of stuff you can do to make it so your codebase is more simple and easier to follow. And there's basically we've been implementing all those security guidelines in our codebase to, you know, give it as best a chance as possible as it can have. And then another thing we'll be doing is lots of bug bounties. I think if the Dow had bounties, maybe if you have like a $10,000 bounty, I think there's a decent chance that the bug would have been found, or the vulnerability would have been found.

So I think stuff like that is key. And then the final thing is making it so that you can recover from an attack or a hack without relying on forking Ethereum is a huge lesson. So it's something like Argo got hacked. You'd want it so that we can upgrade the contracts quickly and resolve it on contract as opposed to hoping that Ethereum will fork around us. You mentioned bug bounties.

Are people finding bugs in auger and claiming those bug bounties? So we've only started one for like the, in this contract that we're going to use for like the Rep Distribution. We started a bounty for that. I need to increase it. It's like at $1,000 right now. I need to make it like $10,000. Nobody's not any bugs yet.

So I probably have a post about that soon. And then the, the, the, core contracts on chain. We haven't started bug bounties for those yet because I'm still fixing some stuff, basically. I'm still making too much of edits to them to make them comply with the NASA guidelines and the automobile guidelines. I don't want to start the bounties until after that's done. And of course, one should say in you guys sort of defense, depends maybe the wrong word, but I mean, so the crowd sale was in 2014.

now it's two years later or around two years later and you know you're sort of still working on it right whereas the Dow guys was like let's just go ahead no raise 150 million so i i think they were a little bit particular in that regard um speaking of that what's the timeline here when is augur going to launch and and are there going to be different stages to that yeah so so the timeline So we did about a year of development from 2014 to 2015. Then we did the crowd sale. It started like maybe 10 days from now a year ago today. And so now it's like a year later.

And yeah, we've been taking our time, just, you know, iterating on it. As far as timeline, what we're looking to do is basically enter security audits once I finish getting this stuff done to make it comply with a NASA guideline. There aren't really any more features to add to the back end besides security ones, which we're taking our time with. So we should be able to start security audits relatively soon. We're aiming to launch a new version of the beta in a couple of weeks that'll have like limit orders and some of the new reporting stuff from the back end in an order book, whereas the old beta only had a market scoring rule. And so, yeah, we're basically taking our time to try to get it as secure as possible, it's as few bugs as possible. And then as far as like phases for the launch, I think the way we would do it is initially

start out where like we have maybe like a multi-sig that allows us to modify the contracts very quickly. So in the early days, something goes wrong, we can push a fix. Later on, you know, once the system matures, then we we could have people, have the reporters basically vote on updates to the contracts. So that's kind of like a difference between, you know, when it would first launch and later on. As opposed to like from a feature level, it'll have binary markets, scalar markets, and categorical markets from day one.

We removed multidimensional markets after the Dow attack because we thought they were too complicated and we decided to just avoid that. And that's kind of where it's at. So I wanted to come back to a topic that maybe we should have discussed earlier, but I guess that this is a good time to talk about. So the ethical aspects of prediction markets. And they have been criticized. I mean, generally, prediction markets have been criticized for this, is that the ability

to have predictions on things like deaths, on things like assassinations, on things like terrorist attacks occurring or not occurring. With something like a decentralized prediction market, of course, it makes a lot harder to regulate whether or not those predictions can make it onto the market or not. What is your stance on that? Do you think that there should be norms around those kind of things, or should we just let people make predictions about whether or the hell they want? I think there should be norms.

So there's like two answers to this question. One's a cop-out, but a practical answer, and one is like how it actually works on Auger. So on Auger, reporters basically can judge whether a market is unethical or not. So when you report, you can also submit whether you think a market's unethical. I suspect that only like very egregious things will get reported as unethical. So things like assassination markets or markets that would cause harm to other people. Like if someone made a market to incentivize robbing a bank or killing someone or something,

I'd hope that the reporters would report that as unethical. I certainly would. From a practical standpoint, I think the concerns are a bit overblown because in traditional financial markets today, these incentives are all at play, and nothing really happens. So if you look at,

so the example of this is, take any Fortune 500 company. You can buy options that expire within a few days. If you were to assassinate the CEO and buy put options the day before, you can make way more money than you could probably make on auger doing something similar. But it doesn't happen. And so these kind of incentives already exist in the real world.

And, you know, it's a nice, I guess it's a nice thing to make you have faith in humanity, I guess, that they don't actually happen. And so there's like, there's also like the idea is like you could also make way better ways to incentivize this sort of stuff on Ethereum, which is kind of another cop-out answer. prediction markets are not the best way to do this. If you did want to kill someone, you could use a dominant insurance contract where all the money goes to the person who actually kills them. On a prediction market, people can buy the other side,

which means that you're not actually doing it in the most economically efficient way because you're giving a lot of money to people who didn't actually do the job. That's like a very cop-out answer, but it's definitely true. Yeah, actually, so I was watching this panel, as I mentioned earlier, your co-founder, Jack Peterson, mentioned that you could have a vote on

whether or not a question was, a prediction was ethical or not. I mean, then you mentioned earlier that reporters will set the bar and define whether a question is ethical or not. Then the question becomes, you know, where do you set that bar? That bar may differ from person to person or from culture to culture. If you could have a way to have votes on whether or not questions were ethical, then you're essentially layering on like the prediction on top of the prediction and perhaps getting a better representation of whether or not a question was ethical. I mean, I think this would only really be useful at like large scales, as you mentioned.

It's a lot easier to make money. There's better ways that you can make money than to go assassinate someone and make a prediction. about it on the prediction market. But on some, like, you know, at the edges, there may be other, perhaps not so, you know, predictions that wouldn't be so grave, but could potentially be considered unethical in certain places where you might want to get some sort of wisdom of the crowds around the ethical aspects of that prediction. Yeah, I think with like, with,

Because the reporting system is kind of so global and distributed, you know, at least now ignoring the pool problem, you kind of really get a global set of norms. It's like the things that people consider unethical are probably going to be only things that are kind of globally considered unethical. So like most cultures are considered, you know, assassinating someone to be unethical. most considered stealing from someone to be unethical. But there's other way less grave things that might not be considered unethical that might make people uneasy. I can't think of any off the top of my head.

I guess maybe like a geopolitical market. Something like the CIA would love to have, like will Russia invade Estonia within the next five years? I don't think that would get reported as unethical and auger. That's like a good example of one that's kind of great. But to the Estonians would. the Estodians would probably consider that to be an ethical.

They'd probably, I don't know, they're a pretty technologically advanced country. I bet they'd want a market like that so they could have better odds, you know, better probability of knowing what's coming or not. Yeah, and of course here, the kind of danger only will be, well, if Russia says, we're really low on money, we need to make money somewhere, oh, there's this big prediction market on us invading Estonia. let's bet they're invaded and take that part, which is, of course, an absurd scenario, right? So that's not going to happen.

So, all right. Yes. Yes, for the assert. And it's also like, it's also like relies on the fact that, you know, people in the Russian government want an insider trade and front run those trades, which, which you know would happen, obviously. Yeah. Yeah. Now, one last topic we do have and we want to talk about it. And it's also a topic that we kind of revisit very, very regularly.

And that's the kind of topic of governance. And again, our listeners probably will not need much of an introduction to that. But the best example, I think, of governance that isn't really working so well. It's Bitcoin, right, where there's a lot of division and no agreement on how to evolve the protocol and no way to come to an agreement here. Now, presumably augurs are also going to need to evolve and to be upgraded, etc. So is there going to be a process for that? How will these decisions get made?

Yeah, so right now it's a very simple governance metric. It's basically just reporters can vote on upgrades to the contracts. So the way it kind of works is there's a registry contract, which maintains the latest version of each of the contracts, and reporters can vote on whether to update them or not. you know, longer term, you have more interesting things like futurearchy and things like that. But for the early days of the system, we're kind of trying to keep it simple, as opposed to, you know, going with something that's probably better in theory,

but much more complicated, I guess. So you guys would propose a new contract, and then the reporters would essentially vote on, you know, do they want to improve that or not? Yeah, yep, basically. Or, you know, if I propose a contract and everyone hates it, some random person from the community can propose another contract that, you know, modifies two lines and the reporters like better or something.

Okay, great. Well, Joey, with that, I think we're at the end of the episode. Thanks so much for coming on. I think it was very interesting to dive into Augur to hear a little bit about how you guys are approaching things. So thanks so much. Yeah, thanks for having me. And of course we are very much looking forward to, you know, seeing it in action. We were playing around a little bit with the beta before.

So that's possible to do, of course. So we will have links to the Argo website, to the application, the beta application where you can play around with test coins and to some other resources. If people want to learn more about Auger. And of course, we'll also link to some of our previous episodes on prediction markets if people want to dive more into that. So thanks so much for coming on and thanks so much to our listeners

for listening. So we are part of the Let's So Bitcoin Network. You can find this show and lots of other shows on Let's Stop Bitcoin.com. So we put these episodes out every Monday. You can subscribe to the show on your favorite listening app, whether that's on mobile or

SoundCloud or whether you'd like to watch the videos on YouTube.com slash episode of Bitcoin. So thank so much and we look forward to being back next week.