Epicenter - Learn about Crypto, Blockchain, Ethereum, Bitcoin and Distributed Technologies - Robin Hanson: Futarchy – Prediction Markets and the Challenge of Disruptive Technology

Episode Date: September 28, 2015

When invented the concept of prediction markets almost thirty years ago, he felt he had stumbled on a concept with huge implications. By allowing people to bet on the likelihood of future events, prediction markets promise to allow better forecasts and better decision making. Research into the area has been vibrant, culminating in Hanson’s concept of Futarchy: A prediction-market based governance system. At the same time, the real-world applications have been few and far. Hanson, an associate professor of economics at George Mason University, joined us to discuss his invention, futarchy and the challenges of disruptive technology. Topics covered in this episode: How prediction markets can surface information The history of prediction markets How futarchy works Whether futarchy could settle the blocksize debate Why prediction markets failed to get wide adoption How Bitcoin is facing a similar adoption challenge Episode links: Science 'The Promise of Prediction Markets' (PDF) Futarchy: Vote Values, But Bet Beliefs Robin Hanson's Blog Overcoming Bias This episode is hosted by Brian Fabian Crain and Meher Roy. Show notes and listening options: epicenter.tv/098

This is Epicenter Bitcoin episode 98 with guest Robin Hansen. This episode of Epicenter Bitcoin is brought you by hide.combe. Protect yourself against hackers and safeguard your identity online with a first class VPN. Go to hide.combe slash epicenter and sign up for your free account today. And by ShapeShift, with no account or signup required, it's the easiest way to buy and sell gems, counterparty, dogecoin, dash, and other leading cryptocurrencies. Go to Shapeshift.combe. to instantly convert your altcoins and to discover the future of cryptocurrency exchanges.

Welcome to Epson of Bitcoin, the show which talks about the technologies, projects and startups driving decentralization in the global cryptocurrency revolution. My name is Brian Kravian Kray. And I'm Meher Roy. Today we have a special guest joining us from Washington. His name is Dr. Robin Hansen. And for those of you who don't know him, he could think of him approximately as the Nick Sabo of the prediction market world.

He has played a role in the technology of prediction markets for the past 20 years, developed some of the main inventions that underlie modern prediction markets, and actually experimented with different kinds of prediction markets in different areas, like with government, with scientific predictions, etc. We'll be talking today about how prediction markets and Bitcoins can, and blockchain technology is linked and his experiences developing these solutions for the past 20 years. Dr. Hansen, we are pleased to welcome you.

Thanks for having me. It's great to be here. Yeah, actually, it's funny because I was, I was, we have been in contact with this guy and about coming on the show. His name is Mark Miller and he's sort of deep in the smart contract space and he sent me this talk that I was watching last night from 1997 talking about smart contracts. And then in the talk, sort of at one point, the camera shrubs over and it shows Nick Salvo, which is sort of funny, right, because Nick Saabu is thought of as this mysterious person with no face.

But he's like, there, one minute or 30 seconds or something, the camera is on his face. Did you see Nick Saabo? It's like, well, not so anonymous, I guess. And then in the end, you come up, Robin, and you ask some questions. So it's like, wow. So 1997, you were already thinking about smart contracts in this space. So, I'm not that young. Yeah, I've been around a while.

Well, not that young, but you were also just there at that sort of nascent stage. And another thing is when I was reading through your blog, you see Hal Finney commenting there in 2008 and 2009, which of course is when Bitcoin was created. And Hal Finney being the first recipient of the very first Bitcoin transaction. So in some tangential way, you have been around. that Bitcoin cryptocurrency blockchain space for a long time. I'm honored to have known these people and still honored. So perhaps we could begin with what is the prediction market and in your words understand what does it do and what is it good at? Right. A prediction market is really just a betting

market but we give it a different name maybe when it has a different purpose. So an ordinary betting market on a football game. give us the game and that somebody's going to win or lose and you could put money on it and your money will increase if you're right if you bet on the team that wins and it'll go away if you will bet on the team that loses. So you make bets and a betting market is a place where people can make offers to bet so that the odds adjust to some sort of consensus of what people betting think about the chance of the team winning or losing. A prediction market is a lot like a betting market or really stock markets, currency markets, etc. They're all really the same kind of mechanism.

We call them speculative markets. That is something you can buy or sell. And the people who buy or sell estimate the value of the thing and the price ends up being an aggregate estimate of what many people think of the value of the asset. And you can try to make money by buying today, buying low today and selling high tomorrow. And a great many of most of the people participating in most financial markets that are thick are actually speculators. They're there to just buy today, sell tomorrow or in a month, and hope to make a profit on the difference. They're not actually wanting to hold the thing for a long time. They're just adding to the information and the price. So a prediction market is just a market like that, but applied to something else where the reason you create the market is because

you want to know the answer to something. So you could create a market in a company, say you might have a deadline of a project. Will we make the deadline? You might have a let me, let me, let me people bet on whether you'll make the deadline. And the net effect of people betting will give you some consensus estimate of the chance you make the deadline, say 60%. And that could be, and often is a much better estimate of the chance of making the deadline than you will get by asking the manager of the project or doing a survey or many other mechanisms for finding out the chance that it'll happen. So a prediction market is a betting market on something you care about so that you bother to make the market in order to find the answer to your

question. So how did you become interest in that and what do you find so fascinating about this? Well, so I was a, you know, physics student long ago as an undergraduate and then I got interested in the, you know, physics teachers tell you claims about how like science is different than everything else and everything, you know, the rest of the world is a big mess, but in science, we've got this way to figure out the truth. Not always, always appealed to me. And then digging into it, you try to say, okay, what is it about science that's giving us the truth that other people can't get? And you get various mumbo-jumbo, but I wanted to figure that out. So I finally went into philosophy of science, graduate school, say, well, you know, what is all this stuff? And in the

end, it turns out it's a big, complicated social institution that doesn't have any particular simple answers for why it works. And there's just really all sorts of things can go wrong with it. And then you finally realize in the world, there's all these people arguing about global warming and policy and everything else. And you realize that these things don't work very well. that the institutions we have for figuring out what we should think about the Pope or Russia or China or anything else, they're just messed up in all sorts of ways. And so you get interested in, well, how could we fix that? And I was involved with the World Wide Web in the very early years before there was a World Wide Web,

but with Project Zanidu, and their vision for making the world better was to say, well, debate would be better if you could find rebuttals of things. So their simple theory was, we have bad beliefs about things because people say stupid things and you can't find the rebuttals somewhere else. So they were hoping that backlinks on the web would make it easy to find rebuttals and that would fix our bad debate outcomes. And so I was hanging out with them in Silicon Valley in the very early late 80s and listening to this vision and it was exciting. But then eventually decided maybe that doesn't work so well. Back backlinks seemed to me a little, they wouldn't be sufficient to make debate work and to make people not believe stupid things.

And so I wondered, gee, what would work? And so hanging around this world, they were mostly libertarians, I naturally, I think, came on the idea, well, what if we were to bet on these things? If we were to bet on these various issues we were debating, wouldn't that create good incentives and even in scientific questions? And so I began to explore that idea back then and started to, you know, write about it and give talks on it. And I find that I was basically a nobody. I was just a software engineer, research software engineer in Silicon Valley. And so I decided I would be going to go back to school and get a PhD so that I could have context and credentials to talk about things like that, which I eventually did. And so eventually that became part of, you know, my agenda to, in my research agenda and elsewhere,

that we could improve the world by taking the issues we argue about and then bet on them in addition to just arguing, and that there was a huge potential to apply that to a very wide range of topics that we weren't exploring. In the past, what kind of prediction markets have you experimented it? What are some of the projects that you have done with prediction markets, and what came out of these projects? Well, what we know in total about prediction markets is, of course, course, the sum total of all the different related projects that have happened over the last 30 years,

say, because that's the period of time where which people have taken this idea seriously, you know, the projects I've been personally involved with are a small fraction of all those projects. And so, you know, we learned from all these projects. But I originally was motivated to deal with science and public policy questions. And so I originally talked about doing that on these things. And then, you know, I was trying to explore how to do that. The very first I got anybody to do was Xanidu group had some internal markets there at the company on whether they were going to make their deadlines for delivering software. And they also had some science questions on that.

And so that were the first internal corporate markets in the early 1990s at Zanadu. Then a group of people who were software engineers read some things I had written and then wrote one of the first web markets, I think it was the first web market, and they called it Ideosphere. Now, at least that's the name. Now, you can still look it up at idiosphere.com. And that was on science and technology questions, longer term. And then other people have, you developed other predictions. There was the Hollywood Stock Exchange that was developed around then and got popular, and it was betting on movie markets. A number of smaller private markets were tried to create it, say within pharmaceutical companies. Then in the, around 19, late 1990s,

I got involved with DARPA who set up a call for proposals to try out markets in the Defense Department, and that created the policy analysis market project that I was involved with, and that was exploring technology to apply to the Defense Department. And in the last few years, I've been involved in an IARPA-funded project to apply these technologies or to explore them for intelligence gathering and first in foreign affairs and then in science and technology. But over the years, I've been involved in consulting with lots of companies and projects trying to apply prediction markets inside various organizations. So I've heard a lot about a lot of projects that I wasn't greatly involved in.

And so your question was, what have we learned? Which is a broad question. So, I mean... So meaning like perhaps, perhaps it's easy for me to specify the question a little further. Like, there's two parts to the questions. A, what have we learned about the accuracy of prediction markets? And B, what have we learned? What are the reasons why?

prediction markets have not succeeded on a on a major scale and the and the challenges faced by some of your projects in specific so you know the idea of a prediction market just to be clear is you have a question and there's some way in which later on you will know the answer to the question and but you want an estimate on the question now and so in order to get that an answer to your question now you set up a betting market basically and the and you subsidize it sufficiently so that the consensus price will exist and enough activity will be. And then you can look at that current price as the estimate to you for future questions. So that idea has excited a lot of people

and drawn a lot of interest. And so a wide range of people have been interested in talking about that idea and exploring it. And that's produced a set of research in academia where people have tried to test the idea in lab experiments and test some problems that could occur in lab experiments. and then develop the technology, i.e. work on mechanics and tools that would make various problems less and very harder cases easier, etc. And there still is a large intellectual and academic pursuit of the idea that's somewhat separate from real organizations who have actual questions applying it to their questions inside their organizations. So that second path, unfortunately, is much less well-funded and difficult, but it's the path that needs to happen. So when organizations like DARPA or IARPA have decided, or even other grant agencies, decided to fund work in the area, they typically fund researchers to go study things and develop tools, but they aren't very interested in funding,

like real organizations to actually try it out and figure out their real problems. So most of the applications in real organizations have been done on, you know, shoestring budgets and by small groups, usually by somebody who just excited by the idea and finds an opportunity to try it out. So in those projects, they usually, of course, wisely ignore most of the complexity of the elaborate mechanisms that researchers have developed. And they, you know, don't go for the sophisticated problems or deep, you know, obstacles or, are, you know, issues. They just opportunistically try the simplest things they can, simple mechanism, simple questions, and simple applications. And they have had a somewhat consistent experience, I would say, which is that, well, first of all, the typical person who wants to set one of these things up usually doesn't have much of a budget. They just have a lot of enthusiasm. And so they might, you know, set up a market, throw some questions out and offer no, not really much incentives. Other than it's just play money or something.

And then when something like that happens, they don't really even have strong organizational support. And they just say, hey, come and play on your lunch hour or something. And, you know, the typical thing that happens there is it crashes birds. Nothing happens. Nobody trades. Nobody's interested. It doesn't do anything. But there have also been markets where somebody hiring the organization had authority.

some budget and could push it and say, you know, encourage people to participate and offer larger incentives financially and also just praise and organizational recognition. And those markets have typically produced accurate estimates in the sense that whenever they've compared the market estimates to some other source they had for estimating the same thing, at the same time, the markets have done as well or better. So there's been a very consistent track record of accuracy in these markets. also a consistent track record of user satisfaction. That is, if they ask people, what do you think about it?

Are you enjoying it? You feel like your voice is being heard, et cetera. People send to like those two. But in addition, there's also been the negative part of this track record is that if you go and consult for these organizations, you say, great, you want to set up a prediction market, let's talk about what questions are important to your organization. And where could the high value be in your organization that you could achieve by asking important questions?

they immediately shy away from the most obvious important questions saying, well, that's a little sensitive and we don't want to start right there, and that could make people upset. And so they shy off to the side of relatively safe, interesting, engaging questions, but questions are not really going to bother anybody. So what kind of questions do they shy away from? Well, you know, the most straightforward of what are your big projects? Are they going to happen on time?

Are they going to succeed in delivering the benefits promised? You know, what are the major products and what are the sales going to be? You know, sort of who are you going to put in charge of the products and whether they make a difference? Are you going to change the requirements, definitions, you know, the key big questions on projects and products. And again, those tend to be sensitive. Those tend to have somebody who, and so, for example, one of the most consistent set of accuracy success, in these markets is estimating whether a project will happen on time. Just a simple deadline forecast. And there's been a lot of examples where prediction markets were set up and the

official management forecast of whether they make the deadline was optimistic and yeah, we're pretty sure to do it. It looks like we're on track. And then you open the prediction market and the odds just go down to a few percent saying, no way. It's not going to happen. And those are, of course, consistently right. They weren't going to happen. But usually this embarrasses somebody, the guy in charge the project. It makes him look bad. He was saying it was going to happen and the market say no way and of course it doesn't happen. And you might think the management above that layer would be grateful to hear this news from a new source, an independent source. But in fact, usually the guy who's mad that the project like made him look bad, complains about it and gets

it killed. And so, you know, and his manager is part of getting it killed. He's saying, no, we don't want this sort of embarrassment to happen. And the project goes away. So you mentioned that there's been a lot of the research interest, and there hasn't been so much on these actual projects that then drive decisions. Is this why? Do you think it just makes people uncomfortable and it disrupts too much existing organizational structures that people just don't want to do it? Well, I mean, this is the key question. So that's why we want to dive in a little detail. I mean, we have a technology here, and we want to apply it in real organizations, and we want to understand the barriers to adoption.

And in general, when technologies have to integrate with organizations, part of the barrier to adoption is that your initial concept of the technology and the way it's implemented was based on some abstract concept of its use that wasn't very, didn't accommodate much the actual details of how these things are actually used and the actual issues. So, I mean, this is true for almost all technologies. If you don't have a very good image of how it will actually be used, and you create some prototype versions of the products and you try to get people to use them, you often find that you just misunderstood how these things are used and what are the important features of the product, what are the important costs and benefits, etc. This product is a product to produce information inside organizations,

and that's a kind of product that we have especially poor understanding of and especially prone to lie about. There's a lot of complexity in how organizations produce and share and use information. And the simple idea that if you just tell them that their project's going to fail and they'd want to hear it is somewhat naive relative to the real political barriers and processes going on in organizations. So I think fundamentally academics and idealists sort of start with their initial concept of what these things are for and useful for. And then they just make a product and then they often, as engineers want to do, just throw it over the fence and say, okay, business, now it's your job to figure out what to do with this.

And at least half of the real engineering task for any product is thinking about how it actually gets used relative to your conceptions and working out the details to make sure that it actually can be valuable. So I would say prediction markets have not yet gone very far in the distance of figuring out what the real desires and needs for information organizations are, figuring out what the real problems and barriers using information organization and finding ways to change and adapt the product so that it can give the things people really want to need. So when you came up with prediction markets and you started experiencing. that idea, what were your expectations? Where did you think this was going to lead? Did you have, did you think this was going to turn out quite differently? Well, I mean, I was young and idealistic, and I'm perhaps definitely older and perhaps a bit less idealistic, but still substantially so. you know, I thought perhaps naively that there was just more broad-based enthusiasm for the generic idea of knowing answers to important questions better. That is, we're in a complicated world, we have all these big questions that matter for our organizations and national policy and all over the place.

And a lot of the barriers to making good decisions is not knowing answers to key questions. and I would have thought that there would be a lot more enthusiasm for ways to get answers to those questions. But that's based on a naive conception of what people's concerns are and how these processes are going on. So, in fact, people are really, you know, focused a lot on how things make them look and how they can support their allies and oppose their rivals and show their loyalty and show. show their ideological commitment and show their, you know, a wide range of complicated social processes that information is involved in. And prediction markets are just naively, I mean, it's basically an autistic sort of lack of social savvy.

They just sort of blunder in and tell the truth when that's not the socially acceptable thing to do. Let's take a short break to talk about our brand new sponsors, hide.me, personally been using them for about a year. so I'm really excited to have them on. You know, we sometimes take our privacy and security online for granted. I know that I did. I often tell people, if you use public Wi-Fi,

you might as well assume that your data has been compromised in some way. There's so many ways people can attack you nowadays. I mean, if you're using a website that isn't SSL, people can on the same network as you can pretty much see anything that you're doing on that website. And even SSL websites, like your bank, social media, or Bitcoin wallets, for example, can be vulnerable to certain types of attacks. So, you know, as a Bitcoin user, in your office or in a co-working space or like if you're a public Wi-Fi, someone could potentially target you just based specifically on the websites that you visit. Now, you want to protect yourself against that. And to do that, you need hide.m.

Hide.mee gives you an encrypted connection between your device and their network of servers. So attackers and even your ISB have no idea what you're doing. This all happens over super fast gigabit Ethernet, so there's no lags. And you have an encrypted tunnel there, which protects you. and in addition to that, Hyde.Me keeps no logs of any of your activity. And the great thing about Hyde.comi is that they have a free plan. The free plan gives you up to two gigabytes of data per month

at unthuddled bandwidth, and that's just enough so you can protect yourself whenever you use a cafe, when you're traveling, you're on an airplane, in an airport, or use any public Wi-Fi spot. And you can sign up and get that free account when you go to Hyde.me slash epicenter. The great thing too is that if you ever decide later to get a premium account, then signing up with that URL is going to get you 35% off.

The premium account includes unlimited bandwidth, access to all their servers worldwide, and they've got lots of them. And it also lets you connect up to five devices simultaneously. You get it on your mobile phone and with all your devices. And of course, you can pay with Bitcoin. I would like to welcome hi.combe as a new sponsor. We're excited with what they're doing. and of course we would like to thank them for their support of Epicenter Bitcoin.

So in the Bitcoin and in the cryptocurrency field, there's this belief that prediction markets have not succeeded because there, like you say, there have always been vested parties that do not want to see or have public the results of what these markets are of these markets. because as you say, they can sometimes be autistic. They give the bad truths to your face. There's this belief in the field that if you make a decentralized prediction market, that is censorship resistant, then the market can go on, even when the person whose life would be influenced by the market doesn't want it to go on. Or the market will go on, when the government that will be influenced by the market doesn't actually want the market to go on.

Will this kind of censorship resistance actually help prediction markets go into the mainstream? What do you think? Well, I think people overestimate the degree to which there is a vast pool of rebels waiting to resist the repression of authorities. Just eager for any chance to show their faith in the ideals of their society. it's a lot weaker than that. So, you know, in most organizations, you know, there are people who at lower levels might want to be honest and tell the truth, who might want to even pay a bit of a cost career-wise or socially in order to say honestly what they think about the organization. But it's easy to weigh overestimate just how enthused they are about that and how eager they are to do that.

So, you know, if in an organization, we could create this prediction, the organization could pay to create this prediction market on the deadline. And then, you know, somebody gets embarrassed and they cut off the funding for that prediction market. Now we'd say, well, what if employees lower in the organization? They can just go off to this website and make and make a market in whether this project will make the deadline. And the authorities can't shut that market off. And so it'll just, you know, shout the truth. Well, I mean, in a sense, you know, they could have already been doing that just going on to an external website with some sort of anonymous market. So it's not necessarily that they couldn't have already done that.

But they show relatively little interest in bothering. I mean, they could have just done that by setting up a chat site outside the company where they just talked about whether the projects was to make the deadline. And they don't bother to do that either. So, yes, of course, an outside chat site in the firm could let people talk about whether they could. the project will make the deadline. They could have set that up. They could have had a secret meeting where they all came later on and talked about it. They could have a prediction market outside that's a play money market or they could have a real money, you know, blockchain-based prediction market in a few years. But the question is, well,

will they want to bother? I mean, it's not enough just to make it possible. You have to make something people care enough about to bother. I guess the use case people think more about this, maybe not so much inside organizations, but then when you talk about, you know, political things. I guess that's also where in trade, the prediction market that has been most successful so far, right? The politics was very big on their

elections. Right. So maybe in something like that, do you think, I guess the question then becomes too, right? Is this just a betting site? And then what's a prediction market about it, you know? If it's just making...

Let's talk about that. So in the world of some place like in trade, you know, it was offshore, but they had, they could do many questions. And so a place like in trade could have had betting markets, all sorts of policy relevant topics that people might care about. I mean, you know, a large fraction of the topics that show up in newspaper articles or in academic journals that people argue about, there could have been bets about those things. but in fact the vast majority of bets on places like in trade or all these is basically sports bets there are really very few bets on anything but sports and then when they are betting on something that isn't sports it may be elections but it's really much in the sports vein of betting on a contest or a race and the large space of abstract intellectual questions that people could bet on there's

just very little demand for them that over the years they did try a number of those I suggested a few and they just produced very little trading interests so so to be specific I got someone to create and subsidize some markets on in trade on some consequences of who would be of whether the Republican or Democrat would be president. So there were markets in the oil prices

and stock prices and troops abroad and things like that. And those are arguably, you know, very basic policy relevant things to argue about in terms of who to elect. So there were big markets, of course, in whether the Democrat or Republican would be elected but these markets in, if the Democrat was elected, how many troops would be abroad or what would

stock prices be, or royal prices be, there was just very little interest in that. And those were markets that could basically settle quick soon after the election. So when people go betting, they want to bet on things that are things that were resolved quickly and that there are things that other people are arguing about. So if you walk into a sports bar and you argue about who's going to win a, you know, which team is going to win this Sunday, you can get a lot of people argue about with you. If you start talking about health insurance mandates and whether those were reduced costs, you'll find it hard to get anybody in the sports bar to take a bet or argue with you about it. I mean, it's just not what they're arguing about, right? So I think it's somewhat

naive to expect that there's just this huge groundswell of demand to want to bet on important policy questions. There's a mild demand for it, and hopefully if we can make the cost low enough, that demand will be realized, but I have more hope about organizations who want to know things paying themselves to get the answers to the questions they want to know, as opposed to they're just being a lot of people out there. So I'm more want to convince firms that they do want the market in their project deadline, and they want to pay for it and they want to structure it and get them to try that out, then the idea that if the firm doesn't pay for it, even resist it and tries to repress it, that the employees from the ground will all demand and create their own offshore

off-site betting market that they all insist on, you know, playing every day because they are so into telling everybody about whether the project will make the deadline. I mean, it reminds me a lot when you talk about prediction markets. And I think that's also why I wanted to spend some time on that. And explore this is, of course, the Bitcoin case. Because I remember when I learned about Bitcoin, I mean, I didn't have any role, of course, in Bitcoin like you had in prediction markets. but when I learned about it, it was immediately,

wow, this is amazing, incredible. And it seems like people are just discovering about that because the price was growing up. It's like it's pretty obvious to me that this is what the world is going to use and it's going to happen very fast. And so far, it hasn't really happened. And most people don't care. And even if people care in the abstract that they like the idea,

they don't actually care to use it. And so it seems interesting that there's, this seems to be a big similarity there. So is this a little bit how you look at Bitcoin in cryptocurrencies? It's a bit like the thing you try to do with prediction markets is like idealistic radical thing that actually doesn't satisfy real demand? Well, I would just say more generally that there's just a large space of, you know, engineering and simple model inspired innovation, which then has to go through a big second stage

of working out what kind of products people actually want and the variations that will actually work. So it isn't specific per se to prediction markets or Bitcoin. There's just the larger world of innovating. Most innovators start with some simple concept of what products are for and what we do with various things and what kind of things would be useful how. And, you know, unless they're staying pretty close to an existing product, if they're moving away from existing products and generating a whole new product category, then they are generating that based on some abstract idea of what people might want. And there's just usually this disconnect where people's abstract idea of what people might want is usually rather far from what they end up actually wanting. And so many product ideas are still born in the sense that you try it out and nobody wants it and you just quit. And other ideas have more promise.

and then you need to adapt them to find the variation that people will want. And part of this barrier is not just ignorance. It's not just that, you know, most people who have ideas for products, just, you know, it's a big, complicated world. They don't know all the details of it. Often, the barriers is we have ideals. We have these theories in our head about what things should be for and what people should want.

And if we are overly confident in that, then we will push too far for what people should. want. So I don't know the Bitcoin world as well as I do prediction markets, but yes, obviously, many of the concepts for Bitcoin and blockchain-based products are based on the kind of product people should want and the kinds of features of it they should want as opposed to what they do want. So I know there's reasons, there are differences, but I'm always struck by the, you know, product from 20 or 30 years ago of like PGP and private email and private communications, because that was based on the concept that people should want private email.

and they should be willing to pay a little of extra trouble to download some different software and go through a few extra clicks to support their private email. And of course, there was perfectly reasonable to believe that many people cared somewhat about privacy. I mean, people all the time talked about how much they cared about privacy and how upset they were that privacy was broken. And in the last few years, people claimed to be upset that Snowden discovered the NSA was reading everybody's emails, etc. And phone calls. They say they're upset and they say they want to do something about that. And they'd rather find, you know, they talk all that. But then, of course, they talk a little bit more than they do. So, you know, the things like PGP have still been pretty marginal over the last 20 years.

And so, I mean, that's an obvious example of people saying, basically people care a bit about privacy, but a lot less than they, the noise they make when it seems like somebody else has taken their privacy without their permission. This kind of reminds me about there was this classic book that was written in the 90s called Mega Mistakes by a guy called Steven Schnars. And he had this idea that most new technologies or concepts are going to fail. And only those technologies that improve on some aspect of somebody's lives by a factor of 10x, like 10 times, are going to succeed. like it should make something that the user cares about 10 times better than the current alternative out there and and maybe like the problem Bitcoin is having is you're not finding that 10x improvement in the life of a Bitcoin user yet well I think with Bitcoin you have you have so strong network effects

because as I think in Bitcoin is you always see like you see this world where everybody uses And then it's very easy to see these 10x improvements everywhere, right? But to get there, it's not clear how you get there. And before you're there, it's not 10x improvement. It's actually worse, in most case, right? In most case, now you have to deal with these, you know, private keys and whatnot. It's actually interesting. Sebancia sent me this radio lab podcast episode, which was hilarious.

It was about this woman, some older Ukrainian woman in Boston that got this virus. It was like a crypto locker, sort of. So basically it encrypted all her files and it asked her to pay a ransom in Bitcoin. And so the woman talks about what an ordeal it was to get these horrible Bitcoins. And now she had to send her passport. And then she got finally like $500 worth of Bitcoin. and then she wanted to pay, but then the price had dropped. And she could not make sense how like this thing was not worth 500 anymore.

So she didn't have enough. And she only had 490. And it was quite hilarious. And it almost sounded like the worst part was not that her files were encrypted, but that she had to go through this ordeal of trying to figure out how to use this Bitcoin thing. So I think that's a problem. problem we have, right? And the other problem is, of course, you see it's like if everybody used it, well, 10x improvements everywhere. But before that, they're not here, right? So I mean,

I don't know if prediction markets are similar in that there's such strong network effects. Maybe maybe not. I mean, there's definitely effects of the form when only a small number of people are using them. It's a lot more expensive to do so. There's a lot less social support, infrastructure, you know, tools, etc. So it's not quite the same as a network effect, but there's definitely things get cheaper when a lot of people use a product and service, even when it's used individually. So I definitely think, so I like to make the comparison with cost accounting in firms. If nobody did cost accounting in firms and you propose that this project and this particular firm have cost accounting applied to it, you'd be basically saying, somebody's stealing here and we need to check and find out who that is. And it might be pretty unwelcome advice or news.

say, you know, you're basically accusing somebody around here stealing. In a world where everybody does cost accounting, and you say, let's just not do cost accounting on this project. How about that, guys? You'd be basically saying, I'd like to steal a little. Is that all right with you? Can we cover that up? And that also wouldn't go over so well.

So, you know, it matters a lot what the usual thing is, whether you can do any one thing different in a way in place. So in a world where nobody does prediction markets, if you say, we should do prediction markets here. You're basically saying, you know, our usual way of talking about this stuff and deciding what is, it's bullshit. We keep lying to ourselves. So we need something else. Not such welcome news. And we're everybody basically always did a prediction market on a project to see if it was going to make the deadline. If you said, I think we should do this project and let's get going, but let's not do a prediction on a market on this one, please. Let's just skip that step. You'd be basically saying,

we're not going to make this deadline, but let's not talk about that. Okay, guys. And that would look bad too. So, you know, yeah, the more we can create a standard of people using it, that's also, of course, true of, you know, encrypted email. You know, if almost nobody's using encrypted email and I tell you, hey, how about you and I use encrypted email? Why don't you download the software? All of a sudden, you're going, you know, what, what's he trying to do here? What kind of secrets is he want me involved in? Do I want to be involved in that? And, you know, I'll set up sorts of red flags. If everybody's email was automatically encrypted, then there's nothing strange about talking to somebody that way. It's just the usual thing,

and it doesn't raise any questions. So a similar issue, of course, with blockchains. you know, the question is, what are they going to be used for? And is that going to be a typical enough thing that it's okay to just use it for usual things so that it doesn't make you look weird or suggest you're doing something different? So at the moment, people can say, oh, I'm just into it because it's new and cool. And that's a reasonable excuse. And so it doesn't look too bad to be using it. If you're saying, I just like, I'm a techie guy and I like things that are new and cool. But after a while, that's going to go away. And, you know, and you're trying to avoid the situation where you say, I'm using Bitcoin and somebody else thinks, ah, you're doing drugs,

you know, because hey, that's what the only thing everybody ever uses this thing for, right? So you definitely have to create this wider set of common applications that are accepted that don't flag you as being something illicit merely by using it. Today's magic word is Huttarki. F-U-T-A-R-C-H-Y. Head over to letstockbidcom to sign in, enter the magic word, and claim you're part of the listener award. Let's take the time out to imagine a world in which people are using prediction markets

and they have broken through into mainstream society, as you alluded to. And you have this proposal of how government would work in such a society, and you call it Futaki. So we'd like to know what a society that has adopted Futaki would look like. How would it make decisions? How would it make laws? Okay. So let's first talk about governance as opposed to government, because governance happens in all sorts of scales. And so Futarky is a proposal for governance, and eventually it could be applied to the largest scale governance issues like national governments and even international. But since it's a governance mechanism, it can and should be applied first at much smaller scales. So prediction markets, again, are markets. that produce information.

And of course, the value of information in standard decision theory is that it helps you make decisions. There's not much point in just collecting information because there it is. You want to collect information that sits next to a decision you're going to make. The closer you can get to a decision,

the more value of the information. So if you have a project that might go over a deadline, the reason why you might want to know if it's going to go over a deadline is because there's something you would do about it. If it's not going to make the deadline, then you can delay the marketing of the product or you can change.

who's in charge of the product, you can kill the product. There's things you can do. If there's nothing you can do about it, there's no point in bothering to find out if you make the deadline. You just will or won't and live with it. So fundamentally, the reason why we want information is because we make decisions. But most of the prediction markets that are typically created, there's a substantial distance between the information we're asking about and the decision you're going to make. And usually people have often just thought, what might I want to know? And they make some markets on that and they don't think very much about what I could use it for and they often realize later on that they didn't really want to know. It wasn't something that was actionable. They couldn't do much

with it. So they quit. They just give up. And that's one of the reasons prediction markets are often don't last that long at organizations is even if they don't bother somebody, they turn out not to be useful because they didn't actually ask the questions people cared about. So decision markets are a way to make prediction markets very close to a decision. So, Let me give you a concrete example of a decision market because it has a lot of ways to explain in applications, which is one of the biggest decisions most firms make is whether to fire the CEO. So CEOs make a big addition to firms. When the CEO isn't doing well or not a good match, it's really important to fire the CEO.

And so it's important to decide whether you should fire the CEO. Now, it's well known that most boards of directors are shy about firing CEOs because often the CEO put them there or if they try to fight the CEO, he'll fight back and he's got a lot of weapons. And so they usually just let him go. And so CEOs stay too long. So you could have a prediction market related to how the firm is doing, you know, and predicting the future outcomes of the firms. But that's not very directly relevant to should you fire the CEO. So a decision market on should you fire the CEO could be constructed and it would go the following. You could have, say, a stock market, you could have a public company, for example, and it could have a stock price,

and that stock price goes up or down as people expect the future of the company to do well or poorly. And you could have a market in the stock price that's conditional on whether the CEO leaves or not. That is, an ordinary stock price, you just trade cash for the stock, and you're willing to pay more cash when you think the stock is worth more. In a conditional market, you make conditional trades, that is trades that are called off if the condition is not met. So you could trade cash for stock for the company conditional on the CEO staying this quarter or conditional on the CEO leaving this quarter. So the idea is to create both of those markets. And that would have two different prices. One price represents what people think the company is going to be worth on average if the CEO

stays. And the other represents what they think the company is worth on average if the CEO leaves. And so now the difference in those prices is the consensus speculator estimate of whether or not the CEO is good for the company. Now, to settle these markets, later, we only need to know if the CEO came or left and whether the stock price later is high or low. We don't actually need to know if the CEO was directly, causally helpful to the company or not. We don't ever have to judge that. All we have to judge is, did the CEO leave and what was the later stock price? But just knowing those two things later, let's us settle these bets. But Xani, when we're making the bets, we are incentivized to reveal what we believe about the effect

of the CEO and the company. We should set these conditional probabilities, these conditions, estimates to our best estimate of them. So this is a decision market in whether to dump the CEO. So you can, the basic structure here, there's two parts. One is there's a discrete decision to make. You can say we're going to do this or this or that and you can lay them out ahead of time. We're going to do one of these things. And the other part is the outcome we care about. In this case, it's the stock price of the company later. So if we've got discrete decisions, clear outcomes, we can set up a decision market. So Futarki is an application of decision markets. So we can, of course, create decision markets on many decisions like whether to fire the CEO, and we could create them with many different outcomes and many different decisions, and they could be typically advisory.

So my proposal, if I had a million dollars, which I posted on, is I would go offshore and I'd create these markets on firing the CEO for the entire Fortune 500. So 1,000 markets, too, for each company. And I still thought that would be a great grill of marketing because all the CEOs would pay attention in the market. You'd get lots of press attention. and after a couple of years you have enough statistics to decide whether, in fact, the companies that are following the advice are doing better, after which point you could probably be pressure boards and directors in the following the advice for fear of being sued for not taking well, you know, being good trustees of their company. Yeah, that's equally. Let me ask one question about the mechanism because I'm not sure I fully understand that.

because so there's a market now for what's the share price if we keep the CEO and there's a market for what's the share price if we don't keep the CEO. And then the higher share price is going to determine our decision, right? It's the term of the recommendation for your decision, yes. Right. But so let's assume that's binding or something. So let's assume that, oh, if the share price, people say, or if the CEO stays, the share price is going to be higher. So they pay more for that. But then, I mean, why can't this be manipulated, right?

Because you never know, you never know what didn't happen, right? So the people trading in the lower market that, you know, of those, that event that's never going to occur. How do the economics of this work? So, you know, the typical scenario here, the simplest scenario, let's say. is where the market is just advisory. And so all you need is that people think there's some chance that either condition will happen. If you're really sure that a condition won't happen, then you have very little interest in betting on the condition. So, you know, we can have betting markets on if Santa Claus shows up tomorrow, what will we do?

But since none of us believe Santa Claus will show up, we just have very little interest, no interest really at all, bothering to bet on those markets. So the market's price will be pretty meaningless. there probably wouldn't be much of a price. So we need a condition that people think has a chance of happening. It doesn't need a 90% chance. It might even only have a 10% or 1% chance of happening. But the smaller the chance of a condition happening,

the less interested people are in bothering to bet on a conditional market because their profits are proportional to the probability that the condition will happen. Other losses are as well. So in order to get people to bother to bet in a market, you need some chance that it might happen. Now, of course, in most companies, there is a substantial chance every quarter that the CEO will leave. You know, the average 10 years is maybe 10 years or something.

So that's 1 in 40 for every quarter. So, you know, there's a 2% chance every quarter that CEO will leave. So that could give you enough of a chance to bother to bet on what would happen if the CEO left this quarter. That's, of course, if you think there's that chance. Now, the more sure you get that one or the other decision are going to happen, the harder it is to get people to tell you about what they think will happen in other conditions. So decision markets to work better when there is actual uncertainty about the decision. Certainly among the people who are advising the market, who are betting on the market, they can be sure.

Now, obviously, one mechanically, one way to just ensure this is just to have some small random chance of actually doing the decision randomly. So, you know, if one and a thousand, if every time, you know, if every quarter the board just, you know, rolled some dice and had a, you know, one and a thousand chance of dumping the CEO just based on the dice roll, well, now, that could give you enough of a chance to be sure that he'll go out. But of course, we know in the real world, there's just people die randomly at various times, right? CEOs just sometimes die. And there's enough of a chance of that randomly happening. That'll give you a small chance that the CEO will leave this quarter. order. And so, and there's other random things that can go wrong, you know, in the world. So I don't think you have to worry that the chance is going to fall below one in a thousand. But of course,

the smaller, the chance, the less effort will put people will put into it. And so the noisier the prices would be. It's time for word from our friends at Shapeshift. ShapeShift is the fast and easy way to trade cryptocurrencies to trade all cons and they now support about 50 old cons in cryptocurrencies. So if you want to trade all coins, there's different ways to do it. You can use an old fashion exchange. It's a bit like using MSDOS and getting angry at your computer. Or you can use ShapeShift to make this the most heavenly part of your day. And now go to ShapeShift.combe, and you can get this done in less than a minute. The great thing true is that you don't even need to sign up for an account. In fact, if you try to sign up for an account, you'll get very frustrated

because there is no way to do it. So if you're an Altcoin high roller, you're going to love their a higher limit option that they've just added. It used to be that you could only trade like pretty small amounts of alt coins, like for example, like four Bitcoins worth maximum. But now you can specify higher amounts. So for example, if you want to buy like 500 light coins, you can now do that on SafeShift. So go to ShafD.O and give it a try. And we'd like to thank Shapeshift for those support of Epicenter Bitcoin.

So the interesting thing about this future target discussion, you know, when we talk about it like this, it seems to, this abstract futuristic concept that's like far away. But actually, when we come back to sort of, you know, the usual topic of this show, which is like Bitcoin and cryptocurrencies and decentralized technologies, then I think a concept like Futurarchy is extremely relevant and extremely pressing. And the reason is that one, you do want decentralization, right? So you want to have some process that can be captured by some entity or government

that determines, you know, first of all, with Bitcoin the transactions and stuff, and I guess Bitcoin has that part semi-handled. I mean, there's, of course, problems there too. But then how is it updated, right? So we've had the, with the block size debate, we've talked about this many times, and we've also been started talking about governance. We've tried to talk about governance. Bitcoin doesn't have a proper governance model.

And the result of that is that decision, there's just no decisions are made almost. It's like a stalemate. And people disagree. There's no way of resolving disagreements. And so I think with Bitcoin, you know, future key, maybe one of the few candidates that actually could solve this, right? So is this an area you're watching? And do you think that could actually be where now prediction markets will break through? because people will see what happens with Bitcoin

and they will say maybe Bitcoin itself will do that or maybe people will do that with other projects to say, well, the next time we're building something, we need to build a prediction market in there from the ground up so we never get into the situation where like no decision can be reached. Right. So, I mean, from the point of view of being a cautious, careful, reasonable researcher or thinker, you know, the obvious thing to recommend is that new ideas should be tried on small scales first

and slowly worked away to larger trials until you have enough evidence to try something big. Of course, often what you have is that the small scale trials never happen, and so we've got all these interesting ideas for changes that people don't even try on the small scale. So sometimes what often happens

is that you create these ideas for changes, and then they just sit there until somebody's desperate, until a crisis. Somebody out there just needs something, And it's on the shelf and you've been waiting for a decade or two with your proposal and suddenly, boom, they grab it off the shelf and they give it a try. You know, when you're desperate, you're willing to try more things than you otherwise would. But you have pretty, you know, tight criteria too. So I would rather we started with small scale trials and worked a way up.

But hey, somebody who's desperate is also something I'm willing to go with and help. So, you know, it's a bit of a Hail Mary pass, if you will, of everything. taking something that hasn't been tried on small scales, you don't have a lot of data. But I do think the blockchain world has this fundamental problem. I mean, the whole initial conception was that you had an idea of something, of how a system would work, and then you wrote it in code, and then you sent it out there, and now that it's out there,

nobody can stop it. Ha! Right? So the whole appeal was, once we get everybody using it, nobody can stop it. But of course, the flip side is that nobody can change it, right? And nobody can stop it or nobody can change it is great if you thought it was great. If you need to have it evolve and change in response to conditions because it wasn't quite right in the first place, well, now you're in big trouble because you don't, nobody can stop it. Nobody can change it. Nobody can anything. So, you know, that's the key question this community is stuck at, right, is somebody made something and they looked pretty good and they threw it out there.

And in such a way. And everybody jumped and said, hey, if enough of us get on this, then nobody can stop it. And here we are. Now it's going and nobody can stop it, right? And nobody can change it. And now you realize it's not quite what you wanted. And what do you do? So, and maybe now you realize, okay, well, we need to do all this back from the beginning.

But oops, we needed to have some governance as part of this. We needed like a process that somebody could not stop that included governance so that it could make itself better when there were problems. And that was something that wasn't included in the initial concept here. It was a system that it did not include governance because it was under the basic engineering presumption. that you've got to design and you work it out and you release the design. And if, you know, you ran your software tests and they, I'm sure these things passed their initial software tests. And they have passed a bunch of initial practical tests too.

But eventually, you know, real systems that get big and move into new context, almost all real systems, they do need adaptation to context. And so how are you going to do adaptation to context? That's the real question. So you're searching for a governance mechanism. And now you realize, you know, most of the government. governance mechanisms people have are also things that let people like stop things, right? A city does something one way and then national government doesn't like it. And, you know,

if you have a governance mechanism to oversee the thing you're doing in the city, then the national government will come and say, stop it. We don't like it. And the city is going to have to stop it because there are ways to pressure the governance mechanism. So you want to set up something people can't stop, but you want to govern it. So you need a governance mechanism. People can't pressure so easily, right? And so now you need to decentralize governance mechanism that's hard to pressure. And there aren't so many of those out there on the luncheon, right? Well, what else you got? Yeah. I mean, so like the current alternative that the community has is just discussing on Reddit and those like those, those are, they have a big, big problem because anyone can post

the discussions, they result in something that is barely readable and you have the strange people coming anonymous people commenting anything on on the block size debate so bitcoin does need a governance mechanism and futaqi seems like a crazy idea to most people but like it's one of the only ideas out there so maybe you could take the time out and just hash out what it rough what it would roughly look like like in your in your futaki proposal you have these values and these beliefs and we'd like just like to understand what values for Bitcoin would be, what beliefs would be, and how they would interact? So we were talking about decision markets, and we didn't quite get to how it would work

as a governance mechanism as opposed to advising governance. So we were just talking about advisory decision markets so far. And then the final step is to make them formally in charge. So to make a decision market kind of mechanism formally be in charge as a governance mechanism, you need to have a little bit more than what I described before. you need to pick an outcome that's the official outcome and or a way in which that official outcome can be changed. So in my paper on the subject in the Journal of Political Philosophy, the slogan is vote on values, but bet on beliefs. So in that scenario, I imagine that there's a voting process

by which it's analogous to something we're doing today. People, there are legislatures and they pass bills and they vote on what the outcome measure could be. And so in that scenario, you could revise and change the outcome measure. In the Bitcoin world, it's less clear how you're going to revise the outcome measure. And so you might need to pick some outcome measure and stick with it, which is also a lack of flexibility that could come in to bite you later. But nevertheless, what we need is a official outcome measure, either fixed forever or the result of some process that can modify it,

that would have to be specified. And then in addition to that, you need a, way to make policy proposals, an official channel by which an idea becomes an official proposal, such that it must officially be considered. And then once there is an official outcome measure and an official way to make proposals, then you invoke a betting market, basically, and you say there has to be some official betting markets, i.e. a place you go to look for the price. And the rule of government is that you look at the official betting markets for the price on whether the proposal would improve the outcome.

And you have to look at that at some time or some average over time. And it has to be, of course, using some whatever the asset is in this market that you're trading that's connected to this outcome. And then at the right point in time, you look at the price and you see if the price is above the threshold, then the rule of government says we adopt a proposal. And boom, there it is. It's now the default. And it could be overcome again by a new proposal.

So those are the extra things you need to turn this. decision market into a governance mechanism here. You need an outcome measure that's official and a way to change it if you're going to allow that. You need a way to make policy proposals that are official, and we can talk about how you limit that so there's not too many of them. And then you need the markets that are official, i.e. where we're going to go to look at the price, and we have to decide when we look at the price. And then the rule is some of the automatic mechanism looks at the price and makes the decision. And if the price is high enough relative to the status quo, you just do it. So that's like a lot of concepts.

Maybe we could just map these concepts to what they would be in Bitcoin for it to make better sense. For example, when you say outcome measure, I tend to think of the price of a Bitcoin in dollar, because that is a measure that's very obvious on the system. Right. So, I mean, more robustly, you might think that if you're going to allow quantity changes in Bitcoin's or something, then it would be the price times the quantity that you're going to want to track. But of course, and scenarios with a number of Bitcoins don't change, the prices is it reasonable proxy. But then you might also want to tie it to some stable measure of value besides the dollar.

It's always possible the dollar will undergo some terrible inflation. And so you might want to pick some more stable value, like some average of world currencies or an average of world stock prices or something. Even though they would be more trouble to monitor, they would be more robust to possible fluctuations in the dollar. You know, those, that would be details that you might want to pay attention to. But yes, it does seem like, you know, just the total quantity value of the coins might be a reasonable, robust value of measure here. You know, you could also, of course, do transaction volume, but you might worry a bit more about that. Honestly, I think, you know, having any decent governance mechanism tied to any decent outcome measure is just such a big gain over what you're likely to actually end up with. I wouldn't like haggle too much about the exact outcome measures.

I'd just worry about whether you can get anything going. So, I mean, like another outcome measure could be perhaps some measure of decentralization. Like how much does it take for me to, for people around the world to run full nodes, for example. So, I mean, if it is cheap for people to run full nodes, then Bitcoin can be a more decentralized system. So like, perhaps you could have outcome measure. like which are price, then volume. The problem with volume, of course, is that some determined attacker could just spam the network

and create volume artificially. You want to be sensitive to whether these measures you choose can be manipulated, but you also want to be sensitive to whether they're like constraints and conditions as opposed to the thing you really care about because it will really try to maximize whatever you give to it. So, you know, if you just made it this cost of doing these full nodes and that was your only outcome measure, then it might be able to just achieve that. And then it wouldn't care about anything else.

And then it might not do things that would increase like the total size and use of the community once that it achieved this thing of just making the cost low. So I would think you'd want to pick an outcome measure that continues to be hard, however far you go along that line, so that you continue to keep moving. So I would think you might want to like, you know, you could do a product of the Bitcoin price times some measure of this cost that you expect, you know, as long as the cost is below some threshold, you're okay with it or something so that it saturates. It seems like this so clearly also illustrates how hard of a problem is and how that probably, to be quite honest, isn't going to be the solution, right? because even if you look at something like where one maybe could agree on like the Bitcoin, you know, the Bitcoin, the total volume as a total value of, you know, quantity times price, like of course, if you looked at the Bitcoin community today, you know, there will be multiple ways of maximizing that quantity,

but of course one would just be by increasing the volume, which of course would dilute existing holders and huge resistance there, you know. So people would definitely not agree on that. And then the problem is, I think as you said... But would that, in fact, increase the total value of all the bitcoins if they increase the volume? It's not obvious. It would. It's not obvious, right? But let's say increasing the volume by a factor of 10 with half the price.

I mean, who would those new bitcoins go to? I mean, if they went to existing Bitcoin holders, of course it makes a difference. factor of 10 and the price went down by more than a factor of 10, then the optimization would be failing, right? So the question is... Right, but let's say the price would go down by a factor 2. Well, in that case, you would have increased the total value by 5, and this criteria would approve of it, and maybe it would be a good thing. Maybe it would be a good thing, but if you ask the current Bitcoin holders, they would probably say, this is not a good thing. I liked my Bitcoin more when I still own more of it.

I mean, let's get back to the desperate point, right? I mean, the whole reason when you're even considering this is like you're stuck in a corner where you don't have that many great options. Otherwise, you know, you would wait and see, like, let this technology develop on small scale and work its way up. But the only reason you're even considering it now when it hasn't been worked out and tested is you're kind of desperate. You need something here and you need it sometime, you know, soon and you need it to be,

you need it to work. So that puts a high priority in just picking something simple that's clear that seems workable and just doing it even if it's not optimally what anybody wants because you need something. Yeah, that's quite right. I mean, of course, opinions differ on that. I was in contact with Peter Todd just recently. He was like, no, it's actually good that there's no governance process because it should be hard to update it. And, you know, if it can be updated, you know, that's that is as. desired and of course opinions completely differ. So in the case of Bitcoin, it's complicated.

Right. I'm sure there must also be the issue of whether pressure could be applied to a governance mechanism. So once you have a governance mechanism that says it allows a change, then the government could say, okay, everybody, we need this change because we need you guys to be under our thumb. And now if you don't adopt the following change, there's going to be this health. to pay and now, you know, it's pressuring the governance mechanism to adopt the proposal. So once you can change something, you are now open to pressuring to being pressured to make changes that somebody else wants if they have a way to pressure you, to make a threat, basically. They can make a threat, a credible threat of somehow hurting the entire system. If you don't do what they want,

then they can pressure a governance mechanism to do what they want. So you have to wonder, are you able to stand up to whatever pressure they will try to impose in order to make the decisions they want. And then when it comes to, you know, let's say mining pools being in charge of that voting power or charge of that mechanism, then the answer would probably be no, you're not able to stand up because actually people can be located and they can send people there with these data centers, arrest people and all. So the ability to resist pressure is actually not that high. Well, you have to expect a governance mechanism will allow you to adapt, but will also adapt to threats. and you will adapt to threats and therefore you will succumb to threats.

It's walking down the game tree, as we call it, as the economists, figure out what the consequence of these actions would be. I think in any case, people will try things like that. And, you know, I know Vitalik or Vithereum is written a long blog post, and they didn't, even if this is a very daring and out there idea. And I think I am sure many people will end up setting up some sort of future systems in the cryptocurrency space and they will make mistakes like that, you know, they'll choose the wrong parameter and it's gone and I'm sure there will be lots of mistakes,

but I certainly look forward to seeing some of that experimentation. So the right, I mean, a compromise would be to identify the parameters that you don't want outside threats on and don't allow governance on those. So, I mean, that's analogous to the constitution of a country or something. You just have some parameters that you make very hard to change. which is why a constitutional conventions and amendments are very hard. You make other parameters

that are easy to change. So if there are parameters that you don't, not very worried about threats changing, they're just parameters of administration and management such that, you know, if you don't get them right,

things just go badly, but nobody's really going to want to threaten you on. You could make those parameters easier to change and then hold off on some other parameters as just out of the range of change unless there's this

really, really strong consensus. So you could set a consensus parameter, like what percentage of all the miners have to agree on something before that thing can be changed. You probably would want to just set some of these things way off to the side so that you need 99.9% of agreement before those parameters can change. So for example, like you could have something, you could have futarchy as a settlement mechanism for disputes that are beyond a certain size? What I'm trying to say is, let's say there's one proposal which is just to add a particular new op code to the scripting language. There's some disagreement, but you could think of the disagreement as like 80-20. Most miners you would ask are in favor of this new op code. And this has happened in the past.

Those things don't go to the futarchy, so that's settled. But when you have a split where the miners are split 50-50 or the core developers are split 64, it's those situations that are really hard and you could fall back on futarchy as an option because both the options obviously are pretty good otherwise a lot of the community won't be supporting each one of them and then you could use futarchy in those in only those circumstances once you have a set of governance mechanisms possible you have a big commentatorial space in which ways you can combine them i mean that's just generally true of all design uh parameters and features I would think that the issue is just more,

do you have any plausible governance mechanism here that meets some minimal constraints? It's not whether you find the optimal. I would worry less about finding the optimal one and worry more about do you have anything that just works at a basic level. So if you've got some voting mechanism and it works for lots of things,

well, now you should be a lot less worry because, hey, you've got something that works. If you have a futarkey mechanism that works and then great, you've got something that works. but I think at the moment your problem is do you have anything that works? Not can you combine it in some optimal way? I mean, in the case of Bitcoin, right, the sort of way it works is that there's code that's run, right? Code that determines what's a valid block.

And it's sort of what the majority runs of that code, that is the authoritative sort of truth and rules of the network. Now, in a way, there's a bunch of people in charge of that code and, you know, they can change it. And then it sort of depends whether people upgrade or it goes somewhere else. But so there is some sort of very weird type of governance mechanism. It's not maybe, you know, is it certainly not perfect? And I think one big problem is that these people certainly accomplished developers, but they weren't elected. They don't really represent necessarily Bitcoin users.

And there's so much inertia there that to try to get people to change something else is extremely difficult. So they ended up having a ton of power from being in that position. And I think that happened sort of a little bit by actually. accident, but now there's like no way out. Well, I mean, you do want some cost of change in governance. I mean, that's why we often make it hard to change rules in government. So it's not a terrible thing if there is a substantial cost to make changes because there are a lot of things that go wrong when change is easy.

So this is just a general thing we know about governance. In different contexts, you try to adjust the cost of change to trade off. the need to adapt from the cost that can happen when the things that go wrong when cost is too easy. So when change is too easy, I'm sorry, then there can be off of lobbying. There can be a lot of what we call rent seeking, people pushing for things to get personal advantages, a lot of attention people have to pay to things that other people are trying to push for so they can protect their interests. There's just a lot of costs in real systems when change is relatively easy. So a standard story we have like in organizational theory is that when you have like a travel office,

they usually have a bunch of rules about travel that are just hard to change. That is, you know, what kind of travel you can get reimbursed for, whether it's a plane ticket or a cab or whatever. They've just got a bunch of standard rules and no, they don't want to change it. And if you've got some exception, they don't want to hear about it. And you think, well, that's not very adaptive. What if my circumstance is different? What if something changed here? Shouldn't they like pay attention to the details?

well. If they start to pay attention to the details, then they'll get lobbied a lot. They'll have all these people who come up all the time saying you should make this exception and that exception and the other exception and they'll have to react to that and people have to do all this lobbying and they just make a decision. No, we're just going to have a simple world. We're not going to change it and that'll minimize all these costs. So by analogy here, there should be some cost of change. It should be not easy to make change exactly because you'll end up with all these costs if you do. But you don't want to make it so hard that you can't ever change if you need to change. So that's really the question is, do you have a way to make substantial change even at a substantial cost when substantial change is really needed?

Yeah, absolutely. Well, Robin, thanks so much for coming on at the end of our show, but really enjoyed this discussion. And I think and I hope that this sort of intersection of decentralized technology, cryptocurrencies, blockchain and prediction markets is going to be. become quite interesting. I mean, I know there are some projects in the space, Orguri mentioned, and there are some other ones too who are building prediction markets in this way. And we'll see, maybe we'll be like a lot of other attempts in that they don't go too far. But maybe actually there's something there. Right. So in any case, I am very, very excited to see how they turn out. So we academics are usually like big fans of innovation somewhere else, right?

In our own little world, we don't innovate very much and we're not very interested, but we like to take credit for innovation somewhere else. So in the same way, I'm sure people stuck in the middle of this, they're really scared about any of these options because they could all go wrong. But I'm on the outside, so I just, yeah, it would be great if you guys tried something because it should be fun to watch. Yeah, so thanks so much for coming on. Of course, we'll link to your blog.

also your Twitter and your book is it out already? Not until March not to March but maybe six months to go perhaps there's some pre-order site in case which is actually

it's on singularity more so not exactly on topic but if you've liked what Robin talked about then you think he looks like someone who would write a good book then you know you can perhaps pre-order that book all right so thanks so much for coming on and you know

thanks Mayher for your second episode It's been a pleasure. Yeah, it was great. Thanks for coming on, Robin. It was really good to talk to you. Take care. Yeah, and your listeners, thanks so much for joining us. You know, we put out episodes every Monday.

You can get the episodes in your favorite podcast player or see them on YouTube. And it's on YouTube.com, such episode of BTC, by the way. Apologies for the terrible lighting. It's going to be better soon. And if you're a loyal listener, then we still have this bribery competition. competition going on which is that you leave an iTunes review for us and you can say wonderful things or horrible things that's irrelevant and we will send you a t-shirt at least as long as we have still have some and if the rate of reviews

keep up the same way and so a lot many of you've done that and thanks so much for all the wonderful things you've said. And so thanks so much, and we look forward to being back next week.