Augur(ing) Assassinations

[music]

00:07 Paul Matzko: Welcome to Building Tomorrow. A show exploring the ways that tech innovation and entrepreneurship are creating a freer, wealthier, and more peaceful world. As always, I’m your host, Paul Matzko, and with me in the studio today are Will Duffield, the editor of Prototype, and Aaron Powell, Director of Lib​er​tar​i​an​ism​.org. Today we’re talking about the ancient practice of Augury, or the predictive science of animal sacrifice. Aaron brought a turtle dove with him, and after I vivisect it, Will was going to divine the future from the patterns left by its entrails in true Greco-​Roman fashion. Don’t worry, no animals are actually gonna be harmed in the making of this episode. But clearly, the tech start-​up Augur didn’t have historians in mind when it chose its name. Now, Aaron, could you just open this by explaining what a prediction market is, what Augur in particular is doing, and what your experience with it was like, ’cause I think you actually bought a… You placed a bet on Augur, am I right?

01:05 Aaron Ross Powell: So, a prediction market is simply that, it’s a way for people to create a market in bets on various occurrences. So we might have the… The one that I placed a bet on was whether the Congress would confirm Donald Trump’s Supreme Court nominee by a certain date. And so you get to… So I can bet either yes or no, and other people can bet yes or no. And you put a little bit of money on it. And the amount it costs to buy a yes or a no is a factor of what people are willing to pay for that. So you get a… You get like basically a stock market style effect. And with this, you can kind of get at what the community thinks is the likelihood of a certain event occurring. How strongly they feel it will occur based on how much money they’re willing to put into it, and the data we have going back, is that prediction markets can often be fairly accurate predictors of what is actually going to happen… They have a decent track record of making accurate predictions and potentially a more, a better track record than just individuals guessing.

02:22 Will Duffield: Because guessing is costless. We don’t suffer any personal penalty when we guess wrongly. And in fact, most people don’t remember when we’ve guessed wrongly, so there isn’t usually an accompanying reputational penalty. However, in a prediction market, you’re putting your money where your mouth is when it comes to this potential future world state.

02:43 Aaron Ross Powell: Right. And so it’s a way to aggregate knowledge that people might have, and how strongly they feel about this, and how much they’re willing to kind of put their skin in the game on it. So, Augur is simply another prediction market and there have been prediction markets. What sets Augur apart is that most prediction markets until now have been centralized, so that a company runs them the same way that ESPN runs Fantasy Football leagues. The result is that they’re… You can regulate them, the government can regulate them, and so the governments have set all sorts of rules. ‘Cause this looks like gambling. There are certain things that it’s okay according to the state to gamble on other things, it’s not, they wanna limit it, so the people don’t lose their shirts. And so the prediction markets that we’ve had frequently have limits that have the effect of them limiting their ability to really get at what the market seems to know. So, one of the obvious ones is they’ll be caps on how much you can bet.

03:53 Will Duffield: Or whole sectors that are walled off. We can’t bet on politics in the United States for instance, which then limits the ability to use prediction markets to predict who will win elections.

04:03 Aaron Ross Powell: Right. So when the presidential election, there’s a terrific site called elec​tion​bet​tin​godds​.com, that aggregates prediction markets on various political things. And so in the run-​up to the election, you could watch like what the prediction markets said was the likelihood of Trump or Hillary winning, or congressional races, but all of that was based on markets, I believe were in the UK.

04:23 Will Duffield: Yes, there…

04:24 Aaron Ross Powell: Because it wasn’t illegal there.

04:26 Will Duffield: You can walk into any betting parlor, and just as you would bet on sports you can bet on politics, usually get kind of a weird look. Is the one nerd in there betting on American elections? But…

04:37 Aaron Ross Powell: So, Augur attempts to use the decentralization enabled by the blockchain to get around that problem, because they can create a decentralized betting market that can’t be censored, that can’t be regulated, because there isn’t any central service that the state can go to and punish, and that everyone who is participating in the market can do so anonymously, or synonymously. So, Augur is what’s called a DApp. And a DApp is basically, if you think about a standard centralized application, as you have… When you go to Facebook, there’s a front-​end that you’re seeing, and that front-​end which a whole bunch of code that makes the Facebook news feed appear in your browser and processes when you click on things, and when you type in posts, that front-​end is behind the scenes, interacting with a back-​end. It’s interacting with a database that is also hosted by Facebook, and that database is then what contains all of the information that’s in Facebook. The content of your posts and what you’ve liked and all of the creepy stuff that Facebook knows about us. A DApp says, “Okay we’re gonna still have a front-​end, and the front-​end will be hosted on someone’s server, but that back-​end, that database is, lives in the blockchain and other decentralized solutions.

06:03 Aaron Ross Powell: So it might be… It’s information written to the blockchain, plus like all of the heavy lifting information, the stuff that actually would take up a lot of space, and would clog up the amount… The limited bandwidth of the blockchain is stored in decentralized file stores like IPFS, or Swarm, or something of that sort that’s spread out all over. What really sets these apart, is that because you’ve separated, you’ve decentralized the back-​end, you’ve decentralized the data store, anyone can write their own front-​end to that data, just like anyone can write their own wallet to interact with the Bitcoin blockchain.

06:41 Aaron Ross Powell: And so, Augur is… Right now, Augur is a front-​end that’s run on augur​.com, or whatever the domain name is. It is an interface to the Augur blockchain, which is part of the Ethereum blockchain, as a set of smart contracts as… And other data’s stored there, so when you make a bet on Augur, you are basically creating a… You’re contributing Ethereum tokens into a smart contract on the blockchain that says, “If such and such event happens, I will… My Ethereum tokens, I’ll either get back a certain amount, I’ll get back some from the other people who bet the wrong way, or mine will go to other people.” But that’s all written into like a publicly available contract, spread out across all the Ethereum nodes.

07:32 Will Duffield: And from a liability perspective, I assume that third parties creating front-​ends for this database, wouldn’t be held liable for, perhaps what their users, how their users use that front-​end to interact with the underlying decentralized database?

07:50 Aaron Ross Powell: That’s a good question, and I wonder if that’s one that will come up as we get into the specific issues that we have to talk about today that are relatively novel, because that’s not a question that would necessarily have come up with the standard centralized where both the front and the back end were the same place.

08:09 Will Duffield: Yeah, you’re pretty clearly liable in that sort of situation because it’s your cohesive whole entity.

08:13 Paul Matzko: Well, and you have the uncertainty introduced by the weakening of Section 230 of the Communications Act, where now… Traditionally since the mid ’90s, hosts were not held liable for content posted on there by a third-​party user. That’s now been weakened, right?

08:36 Will Duffield: Well, only with regard to promotion of prostitution. So I don’t see a whole lot of overlap here though perhaps someone that could come up with a novel prediction.

08:44 Paul Matzko: But the legal precedent of… So someone could say, “Well because we, as a law system, we’ve decided that it’s okay to hold liable a host, a front end or back end host for content posted by a third party user that has to do with the solicitation of prostitutes and sex trafficking, therefore we can also prohibit these other kinds of posts and behaviors and… ”

09:10 Will Duffield: Well, if they were to pass another law, but not as a matter of legal precedent, no.

09:18 Paul Matzko: Well, well, we’ll see. So the question then is whether Augur, with its front-​end control over access to the back-​end, whether or not… It’s trying to create a cut out, so that individual users can bet on all kinds of things that, that regulators or that law enforcement would find unsavory. This is an introduction to some of the… I think some of the early reporting after Augur was launched, there was a lot of attention around the fact that people had placed bets on whether or not unpopular politicians or Supreme Court members would be dead by a certain point, and this was thought of as an assassination market. This is the kind of behavior that make law enforcement uneasy. But Augur could ostensibly say, “Look, that’s… We’re not in control of that. The back-​end data we… On the front-​end, we can choose whether or not… ” And there was some discussion about whether or not they were allowing those bets on, basically assassination bets. How easier they were making it for users to access the bet itself, but they couldn’t actually shut down the bet. It still existed…

10:35 Aaron Ross Powell: They could… So my understanding is they could at the very beginning. So we had first launched out of beta, a couple of months ago that they had… They owned the original key, the original wallet address that had control over the whole of the Smart contract. And so they could go in and they could modify things or they could shut stuff down or they could… They could act as censors, but a couple of weeks, I believe, after they launched, and this was something they had planned. So you maintain control of that until you make sure that everything is up and running incorrectly, because it enables you to make changes and to fix things. But once they… A couple of weeks later, once they’re comfortable, that it’s actually all working and out in the wild, they transferred ownership of that to like a dummy address that no one had access to. And so now, it just lives on its own and the people who created it have as little control over where it goes, as any of us in this room do.

11:34 Paul Matzko: So at this point now, they all have control over access to…

11:37 Will Duffield: There’s the question of front-​end accessibility, but obviously, even if the creators with regard to their front-​end application, removed certain bets or removed the ability to view or access certain bets, other third-​party front-​end creators could allow their users to access those bets.

12:00 Aaron Ross Powell: Right. So this would be similar to, if Twitter ran this way. So if all of your tweets were stored in smart contracts on the blockchain, instead of in Twitter servers, and Twitter were a true DApp, twit​ter​.com might have all sorts of restrictions, so it might prevent you from tweeting things with certain key words. It might have restrictions on what you could see. So if you searched for certain terms, it might institute its own ban or block lists, but all of that data is still there in this smart contract, so someone else could come around totally independent of Twitter and write twit​tert​wo​.com, they could have a completely different set of rules and could access whatever they wanted.

12:42 Paul Matzko: At this point there could be Augur knockoffs that provide different frontend services, but have access to the same backend data, just in case Augur tried to do something like that. So, okay, so in a sense doesn’t matter what happens to Augur as a front end provider at this point in time. I mean, the data’s there, the contracts are self executing, someone could come along and create an alternative that provide the same kind of front end service, which is probably a good thing because as we were prepping for the show, there’s been a lot of buzz since Augur launched about the viability of that particular front end service. One of the things I think, Will, you found was, how relatively few people were using the service? Can you explain a little bit of that for us?

13:32 Will Duffield: Yeah, so in order for prediction markets to function, as you would intend them to, being able to draw upon the wisdom of the crowd as a whole, you need, to some extent, a crowd. If you have a very small number of users betting on some set of futures, you’re unlikely to be tapping into the breadth of knowledge that you would hope to access through this. So you do see a kind of liquidity crunch here, and particularly with regard to Augur, the Augur protocol, in terms of its valuation derived from its set of tokens, seems wildly out of step with both the number of users and the amount of money being bet on futures within Augur.

14:27 Aaron Ross Powell: So these tokens, just to clarify, so when you place a bet on Augur, you are betting with Ethereum tokens, with that cryptocurrency. But Augur has its own token that functions more in a reputation role. So you acquire these by using it by being on the right side of bets, and the tokens are, I believe the main function of the token is one of the problems that you run into with the betting market is how do you decide which side won. And so, you tie it to an Oracle which is the thing that’s going to decide, but that might be contestable and so these Reputation Tokens enabled the community to interact with which side won, whether it was right or wrong. And so, I believe if you can test something and are proven to be right, you gain these tokens, which then give you more authority at a later thing. And so, Augur did a initial coin offering that you can buy these tokens, and these tokens have a trading value, and so that’s what this… The outstanding value of the Augur tokens.

15:36 Will Duffield: And at the moment, those tokens in some are valued at about $300 million, and yet, Augur only has 64 daily active users, according to a CoinDesk article.

15:52 Paul Matzko: And I went on and checked today, it was 45 active users. So I think pretty consistently in that range.

15:58 Aaron Ross Powell: To clarify the way they’re using these, the count of this active users, is they are looking at… So, when you interact with Augur by having an Ethereum wallet that you sign into it with, it has your unique address, and when you place a bet, either buy or sell bets, on the Augur, into the Augur blockchain, those are marked with… Those transactions are marked with your address. And so the way that these numbers of how many unique daily users is, they simply count the number of unique addresses that made a transaction with Augur each day.

16:37 Paul Matzko: So it’s possible there’s many more people who are logging in to check or look at things, but aren’t actually making trades, that’s a good caveat to point out.

16:45 Will Duffield: But it still looks to be a fairly thin market, at the moment.

16:50 Paul Matzko: Given the total dollar [16:51] ____.

16:51 Aaron Ross Powell: Right, if there’s only 45 trades happening a day, and that’s not in a single market, but that’s across every possible… [chuckle] So from my, “Will, the Supreme Court nominee get confirmed,” to who’s gonna win the next presidency, to who’s gonna win the Super Bowl? That means that the vast majority of these things are not getting even a single one each day.

17:14 Will Duffield: And at least compared to other prediction markets I’ve participated in the past, in this case, Good Judgment Project, there I saw most, or many users updating many of their predictions on a weekly, if not daily basis, in light of both new information and the shifting price of these futures.

17:38 Paul Matzko: And when you don’t have numbers, when you don’t have a certain critical mass, it becomes that… Especially with the proof of stake system, like Ether, as opposed to a proof of work system like Bitcoin, and that’s a whole nother conversation, but the proof of stake system, which is what what this Reporters… Is what they call them, people who have some of the Augur tokens who bought a certain amount and who earned them used them and and get rewarded like reporter status, they’re ones who act as a check on like the solving the Oracle problem, they the are ones who kind of police the community in a sense, if you don’t have a certain number of people, the number of reporters who it takes to throw a result is consequently quite low, so it’s easier to monkey around the results. There’s not only wisdom and crowds. There’s safety in numbers, for a system like Augur, so it is a real issue. I’m still not 100 percent convinced, as Aaron, you brought up with the Oracle problem, this has been an issue for prediction markets before, the most famous example is that I could think of was Intrade, which was a predecessor of Augur, shut down… I don’t know, three, four, five years ago, but one of the founders of Intrade actually is one of the execs of Augur. So there was like a filial connection.

19:00 Paul Matzko: But back in 2012, there was the Republican caucus in Iowa and it was an unusual situation, sure, but Mitt Romney was declared the winner by eight votes, eight delegates in the Iowa caucuses, or kinda throw the Iowa caucus to Mitt Romney. And so, Intrade pays out and says, “Look, the people who bet that Romney would win, they get their payout”. But a few days later, in the rules it stipulated that there would be no recount, so if there was a recount, too bad, whatever the initial count was, was good, but a few days later, they found that… Oh, during the basic certification process, so not a recount. We just basically tallied the numbers wrong. And actually, Rick Santorum won. So understandably, people who bet on Rick Santorum said, “Hey, wait a second, like, we should have won this. It didn’t violate the no recount rule. What gives, guys?” and Intrade had to say “Sorry, tough luck. Our Oracle system just didn’t work in this case”, and so that’s an issue, that really is, especially when you have a small user pool like this, when you can’t necessarily rely on your reporter system to fairly adjudicate the system, I think that’s a problem for Augur specifically, and it’s still something that’s always gonna linger.

20:21 Paul Matzko: It’s not necessarily in Super Bowl problem, but for prediction markets in general, figuring out who gets to decide who wins the bet matters, particularly in complicated situations. So, it’s one thing when you say, “Hey the… Who wins the presidency?” That’s an easy thing to… The top, we want the paper of record and the whatever, the CBS and that New York Times says won the day after, this time of day, that’s who wins. But when it’s something a little more nuanced, like, let’s say, I’ve predicted that Donald Trump will will not finish out his term, kind of an implicit assassination bet, or impeachment bet, or you do that for any other individual person, that person fakes their death. Well, right? This is a scenario that really becomes a real issue for assassinations markets. People who… Or someone spreads a rumor on Twitter that gets reported on as someone… These things aren’t reversible once you’ve written it in the contract. I mean, there’s ways of gaming this system, of playing with an Oracle system that is problematic for prediction markets. Will, you look like you had something to…

21:33 Will Duffield: Oh, I mean, we keep coming back to this question of assassination markets, particularly within Augur, given that someone put one up. And the reason that assassination markets have always been a concern, when it comes to any kind of prediction market, is because if you can, it is unlikely that any one but the assassin can correctly predict when someone is going to die. We usually don’t have a whole lot of advanced warning, as to our deaths or the deaths of others. And within the general market for assassinations, because assassinations are illegal, it can often be difficult to coordinate that pay out. However, if you have an open prediction market, and every one can step in to bet on someone’s potential time of death, the assassin is likely to be the only one who gets it right, and therefore, the one who wins the payout such that the prediction payout can function as a payment for services rendered.

22:51 Paul Matzko: But there is utility in that. We should point that out before we talk about assassinations, which is… So let’s say you’re trying to figure out which… The balance of power on the Supreme Court, so you want the most accurate possible prediction of when Supreme Court justices who hold their offices for life or until resignation, are likely to shuffle off this mortal coil, and so that, which party will get to pick a replacement. There is political value in that, and there is the social value in good, the best possible prediction of when they are likely to live. Even if it’s ghoulish to talk about betting on the date of someone’s death, there’s value in this, and we already do it. I mean the political prognosticators always say, Well, Clarence Thomas is this man number of years old, and Ruth Berry Ginsburg is this number of years old. We think they’ll probably live this long”. We’re already doing that, but prediction markets offer the possibility of doing that even more accurately. So that’s, at the outset, there is social utility in this… And it’s not merely an assassination market, even though that’s kind of the hyperbolic way of framing it. At the same time…

24:02 Will Duffield: You’ve just always heard this complaint or concern when it comes to any sort of prediction market back in… Let’s say 2003 maybe six, the Pentagon set up their own internal prediction market in order to better estimate the time and location of terror attacks. Of course, this was then built or presented by some as a ghoulish endeavor such that it was rather quickly shut down, even though using the wisdom of crowds in order to better understand when terror attacks are likely to happen, could help to save thousands, if not more lives. But it’s a kind of constant refrain that is dogged this youthful means of knowledge dissemination.

24:55 Aaron Ross Powell: “Cause it looks like you’re cheapening the whole thing.

24:57 Will Duffield: Yes.

24:58 Aaron Ross Powell: You’re turning matters of life and death, in this case, into going to the craps table…

25:03 Will Duffield: And making money on it.

25:05 Aaron Ross Powell: And making money on it.

25:07 Will Duffield: There’s a profiteering complaint.

25:08 Paul Matzko: Well, and it’s a reminder that a lot of stuff that we recognize, that are now quite respectable forms of prediction markets, were once illegal. So the stock market, literally the market…

25:22 Aaron Ross Powell: Usury.

25:23 Paul Matzko: Well, interest, you taking interest was illegal under Catholic ecclesiological law, but it was also frowned upon as gambling. It was okay to have shares in the company, if at some point we agreed, “Okay we’ll split shares, companies into shares,” the corporation is okay, but the idea that you would then trade shares in the company with strangers who weren’t… Whose only connection to that company was through a piece of paper and… That was considered no better than gambling. And over time we’ve developed as a society to the point where we say, “No, actually that’s a good thing”. It was once unimaginable, but now it’s good. And the same thing has to happen with non-​financial related prediction markets.

26:05 Will Duffield: And like financial markets, when trading the stock of, or ownership of joint stock trading companies first appeared on the scene, it often did look a lot like gambling. There wasn’t a great societal understanding of how this helped these firms to make better decisions in the future, how it helped other market actors to understand the value of these firms, and in time, we’ve come to recognize the importance of stock markets in doing that.

26:50 Aaron Ross Powell: But now we get back to the problem of legitimate assassination markets. Because prior to a fully decentralized system like this, it was harder to get these things off the ground. I mean, if it’s a centralized system, then when Intrade puts up, “Will someone assassinate this world leader by this date,” the government can knock on the door and say, “Hey you gotta knock that off”. Or it can get at the registration information of the people who are betting one way or another, which might end up being a useful suspect list, but when you’ve got a fully decentralized, this is living on the blockchain, no one can, no one can shut it down, no one can censor it. Everyone who’s participating in it, is participating simply as an Ethereum wallet address, that they could have set up on the fly, so there’s no way to trace that to a specific person or it’s awfully hard to do. State agents can watch this thing go, and they can see the likelihood keep going up or someone wins and the person was actually… And there’s absolutely nothing that they can do about it.

28:05 Paul Matzko: Yeah, well, the prospect for this was imagined a lot longer ago than I realized. There was a fellow named Jim Bell, who’s a cryptoanarchist, they called them a cypheranarchist back in the ’90s, was the term and he…

28:19 Aaron Ross Powell: Cypherpunks.

28:21 Paul Matzko: Cypherpunks. Yeah, right. And he wrote an article called Assassination Politics”, and it’s actually quite pressie, when you read it now, back in 1997, and he talked about using digital cash, and at the time people were starting to think about the idea of something that looks like Bitcoin and Ethereum, but it was still all kind of a theoretical stage, but using digital cash, and his scenario was that… Hey, you’re sitting down watching evening news and you hear about the corrupt local county sheriff did something you don’t approve of, I don’t know, they shot an unarmed black man, or they barged into someone’s house and killed their dog during a no-​knock raid or whatever. Some kind of corruption or crime committed by local authorities, and rather than just fuming at the television, ordinary citizens could say, “Hey you know what, I’m gonna pitch in a dollar of this digital cash for a prediction of their end-​of-​life coming in the near future, and you only need a few thousand people. The number you put in the article was 0.01 percent of the population to put a $250,000 bounty on someone’s head, so that prospect, Bell was very excited about, as a cypheranarchist. He thought, here, we’ll make politicians more honest, if they do something that only 0.01 percent of the people are concerned about, they’ll have big bounties on their heads from people who are buying into these kind of assassination markets.

29:58 Will Duffield: That sounds terrible, 0.1 percent of the population is probably insane, or just horrifically wrong about any issue, at any time.

30:04 Paul Matzko: Yes. Yeah, he…

30:04 Will Duffield: [30:06] ____?

30:07 Paul Matzko: Yeah, exactly.

30:08 Will Duffield: Probably 0.1 percent of the population believes in that.

30:08 Paul Matzko: It’s the kind of thing that only made sense pre-​social media, and pre-​Twitter. Now it’s like, you can get 0.01 percent of people to say incredibly crazy things, or… His more fantastical imagining was that you wouldn’t need the military, because you could basically just use assassination contracts to control. So lots of utopian kind of visions around this, but what’s kinda [30:34] ____ about is the idea of people placing bets in particular, acting in a kind of assassination market using digital cash all the way back in the ‘1990s. Bell himself ends up in tax fraud and he ends up in jail, and such. But what’s interesting here is that like, Okay for 20 years on, from Bell writing this series of articles that was very influential in the exact community out of which comes the cryptocurrency community. Cypherpunks are the source. It’s people like… Anyways, I won’t run down the list, but it is from this community that we get our… Things like Augur eventually emerging.

31:14 Will Duffield: I wouldn’t take Jim Bell as emblematic of 90’s cypherpunk culture, either.

31:21 Paul Matzko: Well, we could argue about the specific influence of Jim Bell, but he was definitely a member of the community that was well read and regarded by the kind of anarchist community at the time. Not that everyone would have agreed with him, or that he’s representative in some way. But the thing is, is we’re 20 years on, we have digital cash, we have public encryption, we have all the stuff that Bell predicted would be in common usage to the point where it would essentially get rid of national militaries and control our political system, and by no measure of Augur are we anywhere close. I mean, the prediction market for the assassination bet that was placed on Donald Trump, I went and tracked it down. It’s about 50 shares at 0.015 Ether, which is like $500 at stake. This is a far cry from Bell’s imagining 20 years ago. So how come, even though we have all the tools he predicted, and in fact even more than he predicted, do we not see this happening?

32:24 Aaron Ross Powell: Well, the first one would be nobody’s using Augur. So my guess is that that particular market is not necessarily an outlier, that most of the markets, even a totally legitimate of, who’s gonna win the next Super Bowl, probably have similar total numbers. But the other one is… So this is a real concern, and it’s a concern from the libertopia perspective, because on its face, like we as libertarians, think, what these things are, these cryptobased markets, and what the encryption and the communications and all of that, and the lowering of transaction costs and… It’s opening up markets, it’s allowing markets to flourish where they could have been restricted before. Yay, that’s awesome from a libertarian perspective, but the fact is that people will do bad things in markets, people to glorious things in markets, and probably the overwhelming majority of what happens when you free our markets is glorious, but people will still do bad things.

33:32 Aaron Ross Powell: And the concern specifically with this is so back at the question of Augur itself not having very many users, because we can imagine that it could, if something could happen and it could spike or some other one to come along and they can do very well, and it’s likely that these kinds of decentralized prediction markets will sometime in the future have a lot of users, is that this particular sort of bad behavior, namely markets in assassinations, people respond to incentives. Markets can create incentives that people then respond to. And so even if you don’t have a… Part of it might be that there’s not that many ghoulish people out there, there’s not that many people who want to assassinate people, but if you start betting, all it takes is that to incentivize one person to go out and try to kill someone.

34:24 Will Duffield: I also think for anyone important enough to end up with a crowdfunded hit out on them, the would be assassin probably won’t live long enough to spend their winnings. I could be wrong there, but it just seems as though there’s a kind of importance threshold that you would need to pass, and obviously I’m looking at what is now a pretty limited market, if this sort of thing were ubiquitous in our society, I might think of it differently. But that in passing that you’re there in, “The feds will get you” territory. [chuckle]

35:04 Paul Matzko: Well, there is here, I think we mentioned it briefly earlier, but with the assess the [35:11] ____ of Fight Online Sex Trafficking legislation, it’s a related question which is, advocates of the law were concerned that people were openly advertising for prostitution in such a way that they were essentially advertising sex trafficking, so by having websites like back​page​.com, where prostitutes advertised their services, we were fomenting sex trafficking, so we shut that down. The Congress shut that down, and critics have pointed out, well, if you shut that down, you just… You’re putting… Those services aren’t gonna go away, they’re going on the black market. In fact, there’s early signs, Mike Masnick on Techdirt has highlighted some of this stuff. There’s early signs it’s made it harder to catch sex traffickers because they’re now on the black market underground. The same argument could be made here, which is that it’s not like there is not already a market in assassinations. There is. People hire hitmen to kill spouses, and…

36:05 Will Duffield: Yeah, but competence matters here in a way that it frankly doesn’t when it comes to sex work. You aren’t concerned about sex work being more open because you’ll attract more competent sex workers, when it comes to Assassins, that’s very much a concern.

36:21 Aaron Ross Powell: But all it takes with an assassin is one competent assassin, and I think this is where you get an interesting effect and a potentially troubling effect of these kinds of markets, because you… Take the philosopher Derrick Parfait, have this thought experiment about harmless torturers, where I might do something that causes no… I push a button or in some way, like do something to someone else that causes them absolutely no harm, it’s utterly harmless, right? And they might not even be aware of it. And we basically cut out everything that might make my action morally problematic, but if lots and lots of people do it at the same time, it has the effect of torturing this person. So, you’re maybe pushing a button that puts a minuscule amount of electricity into them or something, but in the aggregate, it causes them great pain. And what’s the moral status of this? And with a market like this, so we’d have standard assassination. Do you have… You don’t have a whole bunch of harmless torturers. What you have is like one person who has to really, really care and really want to see this person dead because they’re going to pony up, $250,000, whatever it costs.

37:33 Aaron Ross Powell: And then you have one person, then you have to find one person who’s willing to do it for that. And you have to be not only really committed, but really aware of what you’re doing, to pony up that amount of money. Whereas with this, it’s participating in a little market like, “Oh do I think this person is gonna die by a certain day? Sure, I’ll pitch in a buck”. It doesn’t… The kind of moral distance is much greater, and so… But it still, all it takes is one person willing to do it for that amount of money, and now it happens, and so we can become kind of participants in greater evils without really realizing what we’re doing. This makes me think of… There was a few years back, there was a tech that also didn’t go anywhere, but present to fairly troubling future called, I think, dark leaks where dark leaks was… You’ve got the WikiLeaks model, of people could send stolen documents to WikiLeaks and WikiLeaks will then put them out into the world, but there’s no economic model there.

38:37 Aaron Ross Powell: What dark leaks did was it would allow… Let’s say I stole a whole bunch of data from a government contractor, I stole a whole bunch of data, and I wanna get paid for it, and it allowed me to mathematically proved that I had that data by basically releasing small pieces of it that people could verify and then set a threshold. So people… You basically kickstarter the release of this, people contribute Bitcoin to something and then when the threshold is met, enough people have contributed, the documents, the rest of the documents that I have get released into the world automatically. And so you end up creating markets in stolen documents that anyone can participate in for a little bit of money.

39:26 Will Duffield: But everyone gets them at the end.

39:28 Aaron Ross Powell: But everyone gets them.

39:28 Will Duffield: It’s not just some buyer.

39:30 Aaron Ross Powell: Sure, but the seller, you now incentivize sellers, so you end up in a world where you can, in this fairly anonymous way, sell stolen secrets to a global audience. And we all like… Sure, I’d like to know what the NSA is up to. I’d like to know what this company’s up to, so I’ll pitch in a buck here or there. Whereas you wouldn’t be willing to spend thousands of dollars of your own money to pay…

39:55 Will Duffield: I don’t know. I think compared to insider trading for instance, the main complaint there is fairness. That yes, it probably drives us towards a more efficient market, but because only some are getting access to this extra information, it’s unfair, and therefore should be prohibited. But with regard to this, I don’t know, from an efficient market perspective, it just seems like a good thing.

40:18 Aaron Ross Powell: But a world in which everyone was incentivized to sell secrets and had an easy way to do it at all times might be a… Might have economic consequences. We can think of, it’s one thing, if it’s Edward Snowden selling NSA documents, but if it’s every corporation constantly has to worry about every employee selling everything that they’ve got, their trade secrets, their upcoming product info, that could have disastrous economic impacts…

40:43 Paul Matzko: Or, the…

40:44 Aaron Ross Powell: But these are the things, is that the problem… One potential problem with these new kinds of markets, especially these new decentralized, like D-​app based markets, is that once they’re out in the world, there’s nothing you can do about them. And so Augur is now out in the world, and not only can’t it be shut down, but it can’t really be modified either anymore, so you can’t say, “Oh we’re gonna try to adjust the incentives”, because it’s… It’s open to everyone. And so I think that’s a real concern. I don’t think it’s a strong enough concern to say, “Well we shouldn’t allow these things”, because any new tech is gonna have problems. But it is something… When we’re thinking through… The hard thing to do when you introduce a change is to imagine all of its effects and it counsels us, I think, at least, to say we can imagine all sorts of positive effects, but we should also think of the ways that people are gonna use this to do bad things.

41:40 Paul Matzko: Right. Well, it’s not like another technology that actually the cypherpunks and Jim Bell mentioned, encryption. Unbreakable, easily accessible, encryption. That was something that was being thought of 20 years ago that is now generally available and something that annoys the state to a great extent, which is why they want build inback doors into systems, etcetera. And there’s that arms race between consumer goods companies and the government, but while we as libertarians typically say, “This is great, encryption provides a privacy, it allows people to gather without fear of surveillance”, we do also acknowledge that there are people who are gonna use that for nefarious ends, but what we ultimately say is there’s net benefit to this technology. That yes, they’re gonna be bad actors who use encryption to plan terrorist attacks, etcetera, but those are gonna be the isolated cases that on the net, society benefits from privacy, from freedom from intrusion that’s provided by this encrypted system. The same thing true here with prediction markets. Did you have something you wanted to add, Will?

42:52 Will Duffield: I guess, getting back to the question of why we don’t see a higher, higher user numbers for both Augur, but this liquidity issue has plagued all sorts of prediction markets for as long as people have tried to set them up. And why is that? I think to some extent, for most of the things you see, most of the futures you see traded, there are other much larger markets with proxies for these futures in them, the stock market. And if you know… You may have a prediction market question as to whether the United States will go to war somewhere, or even if Honda will ship X number of cars by next year, but if you really wanna make… You feel you have the right prediction, and you really wanna make a lot of money on it, you’re, in most cases, better just going to the stock market and either shorting the state industries in the country you think will get invaded, or buying a bunch of Honda stock if you think that they’re going to ship the number of cars, and that those are just more profitable analogues for the specific question in these markets.

44:16 Paul Matzko: That’s interesting…

44:17 Will Duffield: I don’t know what other suggestions we have there, but it might speak to… We’ve pretty good markets and things now.

44:25 Paul Matzko: That’s kind of an argument against the idea of this… It’s not to say that there isn’t… There aren’t interesting things going on with prediction markets in general, there are with Augur specifically, however the company… In this particular company pans out, but that this isn’t as radical as people might assume.

44:45 Will Duffield: Well, it’s nice to get a clearly phrased, like well constructed yes or no answer to some question you’ve posed, but if you dig into how people are behaving in real markets all over the world, you can often find the crowd’s wisdom as to that question already.

45:05 Paul Matzko: Well thanks, guys, I think I can predict that our episode is gonna end in the next couple of seconds, so thank you for coming on and until next week, be well.

[music]

45:14 Paul Matzko: Building tomorrow is produced by Test Terrible. If you enjoy our show please rate, review and subscribe to us on iTunes, or wherever you get your podcasts. To learn about Building Tomorrow, or to discover other great podcasts, visit us on the web at lib​er​tar​i​an​ism​.org.