Prediction Markets and Beyond
01:48:03 Share

Everybody, i'm ci rid all the host of marketplace, your daily download on the economy. Money influenced so much of what we do and how we live. That's why it's essential to understand how this economy works.

A marketplace. We break down everything from inflation and student loans to the future of A I, so that you can understand what IT all means for you. Marketplace is your secret weapon for understanding this economy. Listen, wherever you get your podcasts.

Welcome to web three with a six cy. They show from a six and cy crypto about building the next generation of the inter net. I'm sono toxic. And today's episode is all about prediction markets and beyond. Our special guest is Alice to barack, professor economics that George made in university and chair and economics at the mccabe center, and Scott commoners research garate six san's ecliptic and professor.

At harder business go prediction markets hit the main stage in the recent al election, which recover briefly, especially to these apart the height from the reality there, since people have talked about the promise and promise of these for a very long time. But we also go more deeply ly into the how, why and where these markets work. ChAllenges and opportunities, including implications for designers throughout.

We also briefly cover other information aggregation mechanisms and discuss applications for all these markets, including touching on trends like q turkey, A, I entering the market, D, I, and more. About half way, drew. We discussed where do and don't block chain and cypher or technologies come in.

And as a reminder, none of the following should be taken as business investment, legal or tax advice. Please see a six and c docs leased this lotus. For more important information, be sure to also check out the show notes for this episode.

And we have a really rich set of links, including all the research cited in this conversation. But first, we begin with a quick over review of what prediction markets are. The first choice you will hear is Alexis.

followed by stocks. So I think a prediction mark is very simple idea. The bottom line is that we are interested in forecasting.

Lots of people learn to set in forecasting things and predicted markets, or some of the best methods of forecasting, which we have yet created. They tend to be Better then complicated statistical models. They tend to be Better than polls, or at least as good.

And just want a reason for that. Suppose that a model is Better. okay? Suppose I have the night silver statistical of predicting elections and is Better than the prediction market as I told that we're true.

Well, if that we're true, I could make money. Yeah, I can use names model to go and make bets. And in making bets on the prediction market, I pushed the prediction market closer to the truth. So almost by definition, the prediction markets have to be at least as good and typically, they're Better than other methods of forecasting. And actually.

you a sort of that is an illustration of why we think of these things of information aggregation mechanisms. What are they really doing? They are aggregating information from all of the people in the marking, right? So if many different people are out, they're doing their own private forecasts, like calibrating, their own models.

There's nate silver and like JoNathan gold, melloni bronze will make up all of our our variations. You know they all have their own models. They all have their own estimates, which they trust with some degree of confidence.

They come together, right? They're all buying or selling the prediction asset based on what their model leads them to believe. And so as a result, the acid is sort of like aggregating all of this information. It's a Price discovery just like we think about in financial markets and commodities markets, like everybody's demand together discovers the Price at which the market clears. And here, because what the value of the acid is depends on probability, right? It's like it's a value is sort like the function of the probability of the outcome of the event, the Price aggregate tes people's .

estimates of that probability exactly, I think, is useful that these are markets, and actually all markets do this. And we learned this going back to high x one hundred and forty five article, the use of knowledge in society. This is a nobel prize winning paper, which doesn't have a single equation in IT. So anybody .

can go and read .

this paper a fantastic, you know, prior to hire and people think what Prices, you know, the ordinate and the make demand, equal supply and production consumption. Hi, I done to know you're thinking about the Price system. All wrong.

The Price system is really about aggregating and transmitting information. And he said, look, there's all this information of out there in the world, and it's in heads, right? It's in people's heads, like what people prefer, their preferences, but also, people know things.

They know what the substitutes are, what the compliments of our thing, or how do we increase supply, what the demands and supplies are. It's all in heads and a good economy. You wanna use that information, which is buried in people's heads sand.

How do you get IT out? Because it's disperse. It's dispersed in millions of people's heads.

The information is sometimes, it's fleeting. Information is sometimes task IT. It's hard to communicate to a central planner. So what hix said is that markets do this because markets give people an incentive through they're buying and selling to reveal this kind of information, to pull this dispersed information.

And for millions of people and and people who are buying, they're pushing the Price of people are selling, they are pushing the Price down. Suppliers, consumers are all in the same market. And so all of this dispersal information comes to be embedded in the Prices and kind of remarkably, the Price. It's sort of no more than any person in the market pause .

on that for a quick second because you guys might take IT for granny. But that's a very profound insight. Like what you're basically saying is that it's really surfacing what people know collectively at scale and getting at the truth in that way. I mean, that's a very profound thing. I want to pause on that for a quick second.

Exactly what economists have found is that these markets are actually very good at producing predictions, which tend to be more accurate than polls. So if you go to a prediction market, for example, the recent election with trump and Harris, you can buy a asset which pays off a dollar if trump wins and nothing if trump s doesn't win.

Now you think about how much are you win to pay for that asset? Well, if you think that trump has has a seventy percent chance of winning, and you go to the market and you see that the Price of that asset is fifty five cents you gonna want to buy, because you're buying something you think is worth seventy cents, seventy percent chance the trump ins, and you get a dollar, and you can buy for fifty five cents, so you expect to make fifty and and by doing that, you push the Price closer to seventy cents, so you can interpret the Price as a prediction. And in the most recent election, the prediction markets were tending to predict a trump win, even when the polls were closer to fifty, fifty.

especially the public market, C. E O said. A lot of people trust the market, not the polls, at least when I came to the election. Like do you guys agree with that or no? I'm just curious because if that's a place where we can quickly do some hyvert s signal.

I don't think polling is dead. Pulling is one of the inputs into, uh, prediction markets pretty useful. I do think people need to be more sophisticated about how the poll and who who the polis.

Pretty clear that a lot of people now obviously not answering their telephones, and a lot of people don't want to talk to the pollsters. So there needs to be some new sophisticated techniques, but there has to be ways of drawing information from asking people questions. That's not going away .

best on twitter that like landline poll response rates in the olden days were like above sixty percent, but today the response rate like five percent, which means are getting like a very bad sample bias in terms of who's willing to answer a call on a pool like i'll hang up right away. Someone tries bolling me yeah.

In particular, it's not that like prediction markets will out mode poles is actually they're gona lead to revolutions in technology for doing this. Well, if anything, like the availability prediction markets increases the incentive to conduct polls. Red lake, you know, as we literally saw with the whale, they went out and ran their own hole precisely because they thought they could use IT usefully in this market.

That's fantastic. I have to ask. So so this may seem obvious to you, but the key point is that you're putting a Price on IT where people are putting skin in the game essentially with their opinion or prediction, so to speak, and that seems very interesting and useful.

How is that different from bedding? I mean, prediction markets be incredibly tiny amounts that don't have big value to be valid. Like how is the pricing part of this all work in terms of the incentive design?

Well, so at some fundamental level, the pricing works exactly as alex describe. If you think the probability that the trump is going to win at seventy percent, you see the Price of fifty five cents. If you believe your prediction, you know how you know an incentive to show up and buy.

And like, you know, if enough people have beliefs of different types and they all come into the market and they all purchase, eventually the Price sort of convergence according to the convex combination of all of their different predictions. But when you ask about like the size of the market or the size of a betting market doesn't matter once people are there and they've already formed with their opinions, but IT might affect the incentive together information. For example, you know, if the size of the market is capped a thousand dollars and you think the probability is seventy percent, you're not to invest like ten thousand dollars to get a more precise estimate, right?

Like if the maximum possible upside for you is on the order of a thousand dollars, you can possibly invest more than that to learn new information that will change your estimates and thus potentially sort of like inform the information in the market even more. And it's funny. Mean, we've been talking a lot in the wake of this most recent presidential election about you set of prediction markets as having been very strong predictions in the trend direction of what actually happened. But of course, if you look at, say, the two thousand sixteen election that didn't happen at all, right? The prediction markets totally didn't call trump and they also didn't call braxy, which happens to .

like the preceding summer yeah yeah.

And like there people were asking, like what happened? Like how do these miss this? And at the time I I wan opinion column where I argued that this information thing was a key part of the story that like at least at the time, tradition markets were relatively narrow, both in terms of the total amount that could be the total website, the total amount that was been closed in the market and in terms of who participated in them, right there was sort of like concentrated in a small number of locations. And those participants, because the upside was not necessarily that high, didn't necessarily have an incentive to go out and find out you like what's going on in other parts of the country. And so you end of aggregating information just from the people who are already there, which might not be a good estimate .

that circumstance. And I want to push back a little bit. And what's goys?

Yeah.

that's what I went up parts on the table. I am much more of a light prediction market bear than Alexis were both really excited about them, but there's a stack rank in our well.

So I agree.

you want a thick market, of course, and IT helps to have people willing to bet a lot of money because then they're willing to invest a lot in making their predictions accurate.

The part which I want to push back on, however, is this idea that the market did not predict, well, if IT predicted a forty percent chance of trump winning, and trump, you know, actually one, right? Because this is what people always do and frustrated, right? Because you can go back and look at individual examples and say we will do the market predict well, but that's just like, you know, you flip a coin and is fifty percent chance of coming up, kids and that came up tails you say, oh, well, your your probability theories isn't very good, is that 6本there would have a fifty percent chance came up one hundred percent。 So what's the real test? Well, the real test is you need A A lord sample of predictions, which could be predictions from political markets.

But prediction markets predict other things as well. You need a large sample. And then you have to say, in the sample of cases in which the market predicted forty percent, a win of, you know, the republican, whatever of that, how many times of the republican actually win? And what you find is that pretty close forty percent of the time that the market predicted to win, forty percent of the time republic is actually one in those cases.

So in in other words, there are sort of the linear relationship that when the market predicts a high chance of winning, that happens a lot. When markets predict something with a low chance of winning, that doesn't happen very often. But of course, sometimes IT does happen, right, you know, right? Something that happens with only a five percent probability ought to happen one every twenty.

And that's exactly what you see with these prediction markets. They tend to be more accurate than other methods of forecasting, and they tend to be not, not systematically biased. We can talk about there's some odd biases which are possible, but they tend not to be systematically by.

So it's not the case that something which is predicted forty person of the time actually happens twenty percent of the time. The markets systematically get forty percent of the time. It's predicted forty percent of the time that happens.

Let me give a simple non market example, I think illustrates tes this kind of a famous people ahead of the widom of the crowds, right? And so you ask people, how much does this cow? Does this cow away? And people are not that good, you know, feeling out how much a car away.

Summer too high. So mer, summer too low. But if you take the medium prediction of how much the cow, the medium prediction tends to be very, very accurate. So in a sense, the crowd knows more than any individual predict knows. And in the same way, markets do the same thing, they embed in the Price more information than any single individual knows.

right? And just be super precise that you are specifically saying the media, not the mean, not the mood IT has to be like the exact middle point, literally not like averaging out from the extremes .

in that particular example.

Yeah particular example but I did vary by by context, got IT exactly. Let me build on that. And like illustrated began sort of like through a simple example, but in the language of the Price system. So when you're going around and pulling people about the weight of a cow, you do have to go around and and ask them, and they don't necessarily have a strong incentive to figure that out. But suppose you have a very large amount of money to invest in, in commodities or commodities futures or something of the sort.

And you have a predictive model that tells you what you think is gonna happen to these markets, like you have a reason to believe that there's going to be a big shortage of oil or surplus of orange shoes or something of the sort you can buy and sell in the market in a way that reflects that estimate that you have that pushes the Price accordingly, right? So if you think this going to be a big shortage of oil, you're gona stock pile oil today. You're going to buy a lot of IT today and that's onna push up the Price because, you know suddenly there's there's more demand than there was before.

And so when you see the Price of oil going up, it's like it's a signal that somehow people think oil is more valuable right now than IT was five minutes ago, by the way. Of course, each of these are all hypotheticals. Like not of this is investment advice, like people should duck out like buy a bunch of oil or oil futures or whatever.

But like conceptual, that's how the Price reflects the information. And the more strongly you believe that there is going to be a shortage, the more you're going na be willing to pay to buy right now, right? And thus the sharper of the Price movement even to have, the stronger the inference about the information that that buyer brought to the market.

Yeah if you want to know whether there's going to be a war in the middle east, keep an eye on the Price of oil.

I remember that is a child in the eighties.

and it's still true today. That's exactly right. And though that the oil market is a little bit of a prediction market too, right? The oil market is revealing information about people's beliefs, about things that are correlated with the availability of oil. Yes, like whether this are wear in the midwest.

that actually goes perfectly to the question is about to ask because I still want to dig a little bit more into the economic and market foundations and then we can go more into the chAllenges of prediction market and where they're going.

But on that very note of oil, actually, great example, Scott, the question I wanted to ask you both is where does this break? Because in the oil example, one could argue, well, it's not like a quote, pure market. You have cartels, you have other forces that play. Now you might be saying IT doesn't matter because all that matters as people's opinions, which is what the prediction market is, putting his inputs into the market or doesn't IT matter. I guess my question is really getting at what are the distortions that can happen here, like there are things that can manipulate IT or other distortions where people's behavior changes so significantly that they untether the .

market from reality. I mean, things about these markets, you know, oil predicting possibility of war in the middle east, of course, they are not designed to do that, right? And in those cases, the information is sort of a leakage is an unintended consequence of market behavior, which is very useful.

It's very useful for economists to be able to pull information out of these market Prices. It's with the creation of prediction markets, which was really the first ones go back to the iwa political prediction markets created in one thousand nine hundred and eighty eight. IT was there almost for the first time that a market was created in order to produce information, right? So there's a much of a more direct connection between the output of the market, the Prices on the market and the predictions because that's what they were designed to do.

Now of course, you're totally correct that if you want to get a market to predict the future, you're gonna want a Scott said earlier that have lot of people because you you're going to take advantage of the dispersed knowledge because, you know, there are people in pennsylvania who have extra knowledge, you know about what their neighbors are talking about, you know, that can give them a little bit of inside, right, that you might not have if you're living in new york or send from cisco. So you want a lots of people to participate, and you want the market to be quite thick because you want people to be able to want a kind of invest some time and energy. But the production should be IT, maybe applies some models, perhaps do IT things like that. And of course, you wanted to be free and open, and you have to be a little bit worried about manipulation. Yes.

there are some like funny edges cases that we've seen crop up occasionally. In fact, there were even allegations that maybe that was going on here where if there's some external outcome or even some like internal like behavioral outcome, that conditions on the prediction, right? So if like political candidates are going to decide how hard to campaign and the given state, based on what the prediction for that state says, you might want to influence the Price, not for the sake of earning money in the prediction market, although that might happen to, but rather because you just want to place the prediction in a given position.

Now that's very hard to do because you actually have to change buffs from doing that. In the I think that was the obama versus mccain campaign. Somebody tried to sink a bunch of money to move the mccain e percentage.

And then, you know people who we had estimates that the obama probability was higher just out of arbitration, that out over a year of an hour two right? Like you markets work. If you see something that looks to you like a market anomaly, you'd buy or sell accordingly. Yes, yes.

the market periods here. So that seems like that working.

yes. But if the market is thin or if the information signals are very dispersed, maybe you can convince people, right? If you have enough money to like swing the market in a very sharp way, especially if you're doing IT through sibs, like many identities, if you doing through many identities.

So what IT looks like A A surge of people who have a given belief, you might actually change the beliefs of the market participants in a way that actually distorts the probability and could have various other impacts. And then the other thing is this idea that the oil markets are leaking information was still with that example, the oil markets leak information of a potential conflict in middle east, right? That's a feature and a bug, right?

The fact that it's an oil market that is informative about the middle ast on the one he had, as alex said, that means that the market is not optimized for specifically answering the question, what's going to happen in miles? There's lots of other stuff that affects the oil market like you. How popular electronic vehicles are at that given moment time rate.

So you have this very complicated single extraction problem, right? You see a big spoke in the oil Price is IT because. There's a potential conflict coming in the male's or is IT because there's just been like some new electronic vehicle test that failed and like somebody knows that.

And so they know that oil is going to be more important next month. Whereas if you have a market that's just predicting will there be a conflict in the middle ast, that's all it's but of course, that now is zero. Some market it's sort of a harder market to participate in if you only have dispersed information, right?

If you don't actually know whether there's a conflict in middle ast for coming, but know that some things that are happening like sort of suggest that, for example, you saw an oil Price change, you have to do a much more complicated and you're taking a slightly in in some ways of risk your bed by participating in a prediction market where you're taking everything on this one outcome rather than on something that's like heavily correlated. Many different things that could have related, you know, related predictions could be mostly correct even if your main prediction is wrong. So the takeaway is prediction markets narrow.

This is a feature in a bug is sort of dual to the sense in which ordinary markets of broadness is a future in a bug, right? Because of prediction market is a narrow zero sum contract on a specific event. Many people's information about that event is actually coming for all these correlates.

It's not that they know specifically like is there are conflict coming in the miles. They see lot of potential signals of IT. And so if you're buying and selling in a market that responds to those signals, that certainly ensures you a little bit right. If you get the main estimate wrong, but all your signals were correct, you you're at less risk than if you go into a prediction market and had all the signals right, but the final estate wrong and then you know you're just betting on the wrong side of the event.

I think what's god said also has implications for when we have prediction markets in everything. I mean, if these markets are so great and they works so well at predicting things, you know, we have more than and I think scot was basically giving the answer there. This is how I would put IT.

You know, if you have the market for oil and there are a lots of people who are buying and selling oil who are not interested in what's going on in the middle east. okay. Yes, they're not trying to you know predict that right, but is precisely because you have lots of sort of organic demand and supply that this provides a subsidy to the sharks who go in there in order to make the Price more accurate.

Or take the example of you eat, there are lots of farmers who are buying and selling in the market for weed just to ensure themselves just a hedge themselves. And it's because of that native organic demand that the market is thick enough that you then have all of the sharks who are not themselves farmers, but they go in there and they use models and techniques and whatever to predict which way the market for we is gonna, and they make that market more accurate. Now if you didn't have the organic demand, then you're gonna have a market with just sharks in IT. The sharks and who wants to be in a market where you're only with other sharks, you're right. Is why if I know that the other guys just trying to predict this one thing as much as I am trying to predict that, you know, I don't want to be in a market with Scott, he's just too smart, right?

I would say the same thing about you and .

that's why the market wouldn't work. That's why the market won't work. great. So some of these markets, even though they might be forecasting something which is useful, there isn't enough organic demand or you have to subsidize IT from outside the market in order to get a useful prediction of that in order to get the shorts willing to go against one another to try and predict this thing.

And that's why we don't have markets and everything yet. Potentially, this is maybe jumping ahead a little bit, but I just have to ask at this point, I mean, Scott, you're like a market design expert.

So on the market design front, what does that mean if there isn't organic demand? Is there way for market designers to essentially create markets in situations where there isn't that kind of late organic or existing thing to harness? Like can you actually manufacture that market without distorting IT and kind of create conditions that could design a market into place? That's a great question.

I mean, there are two different ways to get at IT. One of them is which is sort of with the framing the question is pointing at, is could you find a way to create late demand? And alex were saying you could subsidize IT, right? You could basically like somehow subsidize the experience of some people trying to predict this event. You subsidized a bunch of coins developing forecasting models, then they have a lower cost of entering the prediction market or something of the sort. Again, not advocating this specific policy.

right. Although analysts is an example, that subset was not intentionally as subsidy a result of the behavior like IT wasn't like people are trying to subsidize IT was just subsidizing because of .

their natural behaviors. true. You know exactly. So like we started this conversation with the recent presidential election and all of these other social elections.

Those have proven, at least in practice, to be much thicker markets because there are some people who seem just interested in bedding on the right. A lot of people have some amount of information of the ion, and so there's a little bit of that later demand. That sort comes from people's general interest in the question.

Yeah one could try and create that for other context, right? You can try and like help people feel that something is interesting or feel that they have an opinion about IT enough that they are willing to participate. Prediction market.

The other thing you can do is you can use other types of informational alita. Mechanical prediction markets are one of many ways of doing incentivized information aggregation, and others are things like incentivized surveys or pure prediction mechanisms. This a whole class, what are called pure prediction mechanisms.

We are what you're an effect doing is asking people what they believe about an outcome and what they think other people will believe about an outcome. And then you use sort of their beliefs about others as a way of cross examining whether they were telling the truth. Because you survey a lot of people, you get sort of like your crowd volume.

You think of no, the aggregate belief of the population. And you can check whether someone's own belief about the population is sort of the right mixture of that aggregate belief and their belief. So like if you yourself think the trump is more likely to win, then you yourself are more likely to believe that other people think trump is likely to win because the frame you have, the you're and sort of like indicates that at least one more person and at least one person in the market believes that.

And so one can cross examine your predictions with your estimate of what the population believes and what the population actually reveals that they believe. And then you can reward people based on how well they did in, in fact, like how good are you at estimating whatever one else thinks, given what you think and those sorts of mechanism? Ms, you can incentivize.

You can pay people immediately. Incidently, unlike prediction markets where the event has to be realized, the payments only realized at the end here. You're not paying people based on their accuracy about the event.

You're paying them based on their accuracy about everyone's estimates. And so you can do that all at once, right? Collect all the estimates, pay people they go home, you have your estimate. And these have been shown in practice to be very effective for small population. Or like opinion estimates, thinks where there isn't a thick market and like a very, very big public source of signal that really .

answers that question. And by the way, that brings up a very important point that we do not address on the recent example of the election, which is the french quote, whale, who won by using neighbor pool, where, you know, their neighbors won't say what they think, but when you ask them, like, who do you think your neighbors are gonna vote for, it's kind of a way to indirectly revealed their own preferences.

And that's the so called neighbor pool. I don't know. That's a standard singer that just came up in this election, as which I I heard of IT. But it's a great example of something. I did study in grad school when I was doing geography work, which is never trust what people say they're onna do but what they actually do, this goes to your economist world of reveal preferences.

Absolutely.

right? Very similar. But anyway, in that case, that person pulled his neighbors and then use that data, especially off chain, to then go back onto the market, quality market, in that case, to up his bet, and essentially one big as a result.

So like that would be an example what you are mentioning ing. Although in that context, you mentioning in how can we address a case where there is a thin market? This is a case where that played out in a thick market of the election.

Well, you might say that he was using this in the thin market of tried to understand his neighbors, sort of like local preferences and estate.

Here you go. That's more precise. yeah.

Although we actually don't know the details of how we produce these estimates, that doesn't sound like they were incentivize. So it's not exactly like what I was talking about with peer prediction. But you're right, is the same core idea that like using people's beliefs about the distribution can be much more effective than using their personal beliefs a lot of the time.

So I would underline two things there. One, yet the market is a way of bringing all of this disperse information and creating and aggregation, but is not the only way that's of what's got saying right and understanding this is one of the first information aggregation mechanisms, which we have studied and understood reasonably, uh, well. But there are other ones. And so you can think about prediction markets as being one example of a class of mechanisms which take to burst information and out of that, pull some knowledge, which none of the people in the marketer that none of the people you pull, none of them might be aware of IT. And yet somehow IT is in the air as IT were.

That's fantastic.

There are also other ways of subsidizing these markets, which is something that corporations may be very interested in doing because corporations are interested in forecasts in the future. And some of them in the past have created their own internal prediction markets. So one famous example of this is hua packards.

They were interested in forecasting how many printers are gonna sold in the next quarter and in the next two quarters, three quarters. Four quarters and so forth. So they created a market where, if you correctly predicted how many printers would be sold, in which time period you could earn money, and they subsidies that market. So everybody going in, which is just hp employees, got like a hundred dollars to play with. So that's a way of trying to get more people involved and interested in playing on these markets to illicit this information.

That examples actually really interesting to me because when I was at zero ox's park, we talked about that. And one of the things that came up is it's a very useful mechanism to europe, alex, for getting certain things right. But that is not a useful mechanism for actually figuring out the future in terms of what to invent because IT doesn't address the case of you don't know what you don't know, you only know what you know.

And this came up just yesterday. Trump announced his candidate for attorney general. And one of the examples someone cited on twitter was at the first time, they've seen a Polly market contract resolved to zero for all potential outcomes because gets wasn't even listed among the twelve potential nominees in those ranges of possible outcome. So that's an example in that case where you have to have the right information itself in that a prediction market. And you maybe you guys going to explain that a little bit more too really quickly because I think that hp example is super interesting on multiple levels.

Yeah, yeah. So these bargains are good at when you figure out people have got some knowledge and it's hard aggregate that. The other thing they're good at, you know, the people have run these markets for predicting when a project will be complete, right? And this glasses case, but you ask people that if we can be all no problem, it'll be ready.

But here in five weeks, you know, whatever, right? They're very optimistic. And yet they tell the boss is gonna ready in five weeks.

Well, they go back and tell their friends, oh my god, it's delay. You can get all these problems, but if you let people bid anonyme sly in these markets, then the truth comes out. So this is a way of the corporate leaders can learn information that their employees know but are not willing to tell them. But to get larger point, yeah, I mean, nothing is more difficult to predict .

in the future, right?

And you know trump is a chaos agent, right? Whatever he's going to do like IT is hard to predict. And I agree, I don't think anybody predicted that gets well.

And indeed, actually, so this sort of highlights and when we were talking about what prediction markets are good adverse, where you might want to use other sorts of informational alita mechanisms, the two examples that alex gave of within company prediction markets, you know predicting sales, you are sales growth and something that's like you, a metric that many people in the firm are tracking and have different windows of information into predicting when a product is going to launch.

We're like, you know, you might have product managers who know something. You might have engineers who know there's a hidden bug that they even told the product managers about. Again, it's like these are context where many of the people in the company have some information that only they have and that the aggregate of all that information is a pretty good prediction of the truth because the actual outcome is the aggregate of all those people's information directly, right?

It's like how many sales calls are you making that are succeeding? Or you know how is the coding for this c feature going? By contrast, you mention with zia's park, you know trying to predict whether a new sort of totally imagined product is onna succeed.

That's really, really hard. And he doesn't rely on information in particular that the company has, right? Like, yes, the company has some idea of what products people might buy. You might be like to A T N T and event before picture phone or something of the sort. And like you thought that was a great idea, but you don't actually know until you put in the market and see whether people are like interested in using IT. And so the aggregate or of all the information in the company there, there's a product they went through with, right, they concluded was a good idea based on all the signal that everyone in the company could see and it's still flocked.

The total information in the company wasn't high enough to actually like provide the writing answer we even when aggregated, right? But I do think there is a sort of subtle distinction between wisdom of a random crowd and wisdom of an informed crowd, right? Like again, with a hully packet red example hully packager sort of knows that if you're trying to figure out now like IT weather or products to the launch on time, a random person on the street has no information about this.

You don't want to like, pull together a focus group of missoni fulla packet red customers and asked them, when do you think we're going to finish designing our new printers? right? I don't know. Like you release the printer last year, probably next year, maybe, who knows? And so there is this question.

Are you learning things from the right crowd? You know, you can have the best incentivize informational alita mechanism on the planet, and if you only survey people who don't know anything at all about the topic, you are incentivize them. You will learn what they believe truthfully, but you won't be able to do anything with that.

yeah. And then back to the future, like the whole idea of the best way to know, predict the future is to invent IT. Like that was like, just like the jobs of the phone. Like no one you can ask a million people, will they ever use a touch phone? People's behaviors can also evolve and change in ways that they themselves are not aware of.

which is that other that example for the market is like a candle in a dark room, right? I mean, that helps us see a little bit, but they're still areas what you can see very far.

great. I'm going to ask couple of quick follow up questions from you guys so far. So just to be super clear.

So then versus ick, you guys, you're talking about the depth of the market like in terms of the number of participants. Thin is too few, dick as many. Is that characters they're Better, more precise way of defining that?

yeah. So I mean, in the predict market, a thin market is few people betting small amounts. And in fact, one of the problems we've had is that prediction markets are mostly illegal in the united states.

So the biggest one in this past election was Polly market, which he was illegal for U. S. Citizens to bed on that market.

We're slowly changing, but we do have this kind of ridiculous situation. I think it's ridiculous anyway that we have huge markets in sports bedding navels, right? Huge, huge markets and we allow that.

And yet here we have the kind of gambling market, a prediction market where the output is actually really quite useful. It's quite socially valuable, and we don't allow IT. So making these markets legal and open to more U.

S. Citizens would thick in those markets, make them more active, attract more dispersed information. And I think it'd be really quite useful.

But to your bigger point, alex, you're basically arguing that they can be a public good in the right context informationally. absolutely. And interestingly, if you think about some of these prediction markets that are getting serve notices and what not, we don't know why, to be clear.

But it's interesting because in some cases, people might argue some people trying to get information is the manipulation of the market. In fact, to your guys sa's entire point throughout this discussion, it's actually ways to provide more input of information into the market itself too. So that kind of interesting on the public incident.

Let me give you another example on this public good nature of these prediction markets. One of the most interesting fascine uses of these prediction markets is to predict which scientific papers will replicate. Oh, you know, we have this big replication crisis in the scientists, pology and other fields as well of you know a lots of research.

And IT doesn't replicate well. What some people have done is it's expensive to replicate a paper. But one thing people have done is to have a betting market, a prediction market in which papers will replicate, and that turns out to be very accurate.

And then you only have to replicate a few of those papers in order to have the markets pay off. And for the rest of them, you use the prediction market result as a pretty good estimate of whether IT will replicate or not. So this was a way of improving science, making science Better and quicker and more accurate.

I love that I bring a lot of events that weird on open access and science, and kind of like evolving you know, peer review and replication crisis and the whole category and theme. So it's very exciting me to hear that that something that we can do to address that IT leads to a quick follow up question, which actually happens to behind my bliss of follow up questions for you in the lightning round of this, which is when you have you're talking earlier about this, this kind of tapping into this intuition information dispersed across many people into this prediction markets.

One of the first questions that came to mind is, do you need domain experts or does actually distorted the market? And this actually comes up as a perfect segway from europe, alex, an example of scientific papers, because that's the case where one would imagine that people in that industry, or that domain, or just other scientists who have the experience of analyzing research, would be the best at predicting things. But is that necessarily true? And do we have any researcher data into domain expertise in these markets?

I don't know the answer. That last part, let me talk about the first part because IT also speaks to your thick verses.

thin. right?

yeah. good. right. So when alex said a thin market is small number of participants spending small dollar amounts, why is that a thin market is because the total information is small in two ways.

One is that there are few people bringing their own individual estimates. Just have like a small number of people saying things. And second, because they're betting small dollar amounts, sort of a signal that the information is not very like strong signal or confident, at least relative to what I could be otherwise.

You know, if you are staking a very large amount of money on this, the market inference is that you have done the research, you know, and indeed, you have the incentive to do their research. Know why is the influence that you've done the research? Because if you're taking a large amount of money, you should have done the research because otherwise you know you're putting .

money in risk without sort of full information .

like the french whe who did the neighbor find out you how good or the that people. He should have trusted his information that much, but it's unambiguous that part of his confidence and he said this part of the confidence that he had to make that huge bet, was that he thought he had a signal that was accurate and the market had missed.

And so like dicis and thinness, like the proxy for the way we think about measuring IT is how many people and are how much are they staking, like how much value are they putting behind their beliefs? Thick and thinness is really in terms of the information, it's do we have a lot of different signals of information that are strong coming together and mixing to determine the Price, really just like a very small number of pretty uninformed signals that this tension, when Alexis saying it's a problem, that the biggest prediction market for the U. S.

Election was not actually in the us. And and was not legal to participate in the us. Well, yeah, a lot of the information, a lot of like real signal is in the united states.

And so without those people being able to participate in the market, you miss at least sort of a lot of that to a first order to radio. People internationally will be figuring out ways to agree and sort and try and use IT. But like you miss a lot of the people who have that information already of their fingertips.

And so you ask about domain expertise, it's not exactly domain expertise for is not, but rather information richness. And for example, in predicting scientific replication, successor failure. So main experts are especially well equipped to do that, right? Like a random person chosen off the street.

Now you can tell them a scientific study baby will have an instinct, one where another, whether they think they believe IT. But like a lot of the detail of figuring out whether something will replicate comes from knowing how to read the statistical analysis, trying to understand the set up of the experiment, and like they were surrounding literature. And so they are domain experts have a particularly large amount of information.

If you think about something like a political batting market, maybe the domain experts who are focused in the world of politics and polls. And so we have like a big slice of information they do. There also might be other categories of people, like people who know that their nights or hood has like recently switched its political event ation in the way that isn't a captured in the national polls. Or our french whale who went and ran his own sort of pole using a customer chosen method. And so the context of the question, the prediction market is trying to and this actually for any information alita problem, this is just about prediction market for the context of the type of information you're trying to learn tells you something about who has the most information to bring to the market and does who is .

important to agree. Everything is god said. One of the interesting things is you often don't know who the domain right until after the market has being run. So of course, is absolutely true that you know, if you're gna be predicting political events, you want people who are interested in politics. If you're predicting scientific articles, people need to be able to read stats and things like that.

But one of the guys and the scientific replication paper on markets he made, like ten thousand dollars, was just one of the super obsessive guys, right? You just really got into IT and you know, was running all kinds of regressions and was doing all kinds of things and stuff like that. And so when you say domain x, where I think one of the virtues of these prediction markets is that they're open to everyone and they don't try and say no, only the experts, you know, get to have a voice, right? It's more only expose do we learn, hey, who who really makes a money at these?

right? absolutely.

I'm so glad I asked you guys about the definition of thick sus because you guys gave me so much interesting nuance to that because people, I think, following this pad guys definitely understood what you meant about thin versus tic early on. But you guys just took IT to a new level.

if you so smart. Why you rich? A I am rich. Yes, I made the money in this .

market well and that again, that's about the incentive we talk about, like the dollar value stick, like the amount of money someone is taking on their prediction again in eglise. Um IT should be a measure of their confidence, how confidence they are in their own beliefs and how much effort they have put into. Learn the information to be precise. And so exactly as I exists, one person who might be really good at predicting a scientific replication failure is someone who works in that exact same area. Another one who might be someone you just like enjoys doing this for fun and like has never had a really incentive to triple down on doing IT.

But now suddenly they can, right, right? And by the way, god, does that have to be dollar and Price incentives? I'm asking you this questions specifically this. And I have done a lot of pieces in the past on reputation systems. And I almost wonder if the skin in the game can just be calm up points and not even any money because I think from .

a pride perspective, one hundred. So like alex mentioned, subsidy, right? Like one way that you can subsidize. I think he's that hute packet ard subsidized by giving all their employees a hundred dollars and saying spended all on this market, you can subsidize people with cash, but you can also have subsidize them with tokens or your reputation or like various other resources of value.

And one of the advantages of using tokens is that that way you can deliver a subsidy that sort of only useful in this market, right, if it's like a personal non transferable token. But I give you a bucket of them. And the only thing you can do with that is used IT to enter predictions that you just choose which prediction markets you choose to enter into and how much you spend in each one, right? And then you earn payoff.

Payoff s are also measured tokens. And maybe downstream, you might get prizes for having large numbers of tokens or something. You get to join the elite predict force, or even just serves the measured of your reputation, how good you are making predictions, which maybe you leveraging into something else, right? Like people who win data science contest, leverage that into data science jobs. Maybe you like lever to this into a forecasting job or something, all of that, so long as you find people who are willing to be incentivised by those types of outcomes, you can subsidize their participation in a unit that locks them into the market, right? That they're one thing to do with IT is to participate in the market and reinforces more and more participation among the people who are most successful and most engaged.

That super interesting, and i'm going to push back on you and that you actually wonder necessarily needs to be crypt of bed .

and you can just do any kind internal mark. But for all the reasons we Normally know, like it's much Better to do this in an open protocol form because, for example, if the tokens eventually going to be leverage for reputation, you want anyone to be able to verify .

that you have to do to see a hence block chains. Got IT. great. And we will talk a little bit more about that, just more relisting questions.

So where do super forecasters like Philip tut locks work come in to all of this? Like are they especially good at prediction markets? Because that's a case where they're like generally Better at the general public in sort of quote, forecasting and making predictions.

Is there a place for them in this world door? They kind of the outliers here? Or does that not even matter here?

I think there's two things. One, I think the basic lesson of tetlock work is most people, even the ones who are in the forecasting business, are terrible forecasters, right? I mean, he's first started tracking so called political experts and seeing what their forecast were.

You know, ten years later were the right, five years later, and they were completely wrong. So he then shifted into looking for, is anybody ever write are the super forecasters? And yes, he found that some people, you know, not typically the ones in the public eye, but some people can definitely forecast Better than others.

One of the things those people can do is then participate in these markets and buy their participation. They push the market Price closer to their predicted probabilities. So forecasters have an incentive to be in these markets and by being in these markets, so they make the markets more accurate.

Now is the market always going to be more accurate than the super forecasters? no. I mean, warn buffett, you know, he has made a lot of money even though markets are basically efficient, but warm buffer t has shown that he, in many cases is able to predict Better than the market Price itself, add more power to him.

And so they'll going to be some super forecasters, but they're hard to find. They're rare. And a virtue of the Prices is that everyone can see that right is your public.

So this actually gets out of bigger maybe more obvious point to you guys. But a recurrent theme i'm hearing is it's not that the prediction market is only taking in like guesses and people's intuitions and bets and opinions and any information IT has, but theoretically done well. It's taking in all information.

IT could be super forecasters contributing to IT. IT could be made silver, taking as eight thousand simulations and feeding his inputs and adding that signal into IT. IT could be people who are pollsters, putting their data and predictions basically doesn't matter where how people get at their intuition. All that matters is that their pricing out information into that market essentially.

Do you know the wall street bets is a famous everything is a Priced in the post?

No, I don't actually, I don't know either.

Let me read a just a little IT. It's a fantastic post is like five years ago, it's skilled. Everyone is Priced in and he says the answer is yes.

It's cried in think amazon will be the next earning that's already been Priced in. You work at the drive through for Mickey and found out that the burgers are made of human meat that's Priced in. You think insiders don't already know that the market is an all powerful, all encompassing being that knows the very internet workings of your subconscious. Your very existence was Priced in decades ago when the market was valuing standard of worlds expected future earnings based on population.

That is so great. okay. You have to tell you that lean gala's and then all these in the show notes. So you're basically agreeing that it's the markets you .

Price everything in. Yeah, I mean, that's an exaggeration. But yeah, I mean, anything is is fair game.

I want to push back what we're fine here because anything is fair game. But you have to wonder who's going to show up to those markets and where the signals are coming from, right? like.

If you're a super forecasting, maybe you work for like a super secretive hedged fund. And the last thing you want to do is directly leak what IT is you believe? Yeah, yeah.

And in fact, you would prefer that the market be confused by this public signing. We meticulous, you might show up and tank the prediction in one direction of the other, just to take advantage of that in the financial market off to the side. And so while in principle, these things can be very comprehensive, you still have to think about who participates in which market were just like we see in other markets. We're like some people trade in dark pull, some people trade public exchanges and that selection sort of a what information Prices really aggregating where .

it's fantastic. yeah.

The other thing about public forecasters, super or otherwise, is that they're very silent to the average person. And so another thing we see in prediction markets is heard behavior.

Again, just like we see in other types of markets like, you know, if a lot of people are suddenly buying oil futures, does that mean that they all have knowledge that there is going to be a conflict in the middle east, or does he mean they saw other people buying oil futures and are like, o gosh, like, i'd Better do this too or you know, did they see one analyst report and they also w the same analyst report and as a result, they all even bought oil futures because they believe the report? Or worse, did they see one analyst report that said, you are like, oil is going to be an expensive next quarter and they went bought oil futures, not because they leave the report. Maybe they even have information that is not true, but they know everyone else is going to see the report. And so there will be purchasing pressure.

Yes.

there's a very famous paper by Morris and shin and the american economic review called social value of public information OK.

I want to put them there.

So now IT talks about information hurting, right? The ideas.

Basically, if you have a market where everyone has private signals, and then there are some very silent public signals, and people have to record neighbors, are you going to run on a banker or not? Or like, what do you think is the probability of this thing happening? People might ignore their private signals if the public signal is strong enough that they think other people are going to follow IT? yes.

And so when a very prominent forecasters makes a prediction like as the sort of polls were coming in and the week leading up to the election, a new major poll would drop and then the prediction markets were judged or around and sort of viewer off, at least briefly in the direction of that pool. And that's this like public information effect, right? This is a sailor. You expect a lot of market movement based on this information. And so the market actually moves even more, incorporates not just the information, but also the fact that other people are incorporating the information too.

And are there any market design implications for how to avoid that happening like if you're setting up the conditions of a perfect, great reduction market?

Oh, gosh, that's a great question. I mean, first of all, it's not completely avoidable. You can't have a market where a sufficiently strong public signal doesn't generate some held behavior, right? Is just and that levels unavoidable.

But you can try and do things to damp in the effect of the time I had. I can think of two. There are probably others.

One is you could basically like slow trading a little bit. You can sort like limit people's abilities to enter or exit positions very, very quickly. So it's sort of forces people to like average.

Well, it's also kind of an example of slowing condition, right? Like an infection spreading very fast, kind like the hurting becoming viral.

A condition is a very good example. What's got talking about, you know, in stock markets, we have a break.

break. Yes, exactly. Circuit break ers. There we go. So that's one of the ways.

Another thing you could do is try and refine your market contracts in a way that orthogonal es, by which I mean, it's sort of extract out the signal that is independent of that signal, right? So a prediction market contract somehow in corporate, the information like adJusting for whatever needs silver claims.

Let me give you an example because my colleague, Robin hanson, who is one of the founders of prediction markets, rob, is usually many steps ahead. He is a very clever proposal for this, which I don't think anyone is ever implemented, but he says, you have a prediction market and then you have a second prediction market on whether that prediction market will revert in the future and is something else.

yes. Oh, so genius. Yes, exactly.

That's the way you do IT. That's the way you thought glides. perfect. It's way Better than my example like that.

So great. Because I was actually gone to guess something like combining the reputation thing. And this is essentially a way of combining reputation by having a parallel market that verifies and valid exactly totally.

That's so interesting.

by the way. That's not a few turkey, right?

His new thing, one of the criticisms of few turkey was precisely the point what's got made. And then Robin's response to that as well, the solution to a problem with a few turkey is .

more few turkey, okay, that you quickly define few turkey for me. yeah.

So Robin hansen's idea is, let's take these decision markets and apply them to government. Let's create a new form of government. You know, there are not many new forms of government in the world.

Deploy racy monarchy. You know, few tarka is a new form of government. And the way that would work is that instead of having politicians decide what policies to have, politicians and voters were just decide on what our metric for success is gonna be.

So IT might be something like GDP would be one metrics of success, but you won't want to adjust that for inequality or for environmental issues. So you're gona create some net ccs stic GDP plus. Then anytime you have a question, should we pass this health care policy?

How should we change immigration rules? So we have this new immigration rules. You have a market on whether GDP plus would go up or down if we pass this new law.

And then you just choose which one. If GDP plus goes up, you say, okay, we're going to do that. And so people would just submit new ideas to the few parking.

Here's a proposal for immigration. Here's a proposal for health care. Here's one for science policy. And then you just run a prediction market.

Would GDP plus go up with that? Or would I go down and then you choose whichever comes out? So Robin expands this idea of decision markets to an entirely new form of government that's fascinating .

and that relate so much to one of our partner's collaborators, where Andrew l. Is done, twenty cities, a lot on on chain and kind of liquid democracies and more. That's very interesting.

Thank you for explaining that, alex. Because i've actually never fully gotten what few turkey is. People tossed IT around and i'm like, but actually what is IT? I still don't get IT.

So that was very possible. IT also sounds like could be .

the subject of like a boy short story or oh my god, yes, absolutely goes the that put in the last reading that for the founder MIT was that the .

laboring sort of like, yeah, yeah.

So so funny. So give more questions and I want to switch to crypto. So since I was hacking actually about like kind of market theories and practice in this recent segment, alex, did you want to say a little bit more about efficient markets?

是， 是。 So another fascinating example of how markets could leak information, which then could be used for other things, is, if you ever seen the movie trading places, you probably know that the main determined of orange is futures is what the weather is going to be in florida, of course. So Richard role, who is a finance economist, I had this interesting question.

Well, can we use orange use futures to predict the weather? And when he found is that there was information in those market Prices which could be used to improve weather forecasts in auda, kind of an amazing example because, no, I get knew this. No one was even predicting this, but this was kind of A A leakage of this amazing information.

fantastic.

Another fascinating one is, you know, Richard find man famously demonstrated that IT was the old rings which were responsible for the chAllenge disaster by dipping the otherness in the ice water and the rational community, however economic ist went back. And when they looked at the Prices of the firms, which were supplying inputs into asa and of the chAllenger, they found that the stock Price of more vial, which was the firm which was produced, the o rings that dropped much war quickly and a much larger amount than any of the other firms. So the stock market had already predicted, in fact, in that he was probably the o rings, which were the cause of the chAllenger disaster, even before Richard five man had figured this out.

And by the way, it's another that ties back to your hp example in a way, because if I recall, part of the back story with the chAllenger was also that IT was a case of death by powerpoint because of the way they were communicating information internally and that the format and the structure kind of constrained how that information was presented. I think tough day gives a famous k study of this .

in one of us many books. So another way of hooding that actually, which is kind of disturbing. But I think you're right in that the people on the ground, they knew this wasn't a good idea. They knew this was not a good idea to launch the chAllenger on such a cold day.

And if there had been a prediction market of like what's going to happen or or should we do this, then I think this is quite likely that that disperse information, which no one was willing to tell their bosses, you know, no one was willing to stand up and say, we should not do this. Instead, IT got buried in powerpoint intent. This first information might have found its way to the top if there had been a prediction market. And is this launch gona go? Well.

exactly. You are said another way, another definition of a prediction market, IT would have been another way for management to list IT Better information from their employees. And using just that is a mechanism for communication essentially exactly.

Yeah, the hp thing really kind of struck me because I just remember red that is like a communication no no, for how information is presented. And that's actually a good segway, by the way, to the crypto section because I wanted ask you guys and this are going to help me break some. You know, I love doing a good taxonomy of definitions in any broadcaster, because one of the things we talk about encrypt s eat of decentralizing, sometimes the information is public, and a public blockchain.

It's often open source distributed IT can be real time. I don't know this necessarily act your information, but the information can be corrected very quickly, which then makes IT more likely to be accurate because of the speed of revision, which by the way, we also saw on the recent election. I think compared to media, one of the observations people made is that media didn't move fast enough to, or even one, two because of biases, their polls and predictions, where as the prediction markets were faster off correcting.

So one question I have for you guys gonna kick off the this section about the underlying technology and how that works as, first, let's teese apart. All those words I just gave you like a big buzz d bingo soup. But words, what are the words that actually matter when IT comes to this context of illicit Better information and aggregating that information in a market? Like, what is the key qualities that we should start with? And then we can talk about the technologies .

underlying that. One way of answering that question might be like the largest prediction market was the Polly market, crypto prediction market. And the question is, is crypto and necessary part of this? And I think the answer is probably no.

I think why was the cypher market particularly successful? Well, because I was open anybody in the world bar U. S.

Citizens, right? yes. And the market, because of that was much thicker than the other markets. So there are some prediction markets which limit people's best to a thousand dollars and the crypto whale was betting millions of dollars, uh, on these markets.

So that's why the encrypt market, I think, as a kind of regulatory arbitrage came very important. And you know now the FBI is going to looking at this, the french are looking at this was illegal, is IT violating some laws, but I think the crypto part of IT was not actually necessary. Yeah.

i'm glad to point that out to alex because I think people have been kind of helping and over inflated the crypto part of and I actually do agree with you completely like I don't know crypto was at the heart of the way that, that market works except in those qualities you mentioned. Scott, any thoughts on that point?

So I totally agree with all of that. One of the crypt og does very well on top of being open and interOperable and transparent is enables commitment, right? You can write a piece of software there's going to run in the exact specified way that can be audited by all of the users, and then they can be convinced that it's gonna n correctly. And some ways we do informational alita have chAllenges with commitment. If you are going to survey people and pay them six months from now based on whether their survey estimate was accurate or not, they might be worried that you're not going to show up and pay them.

And so long as whatever the information is can also exist on chain, where the resolution, the uncertainty can somehow be visible on chain, either through an oracle or if they were like an on chain function to begin with, like just what is the Price of this acid or something you can commit in a way that you can't necessarily or you can't do easily without complicated contracts? Yes, you can just commit that it's going to run as expected now in order for that to work, your informational vital mechanism has to be fairly robustly committed and often also decentralized. Like Polly, market, by contrast, famously changed the terms of a couple of their resolutions because something happened that didn't quite make sense in the context of the way the'd said they.

We're going to evaluate the outcome. And so they post talk. This is after people have already bought in under the original terms of resolution, change the terms of resolution. And so that's like A A lack of commitment. That's actually you know harvard markets to like form when people don't trust that are going to be resolved as described.

right? I mean, is that the most basic rule of markets like you can just suddenly change the rules under you in that way? We always talk about why we don't trust governments that don't enforce property rights and what not like.

You just can't mess around. No, you're exactly right. In the same way that block chains create a form of property ride that you can trust even without sort of a very trustworthy entity having established because the property right itself lives in this immutable table ledger.

Same thing here like you can, at least in principal set up resolution contracts that are trustful and immutable, able and therefore expand the scope of the set of marketplaces we can configure, right? Yes, you know, it's not just the set of tools we had. When you have to be a little trust, the market organiser. But actually now the sort of like commitment enables you to go further just to break this .

down a little bit more because I think you said some really important things in there. I want paws and make sure we flushed out for our audience. So first of basing what alex said earlier, in the case of poly market, one of the key points was public and their information being out there.

That one I mentioned earlier, the example of that being updated quickly as compared to media, at least you just mention the importance of credible commitments. And we've often describe blockchain as a technology that book chains are computers that make commitment. So that's a third or forth.

I don't run on the number count, but i'll just keep leasing the features. And then you also mentioned potentially decentralized, but I couldn't tell if I really needed to be decentralized. Can you give me more bottom line on decentralizing where you down there?

He has a great question. And actually, maybe we should have started here. The necessity of all of these different features moves around with the type of market, the world couple your information indication mechanism is and and this is especially important for the context. We're sort of pure information markets don't work. The more complicated your information alita that IT is IT is the more likely IT is that you want something .

that looks like cricket reals. Ah that's really good.

okay. So like if hulia packet d is running an internal prediction market first, well, he doesn't have to be opens in the entire world because you're only trying to learn information from your employees, right? So openness is important within the firm, right? Maybe there's someone in the male room who knows something that you don't know, they know. And so you actually want that market of people to be able to participate.

But hui packager does not necessarily care what a person on the street thinks about printer cells and certainly doesn't need to build the architecture to bring in like random people's estimates of printer sales, right? And so you know, you need some of out of transparency because if you need people to be able to see what the current Price is and like see whether they agree or disagree, and they can sort of move the Price around, but in other types of elicit ation mechanisms, maybe you don't need transparency, right? If you're just going to pay someone based on the accuracy of their forecasts down the line, you don't need them to be able to see what else is happening.

You just need them to believe that you have committed and that the final accuracy is gonna transparent, right? That they can verify that you didn't just stiff them by like the thing they predicted happened exactly, but you just said, no, I didn't and then you don't pay them. And so transparency is important only there with respect to the resolution, not with the respect to the interm states. By contrast, like commitment is incredibly essential and needs to be believed or else the user won't even participate.

right? By the way, great that you gave the example of the transparency, and i'll let you finish your example in the second. But i'm just jumping in because IT reminds me of how we talk about the things that can be done on chain and off chain when IT comes to scaling blockchain and the covers versus verifiers when IT comes to zero knowledge or what not. And it's really interesting you pointed that out because I want to make sure people are listening. Who are builders listen to that because that means you can do certain things on chain in order to whatever your goals of the design are and then put other things off chain like you don't have to have this purest view of how truth must be transparent is very smart to point that out and you keep going with .

your other example. yeah. And I completely agree, by the way, I like one of the things when I talk to teams, i'm constantly trying to get them to think about which features of the marketplace are the most essential for market function. IT IT varies by basket context. And even even if eventually you're planning on having all of these features, yeah right like as you're deciding like which thing do we build first or like as we're progressively decentralizing, like what do we prioritized, you actually have to understand the market context.

You're working in that so smart because it's basically another way to hit product market fit too because you're not like overbuilding and over features something anyway.

keep going with other side. So to gets to the question of like when does the centralization matter? The centralization has lots of different components that might make IT matter. One of them is just the ability like make these commitments even more enforceable like IT makes IT possible to be confident and function and lively and soft.

All of those things are important for a market because if your prediction market goes down the night before the election, you do, first of all, you lose the information signal from its second, will you lose the ability for people to participate in the market, which which sort of adjust the Price and move the signal around? Similarly, if you lose the ability to like, resolve the truth, then maybe you can finally resolve the market. And you have all of these bets that are sitting in limbo because the market doesn't know what happened.

The key is everyone is bringing in their own information. But in order to finally resolve the contract, to determine who gets the payout for the best, you have to have the chain have a way to know what actually happened. Another placed centralization is sometimes very important is in that resolution function.

Like, you know, if the market is on chain, you somehow have to get what actually happened onto the chain. And maybe the biggest Better happens to also control the one resolution function. And so they can now sort of rob the prediction market by just lying about the resolution of the event.

They tell the system like candidate a one when actually candidate b one. And then by the time people realized that this wasn't correct, they might got to have a way to fix IT. But even if so, person might just be gone.

So the centralization and resolution, just like we think about decentralized oracle sort of mechanism, this is basically an oracle, right? You have to. Bring off chain information on chain in a lot of these context to resolve the contract. Or if you doing this an the centralized platform, the users have to trust the centralized platform to resolve the contract correctly. By contrast, if the information does not need to be brought through an oral, if IT already lives in a system that's verified and the resolution is like provably going to do IT, it's claim, then you don't actually care about the centralization, say, in the discovery of the resolution and you're actually just like reading information in your commitment contracts s care of everything else.

And just really quick, it's got you've said oracle a few times. Can you actually probably define what you mean by oracle in this context? I know we talk about a lot .

in crip to yeah an indeed oracle is not a completely uniformally well defined term in this context. I'm talking about oracles as like a truthful source of information about what the actual of resolution of the event was. So if trump won the election, the oracle tells us trump on the election.

And if Harris on the election, the oracle tells us Harris won the election. And the reason we're using that is because the election is of course, not being conducted, at least maybe in the future we can dream. But in twenty twenty four, the U.

S. President, the election was very much not conducted on a blockchain. And so if you're going to have an on chain prediction market, you somehow need the chain to be able to learn the information of what actually happened in the off chain election. And so the oracle is like basically the source .

of that information. The key of the oracle has got, say, is to bring IT off chain and bring IT on chain. I mean, the thing about off chain is that people can look at the new york times, right? And so the new york times is often considered a oracle.

And that you go by whatever printed in the new york times, that would be a way of resolving a lot of bets. Like did the new york times report the trump one? That might be one way of resolving these beats.

Yeah.

but the key problem is to bring that off. Take knowledge on chain in a way in which the information is not distorted in the transmission. And the reason why that transmission you're worried about IT being distorted is precisely because it's the revelation where all the money is, right? So so there are big incentives to distort the transmission of that information.

In fact, a lot of the crypto hacks which have happened have happened because people found a way of distorting the oracle and then using that on the crypto market. So you know, the market resolved in one way. And if you can change the oracle, then you can make a huge amount of profit out of doing that. So there is a big incentive to mess with the oracle. That's why is really difficult.

And we can stick k with the new ork time to example, right? A lot of people are going to make their morning trading decisions based on what they see in the new york times, on the bloomberg terminal and so forth. And so if you could, in a coordinated way, feel the wrong information to that, I would change many, many people's behavior. And you could trade against that because you knew that they were going to get wrong information.

exactly. So this can happen in the off chain world. And indeed, we saw there was one tweet, right, that the S.

A C is going to legalize, know etf bitcoin contracts. IT look like, you know, is an official ruler. And IT turned out to be a hack to be correct, but that was not revealed until days later. But yeah, so if you can restore an oracle, you can make money totally.

Or I mean, retired at the new york times, IT would be remiss to not have to like dui defeat tremor, right? Famous, famous front page, like huge text headline that just turns out to be inaccurate, right?

That's a famous case of what we did immediately wear too. It's called the pre right. And then you accidently prepare sooner and you get IT wrong there.

You actually have been cases of someone who write their a bitter day months or years in advance, and IT goes out and says they're okay. You conflate IT earlier. And I agree there generally connected and similar, but there are some new ones as between decentralized and distributed.

Like distributed can just be like abundant systems that have multiple, like the system going down. What you are giving the example the night before something that's the case were being distributed matters. But IT doesn't have to be decentralized necessary.

Like I, E. There could be distributed nodes managed by a centralize entity, for instance. Lutely very clear the distinction between centralized as where are as.

By contrast with the oracles, for example, you might really care about being decentralized, right? You might care that no individual entity can sort of unattentive of change how the .

contractor resolve exact just one other point.

another advantage to doing all this stuff and block chains is that it's composer. It's not the word just like intrinsically interested in some of these questions like maybe so some people are just like, no, actually curious like who's going to win the presidency in a month, but rather like lots of other stuff.

Depends on IT, right? If you are making decisions about which supplies to order in advance, you need to have beliefs about the likelihood the terrorists are imposed under the next administration. And so having these things live on open composer architectures is useful because they can be wrapped with other information and other processes. You can tie your corporate Operations in a very direct way into these of information aggregation mechanism.

Yeah to put you in a more basic way, I know if I know composes the we talk about it's like the lego building blocks, the markets on chain or the information on chain is a platform that people can build around, build with bringing pieces of information, combine that with other tools at sea and you can create like different things. And that's a composition order. And i'll put a link in the shower notes oppose explaining compose base as well.

And then the other quick one is open source. Does the code itself have to be open source? Auditable public good.

Again, that depends how much you trust the the market creator yeah. And again, this this is true across the board for applications that can be run on block chains or not, like are always making tradeoffs between, you know trust through reputational incentives and institutions and trust through code.

For example, like in actual commodities markets, there is a lot of trust through institution and contract, but there is an architecture in place to establish the trust between the institutions and the contracts and their enforceability via the institutions for those contracts to be really enough that people believe in them, enough to pay money for them and to have these market features, blog k chains enable these sorts of trusted activities in lots of context, where the institutions are not strong enough or present enough to do IT for you if you're having like five dollar bets, like small money bets, on some incredibly minor question, like will the horse that wins the conduct, ky. Darby, have a prime number of letters in their name or something like this, right? You're not going to have necessarily an institution that is even able to evaluate and look set up that contract in a way that is worth doing at the amount of money is going .

to raise like a scot changes that can take you jervy into something you would be interested in that involved five numbers. This is we get our senses.

But five numbers, I love how I will help you .

know the the coachy darby is also interesting because as all sorts of call statistical questions .

going .

on and pool hats, fascinating heads, absolutely fascinating hats, and definitely intended. I love IT. So like substituting code for the source of trust for these like very unusual or sort of like micro or international, there's not a clear to restriction or all of these context of push you more into security via code rather than security .

via institution. Let me add one more point on the black chain. So I think generally speaking, as I said, the black chain is not necessary.

However, as we are looking towards the future, IT may become more and more useful to have these very decentralized rails so vital. c. Butterine recently wrote a post info finance talking about prediction markets.

and he credited you at the top as well. People refuted, but yeah.

you going exactly. And so one of the interesting points which he made is that a eyes may become very prominent predictors. They may become very prominent participants in these prediction markets. Because if you can have a lot of a is trying to predict things, well, that lowers the cost tremendously and that opens up the space of possibilities of what you can use prediction markets for. And so the blond chain, you know, is very good for, you know, nobody knows you're so if .

we're going to have .

a lot of a eyes interacting and acting as participants in markets, then the block changes very good for that.

That's absolutely right. And we have a lot of content that's already on this topic, which actually gets at the intersection of cypher N I and where they're match made in heaven. In fact, not only because of A I centralizing tendencies and crypto s decentralizing tendencies, but because of concepts like proof a personhood, being able to, in privacy preserving ways yet even if IT on a public blockchain, find ways of adding attribution.

And there's just so much more that you can do with cyp to agree, alex, and i'm so glad you're brought that up. It's funny because when you are saying earlier that in the early definition of our prediction market has this way to kind of illicit information that's dispersed across many people, I immediately went to like, oh, that's the original agi. If you think about artificial intelligence, this just talk about human intelligence sets scale.

Like that's what a prediction market can be. I do want to make sure we also touch on other applications a little bit on the future. One quick thing though, before we do that.

So now we've summarized some of the key features. We've talked about the election. We've talked about some of the underlying market foundations and some of the new answers. We've talked about what doesn't doesn't make prediction markets work and also mentioned earlier that they are part of a class of mechanisms that can aggregate information. I want to really quickly before we talk about applications in the future, near future, I want to quickly summarize what are some of those other mechanisms that could get at this kind of information aggregation that aren't necessarily prediction markets?

awesome. So first of all, like again, just to think about what is this class information aggregation mechanisms that alex defied IT earlier. It's these are mechanisms ms, that bring together lots of dispersed information to produce like an aggregate statistics or said is statistics that combine the information of many different sources and ideally that that aggregate is informative.

Now there are lots of ways to do that, right? Like some of the symbols, ones we actually talked about earlier, are just to like ask people for their predictions and later pay them based on whether they are correct, right? And you can do that with random people widom with a crowd style, or you can do that with experts, right? And so like very simple types of information agreed, because they are just like incentivize people to tell you what they know, or even just go and survey them, right?

Surveying people like in an unintentional vize context, but where people have no incentive to lie and just like have an opinion where they don't have to do any researcher like invest in any effort to know their version of the answer, you just run to survey. But then you know sort of there's a whole mineria maybe of incentivized ilitch mechanisms that are designed around different listings, ation chAllenges. So I mentioned earlier, peer prediction mechanisms.

These are the mechanisms where you ask people for their beliefs and their beliefs about other people's beliefs, and then you use people's estimate of the population believes to infer, like, whether they were lying to you about what they believe, and or like how informed they were, an aggregate. So if we can use that to figure out where the person fits in the distribution. And pure prediction is like an incentivize version of that is you're going actually to like pay people based on how accurate they are, but you're not paying them based on how accurate they are about what actually happens in the future.

Rather, you're paying them based on, you know how actually they are about the population estimate, right? And so that enables you to pay people up front immediately. These are used for like objective information or sort of like informations disperse among small populations.

Maybe it's not big enough to have a thick prediction market, but people are informed enough that if you can directly incentivize them to tell you the truth, then you can actually like agree the information usefully. A couple of my colleagues in hbs, reshma has some the tire regal and bena have this beautiful paper where they use these peer prediction mechanism. Ms, in the field, in developing country context, where they asked people who in their community is likely to be the most successful microsoft, and then they allocate sort of funding according to these predictions. And IT turns out that, like the predictions are actually quite accurate, like the incentivize peer prediction mechanism, so that produces answers that line up with, like who actually ends up being successful in these businesses down the line in a way that is more effective, say, than just asking people and telling them what we're going to allocate the money according to whatever you said because then people will lie and say, oh, my neighbor, my friend is like, you know the best.

I'll put that paper in the show notes too.

Yeah, a great paper or super funded, very readable, able too.

So one way in which the wisdom of the crowd doesn't work, of course, is when the crowd thinks they know the answer to a problem, but he actually don't.

Oh, okay, of course. yeah.

So there's a great paper by fly like and song and mccoy and they give the example of, suppose you ask people, what's the capital pensylvania and most people, will they go? Well, it's probably filled delpha right is the biggest city, popular city, you know, american heritage livery bell, all that kind of stuff. But IT actually is the wrong answer.

So if you go just by the wisdom of the crowd, you're going to a get filter alpha, and that's wrong. The correct answer is actually harassed, which most people don't know. However, a small minority of people do know the correct answer. So how do you eliminate this? So their mechanism for doing this is what they call the surprisingly popular mechanism.

And what you do is you do, what god says is you ask people, not only what do they think is the correct answers, but what do they think other people will say? And most people, of course, will think, well, I think the correct answer is phildee pha. Other people will say, philadephia.

But then you're gonna see a bump, right? Of haris burg is is going to be very surprising. I could be a substantial number of people say harsh berg, and that will be quite different than what people expect.

And if you choose that, the author show that this can improve on the wisdom of the crowds. So the surprisingly popular answer, the answer which a minority chooses in contrast of the majority, that can actually get you more information out. So depending upon the question, there are these clever ways of pulling this ino hate information out of the crowd and illicit the truth, even when most people in the crowd don't know the truth?

That's fantastic. I've obviously include all these things, reference ing in our show notes .

that is really in that. And then maybe one other piece in the massari. Of course, the visitors this podcast will be very familiar with are simple options where IT options and information education mechanism ms, too.

We talk about Price discovery in an ordinary like sort of very liquid market as being an information aggregation source. But some markets aren't like big and liquid all the time. They don't have like lots of flow transaction.

Maybe it's a super rare piece of art, but an auction is still exactly useful for figuring out what the art is sort of like worth in the eyes of the market. And you can often discover things, right, like there's some artists that was not popular to the best of your knowledge, and then they have a piece without a major sale. And people's estimates of the values of all of their other works change accordingly because of the information that's been revealed about people's change in taste or whatever from this one sale. What we're thinking from things to the show notes, there is an incredible book. Or actually, I think this is my very first a sixteen ee cyp T O booklets contribution called auction, the social construction of value by Charles smith, which talks about actions from a sociological perspective as a way of establishing an understanding of value with a bunch of .

different context. That's great. And by the way, I A plug, the episode ref. Guard and did where we literally dug into auction design for for hours. That was so much fun, so was even like army .

through these different types of mechanisms. It's a really good reminder that the type of question you're asking, the type of market participants you have like this, we're just saying IT shapes your decisions about how to like structure your market mechanism. IT also shape your decisions about what type of market methods to use, right? Like if you think that the population is not super informed on average, but like informed at the second order level, then this mechanism alex was describing as like perfect because the information there is just not like immediately apparently there right when .

I love that you guys are are talking about and we can now segway into some quick discussion of some applications in the future, and then we can wrap up, we've been talking about implications for design throughout this podcast. But I think IT is very interesting because you've been saying throughout both of you that IT really depends on the context in your goals and then you can design accordingly. And that's actually what incentive mechanism design is all about, as I ve learned from you in tim rough garden and seen over and over and over again, but too quick things, just lightning round style that I want to make touch on one multiple times. You both evaluated to this payout fee back loop like i'm inferring from what you've said, the payouts have to be almost quick that you get like an incident feedback loop on your outcomes because you you give an example earlier where if it's like delayed by two weeks or so and so that may be less effective is not necessarily true.

Depends on trust and attention rate. Some people who said that one of their concerns about prediction markets is that people like betting on sports because, you know, it's happening in real time.

You know, the answer within a couple of hours, or in the case of a horse race, within minutes, whether these prediction markets often take months to resolve the final answer or the time of resolution might not even be known, right there might be who will be appointed to this position. So there's possibility that speed is relevant for who chooses to participate in some context, whether they find IT fun. The other context we were talking about is when time matters for trust.

If you are in the developing world trying to figure out how to allocate grants, people might not trust or even just have the infrastructure support to participate in. Mechanism where they're going to be paid six months out is based on the resolution of some confusing outcome, whether if you can pay them today, i'll participate today. Hence why the experiment with peer prediction mechanisms in that context of the first place sort of setting where you could, in principal, pay people based on the outcome, like you know how successful their neighbor was being an entrepreneur with whatever grant the'd received. But a lot of complexity goes into actually doing that in practice because you would have to track down the people again and all of that.

One other quick builder y thing that came up that again seems so obvious to you guys probably。 But the best systems are where the prediction markets and such systems work when there is a disagreement event like an election or something to be resolved, IT probably wouldn't work for some ongoing kind of loosely define non discrete event or so.

The prediction market mechanism to onic prediction market, as we've described, is a mechanism where you're buying like an asset that has a pair as a function of a discrete event. But that is, of course, not even the average case of markets, right? Like you know what, you're buying oil futures or something.

Most of the transactions in many of these markets are actually sort of in the it's on changes in people's estimates. And so if you have a market where you know it's possible to sort of continually update and traded know as estimates change, then like you can still gather a lot of information even if the value attained is in a flow or in stages or something of the sort. IT doesn't have to be sure of a single cut off date.

I think you can design them in different ways. They do have to resolve at a point in time, but the way that they resolve could be based upon a stock Price or something like that.

Yeah and you can have like dividends or something, right? You can have things that pay out over time based on sort of inform steps, like a lots of things have continuous payout based on like the growth of a company or something of the sorts you could imagine, like prediction security .

that are kind of like that.

I the stock market, I actually I stock .

the hp example I gave earlier, divided the time into two months periods, right? So is that made a juniors. July to August and september ked up. So you know you can always take a continuous event and junk in yeah five or six scree periods yeah yeah even .

if somewhat arbitrary, that makes some sense.

So so far, these prediction markets have been used just for what we've been in for predicting something. But you can also create and here i'm gona riff off Robin hanson, again, my colleague, on these questions. And he says we can also create these conditional markets. So the question would be something like, as I said earlier, with few targui, what would happen to, uh, GDP if we put together this science policy now, we might not want to jump all the way from democracy in a few darky yeah in one go. We're probably not ready for that.

We're not ready for the full hand, not quite ready for prime time, I think.

But here's a fascinating idea of Robins, which I think we are ready for, which we should use. And that is what would happen if we fired the CEO. So this is a huge question that companies want to know.

You know, we saw a few years ago, I was kind of remarkable when Steve boma left microsoft and the start Price one way up, you know, suggesting that the market thought that bomber was not a great C. E. O.

Or we just saw, you know, with a brian eckles. He moved to starbucks from five guys, been extremely successful. Add five guys who moved to starbucks on the day that starbuck announced that they were hiring ri nickles as, see, you know, the Price of starbuck jumped up.

So why, however, do we need to wait? How about creating a continuous market, which says, at any given time, would the Price of starbucks be higher if they fired the C. E O? And so you can create these decision markets, prediction markets.

You can a prediction market in. Would the stock Price be higher if we had the same C. E O? Or would the stock Price be higher if we fired the C.

E O? Now that's an incredibly useful piece of information. Yeah, no. Companies, this is billions of dollars every single day are based upon exactly this question. And that's a question which I think decision markets, prediction markets would be really good at answering that will already have the stock market people already invested billions of dollars in exactly this question. And we can make IT more precise and more detailed and more usable.

What I really like about that application is that leverages a type of information that people are already developing, right? right? People are spending a lot of time reasoning about what's going to change the stock Price of starbucks ks.

And they have a lot of different refined ways of doing IT. But he uses IT to address a question that's like useful sort of as a practical hypothetical. As already said, that brings the information forward in time. You Normally in the current market context, we can only learn what happens if starbuck replaces the CEO when they were placed the CEO. But actually, that's like the least important time for us to learn that we actually want to know IT like when they're deciding .

should they replace the CEO yeah exactly you to know you.

And so being able to use that same effort that people are putting into understanding what affects the stock Price of starbucks and like you, which companies are well run in which aren't and like pushing IT towards this question can reveal important information at a time when it's more useful leveraging things. People are already good at predicting that.

that such an interesting and such a useful and extremely real and possible right now thing to do. We're not just being crazy future eristic like ten, fifteen, twenty years from now.

That's so great. Can I be crazy futuristically which yeah.

yeah. We actually a little that good.

You are absolutely right. They should be far as the C. E. O. Market could be implemented right now and I would be extremely useful. And it's the first step towards making more decisions by like down by A A watching consents, right?

I mean, so if you can make a decision that should before the C E O, should we expand into tina or into china, should we have a new model this year, right? You can start asking the market lots of different types of these types of questions. So let's start we should be far to see one of the biggest and most important, most salient of these questions where Scott says it's an information rich environment, people already collecting lot of information on exactly this question. And once we've got some experience in this market, we can start applying IT to further markets .

down the line. Note OK, I love that application and that the importance we talked earlier about, you know, maybe running these markets in like an internal currency, you know, an advantage there is you can use IT to put everyone on the same footing at the outset, right right. Know the starbuck CEO question. There are many different sort of like very high value and high ability to trade entities that already are like participating in the style of question. We're as for a double, you actually might have tremendous inequality in wealth of the participants, but you can make them wealthy in proportion to the reputation or something in the internal token, which can then be used to like you to have them all participate equitably at the entrance to .

these decisions. I love this, and i'm very proud that published a deep body of research across many people, not just our own team. What makes them work? Work what's effective governance mechanisms.

I'm going to link to that in the show notes because also we're arguing that sometimes you can do a lot of these things, not just in the cypher to world, but you can apply them to other deal unities. I want people remember that that's a useful use of downs, which are just do centralized authority organizations. Are there any other pet applications, either current or future issues that either if you have I have one, but i'm going to wait to you guys, you're done.

I mean, two other very quick hits we haven't touched directly yet in the podcast on the idea of markets for private data, right? For like you know, another four of information aggregation is, you know, maybe a lot of people have information that will be useful in designing a new pharmaceutical or medical treatment, and they have their own private information of this form.

And we'd like to be able to illicit from them in a way that also fairly compense them for their participation or something of the sore. And we have some mechanisms. Ms, for this already, like you might have, you surveys managed by a health center, and they pay use of a shop fee for participating in the survey or whatever. But there is a possibility for much richer markets of that form that leverage sort of like individual data ownership, like permissions ing. And so h.

yeah one example, by the way, just concretely, is like in the design movement decentralizing ence people are putting their information, like medical data, using block chains to bring more ownership transparency consent, which they don't have. That's just one example. What's the other one you had, Scott.

The other one, you know, is getting incentivized subjective beliefs. S we talked a lot about like predictions of things that are having objective truth.

But another big frontier for information aggregation is getting really good estimates of things that people believe that are fundamentally subjective, right? And like, you know, if you are trying to do like market research for your product, you to do people want this know one of the advantages of crowd funding, for example, is this a Better information assitant ation medium? Or you could go and ask ten thousand people, do you want to buy this? And some of them might say, yes.

But unless you're actually taking money from them, you don't know whether that's like a truth. Full representation. yeah. And so crowd funding lets you learn about the total market for your sort of initial version of the product in a way that's incentivized more broadly. I think like subjective vitality is like a really important direction to go into.

Can you quickly maybe give a very short definition in the unique ecri pto blockchain context of a beige troop syrian here? Because isn't this where asian truth serum supply? sure. I mean.

the asian trust is actually an example of those pure prediction mechanisms we described. And there are many different versions of IT, but loosely, the idea is, if I ask you your opinion on something, did you like this movie? And then I ask you, what's the likely that, you know, another person I ask, we'll say that they like the movie.

You might have a reason to lie to me about whether you like the movie or not. Oh, I really liked IT because you you produce IT what I do you but you actually hate IT IT. Your estimates of everybody else is believes, will be sort of tilted in the direction of them, mostly disliking IT.

It's so long as i'm going to reward you to proportional to your accuracy that you know that you dislike that and so ever what else probably will too, because your abeyance. And so I can detect looking at everybody else's responses. I can detect whether you sort of like, told me a distribution of other people's believes that's consistent with what you said.

Your belief is great. One of my quick applications and kind of an obvious swing, but I wanted just all IT out because I find IT very boring when people say the same thing. I like oh, media, whatever.

What I find very interesting is people often talk a lot about having mechanisms for quote, finding truth. But sometimes I find to be very pedantic and moralistic and equally as grading as a way that the very people they're trying to bring down. And so it's a pep of mine when I among the twitter discourse like a god, i'm so bored by this. But I do find IT very interesting that some of the commentary surface at prediction markets are basically resolving more accurately and faster than mainstream media, but not having some of the same filtering of part of an interest. I mean, although this might be different with certain communities of downs, if you do predictions limited to certain downs.

yeah again, that depends who's in your market.

Yeah exactly. This gets back to your point about thick then, but it's also interesting because it's a way to put a little bit more skin in the game, which is one of the biggest drawbacks in current media, is like the people writing don't have skin the game, which is why I will always not having third party voices, but the experts write their own posts and then editing them is more interesting to me. So I do think is very interesting to think about this use case of reinventing news media using prediction markets. And and vito's post actually had a great headline, which is that think of a prediction market as a betting site for participants and a news site for everyone else that be my application.

So I think more generally, IT is owne how we do quite a bit of journalism. So for example, it's totally standard practice for a financial journalist, right? But IT to be against company policy. For them to invest in the companies, which they are recommended, right? And as an economist, I kind of think, wait a second, don't want the exact opposite, right?

You want more skin in the game example.

you more skin in the game, right? So you know, I say that a bet is attacks on bullshit, right?

I like that line, a great line. I love you. So.

you know, how about you have to be up front about IT? You have to be honest about IT, transparent about IT. But maybe journalists should say, this is what I think will happen and these are the bets which i've made.

And you can see my bets on chain, right? Yeah, yeah. And let's see what they are. Past track record is right? Like it's kind of amazing that we do not have any track record of opinion editorial alist whatsoever, only text lock.

You know, started to create that and found that they were terrible, right? But how about let's create a series of bets and on chain? And this would, you know, change the types of people who become your editorials who get these jobs in the first place, right? So let's start making sure you bet your beliefs, and then let's promote people whose bets are not to be accurate. And that's going to change journalism entirely if we were to change the metrics by which journalists are. If I waited.

I agree, any do talks a lot about this study is not just, but like an a binary, true, false way. But bets that are waited in terms of likelihood, probability of act, like you don't have to make a binary like IT will be this or that be, I believe, eighty percent that x will happen. And that is also another way to kind of assess in a more nuances way. And that gives a lot of room for the nuance, is that are often true when IT comes to guessing the truth.

Absolutely exactly. There is a big incentive to say this is never gonna happen. This is impossible, right? But then if you ask them, well, if it's never onna happen, are you willing to bet ten dollars that IT might happen?

exactly?

We should all be willing. All, of course I will, will never willing to make those bets.

That's right. Even people who hate elon mass journalists will then starting. Well, actually, i'm gonna guy for building x to happen because I saw that, you know, shuttle launch and now i'm thinking, okay, maybe i'll increase that from ten to twenty percent or whatever.

Yeah, exactly. So bedding could reduce the hypership ah exactly .

yeah totally by the .

order on some other really critical information alita mechanism that uses a different version of the sort of cross examining some people's belief against others. Do you think that community notes on twitter that's an information aggregation mechanism, right? It's like getting a lot of people's opinions and then only deciding that they're correct if you have agreement from people who usually disagree.

Yes, exactly. Because that's where a wikipedia failed when they had the cobble of expert review. They didn't have that kind of jack and baLance .

mechanism totally.

And I have one last question for you guys because we don't have enough time to go into policy in general like some of these became popular because they're offering contracts that we're banned from the market. So big question is whether the offshore crypto markets will follow the rules or not.

So how do you sort of create like innovation, obviously, in that environment? To me? The core question here is what's the difference between gambling and speculation? Is there a difference? And curious ly you guys have, but unless a party note .

on this being a one very important thing to remember is that depending on the context, like you may be in a different point on a continuum, right? Like part of what what makes sporting events like exciting and suspenseful is that there's a lots of to custis ity. And like you know of the amount of information that any individual has is reasonable, small, even if they put a lot of effort into figuring that out. But there might be some amount of like you would have informed betting in sporting events.

And then as you move towards things where there's a lot of information to be had and a lot of like value also to knowing the answer, a lot of market value to actually figuring that out, right? Like how do we allocate ate goods? And markets are going back to the very beginning when we were talking about like the role of markets and, you know, determining the value of something and clearing supply and demand, right? Like there, there is value generated through the process of people engaging.

Now just one, a really important cavy ot about speculation. We talk about this like a lot encrypt ly. And right there is speculation of the form I have beliefs and you know, i'm investing to support the product I think will exist, that I want to exist and that I think other people will want. And then there's also speculation on speculation where you're actually not so much betting based on your own belief. You're betting on you know what you think other people will choose to bet on, like we talk earlier about hurting, you know, you might place bets because you think other people are going to place bets and to give in direction, not because you actually have any information about what's going to happen, just because you have the information about how the market might move.

That's right. That's speculating on speculation.

exactly that, speculating on speculation. So there's this sort of like valuable type of speculation, which is people moving resources around in a way that reflects their beliefs and sort of like can help us make markets work Better and achieve Better outcomes like that sort of in this mid space between the randomness ess, where moving the money around has no impact on outcomes, rather just betting on coin flips like a year money does nothing, and this other edge, where moving the money around becomes sort of its own project that is independent of outcomes.

And so again, like sort of doesn't provide information, right? Like this prediction markets are particularly well. Architect, again, at at least in the cases where they're very large and thick and all the things we talked about that you need to make them work. They're particularly well architecture to try and be in that mid space where the information provided is valuable and comes out of like real knowledge and activity in a way that actually sort of means the market does something valuable.

yeah. And by the way, on the earlier example, when we talk about a lot, the obvious examples. Es, where plays out is like the car lot. A para's framework of like speculation phase followed by installation phase is like a driver of technology cycles. There's also the example burn. Hobart wrote a piece for me a few years ago on how bubbles are actually a good thing when they have a certain type of quality in this case. And he also wrote a new book about IT recently for stripe press with the tobias huber, which they go into greater detail about that what should read that it's basically in an example of quote, I don't want to put moralistic c terms on IT necessarily, but useful speculation that kind of leads to other things as an outcome versus speculating for the sake of speculating, which is partly the .

division you're pointing out. Well, I think people know in last vegas who are at the slot machines they're gambling yeah because they have no way of influencing or of improving their predictions. But the slab machine is gonna show up right is just pure random chance.

On the other hand, there are many, many areas in which we are trying to predict the future and in which investing can help us improve our predictions. And this is why I think prediction red markets should be completely legal, should be legalized because of all the forms of a gambling, of all the forms of speculation. This is one of the most useful forms.

So we want to incentivize the type of speculation or gambling, which as a side product produces, you know, these useful public goods, which is trying to predict the future. This is incredibly important. You think about all of the questions that we have.

You know, what is happening with climate change? Which of these scientific predictions are accurate? Who is with the best candidate for the presidency? All of these questions we have. Prediction markets can help us answer these questions in a way which is more objective, more accurate and more open to everyone. So I think the case for a legalizing this is very, very strong.

That's amazing. I'm going to give you the last word on that, alex. You guys, thank you so much for drain this episode.

That was so fun. Thanks god. Fantastic being here.

Thanks so much. Really fun conversation and Q, V, D.

Thank you for listening to web three three six and see you can find show notes with link to resources, books or papers discussed transcripts and war at a accents eclipsed to do. This episode was produced and edited by sono toxic. That's me.

The episode was technically edited by our audio editor, Justin golden. Credit also to motion t design for the r and all thanks to support for a ecru to follow more of our work and get updates resources from us and from others. Be sure to subscribed to our wet three weekly news letter.

You can find IT on our website at a six and cy cypher dot com. Thank you for listening and for subscribing. Let's go.