Robin Hanson

A one-day summit focusing on the theme of d/acc: emphasizing the values of decentralization, democracy, differential accelerated progress, and defensive tech including crypto security, public epistemics, bio defense,  neurotech/longevity, decentralized ai and physical resilience.

Breakout 3

d/acc closing panel

Conditional markets allow predicting outcomes from potential decisions, enabling what is called futarchy governance, but key design questions remain open. We'll examine specific challenges: aligning founders with investors in protocols, encouraging meaningful participation in decentralized governance, and integrating futarchy modules into existing governance systems.

Stage 1

Designing Conditional Markets and Futarchy

Prediction market panel

 Hello everyone and can you hear me? Awesome. I'm Vaughan Mackenzie-Landell. I'm one of the co-founders of Butter. We've been working in the governance space for a very long time and Obviously very excited about futarki and prediction markets and that's why we're all here We have this incredible panel full of lots of people who my love and admire And I'm gonna just hand over to each of them to give a very short introduction and then we'll kick off Yeah, I'm Martin. So in a previous live venosis, we started with prediction markets and Wanted to implement food hierarchy and other things. So in general, I'm absolutely a super big fan of that, but I think in this panel I will a little bit play the skeptic here because I think as Robin mentioned, you have the idea and then you have all the details that need to be figured out. So I think I can a little bit bring in that perspective. I just talked, so I'm Robin Hanson. So I'm Calvin, also CAS.if. I'm the co-founder of Utarki.fi. We just announced our project a few days ago. Our plan is to help DAOs and other organizations make great, impactful, positive decisions. Hi, everyone. I'm CJ from Limitless. We're building, like, basically we're processing around a million dollars a day in bets right now for, like, zero-day contracts on financial markets. I believe that believe that like with crypto, we can build the world's largest economy on chain outside the jurisdiction of any one particular nation state. And I think that as well as that, we need to build really efficient global marketplaces on top of the infrastructure. And so that's what we're doing at Limitless Labs. Thanks, guys. Really appreciate that. So we're here at DIAC today. And so the first question is, how do prediction markets tied to DIAC, how do prediction markets and epistemic tech make the biggest impact in the vein of DIAC on differential progress, defense, and democracy? Over to any one of the panelists. One ring to rule them all. If you could have reliable prediction markets on whatever policy question you have, then you can just do better on all of your other areas of life. You could better know how to pursue longevity, better know how to pursue decentralization. Whatever it is you're trying to do, having better information on that can make you do that better. So could we do much better? And how does that map directly to, say for example, defense? I can just replicate that. I think the promise is very clear. So the promise is whether we we talk about Even simple things like this community node or so from micro decisions whether to show that community node or not to show that to Extremely large macro macro decisions should we raise the interest rate or should we make this huge? Well defense spending here or there. In general, the promise is that this is a tool for more robust decision-making to bring in, yeah, to have a better quality of bringing information that actually shows the true impact of a decision. Yeah, so I'm going to make a comment here. So over the last few weeks, everybody has seen a lot of people are talking about prediction markets, of course predicting who's going to win the election, and this generated a lot of excitement. But actually, the most interesting thing is not only predicting probabilities, that's what normal prediction markets do, but predicting the consequences of things. So if you could have, for instance, conditional markets on inflation, on GDP growth, on spending, on unemployment, conditional on candidates, right? This would be even much more interesting for democracy. You could actually see like a scorecard for the candidates, right? Like a game, you can pick one. This one's better on this one. This one's better on this other thing, right? So this is the promise of using conditional markets for decision-making. Yeah. Sorry, Monik, go ahead. Maybe let me throw in one kind of concept to challenge this idea. So to say, I mean, I'm just pulling up here CoinGecko and look at prices of crypto. And, well, they change by 6%, 7% a day. So I think prediction markets rely on the assumption that markets are so efficient that out of those prices, we can get reliable information. So I think prediction markets rely on the assumption that markets are so efficient that out of those prices we can get reliable information or better information. But I would argue well there is also a lot of noise in markets and I would say that markets here, prices change by six percent is largely not related to any meaningful signal, but is... Well, I need to be careful in the statement, but there's still a lot of noise. So that's, I guess, my question. How can... Is there enough signal in the noise? So when you compare head to head this mechanism to other ones, consistently this wins. So that means all those other mechanisms have more than 6% mistake. Think about any committee you've ever been on. Think about the gossip network. Do you really think that's within 6% of the truth? Come on. Like being 6% close is great compared to what we usually have. Yeah, I would, I guess I would come with the, like the whole question from, from a little bit of like a different angle. I think that definitely prediction markets have a lot of power in terms of forecasting and decision making. But I think in a crypto context, what's really interesting and powerful is the fact that essentially we can build these global efficient marketplaces. And so I really believe that markets in general accelerate human progress. If it wasn't for markets, we wouldn't be where we were today in almost every field. And even they have positive trickle-down effects to science funding, for example. Brian Armstrong sells his Coinbase stock to fund moonshots. And science funding, that's a trivial example in a global context, but still kind of proves the point. And I think that building prediction markets onchain can be defensive in the sense that we actually have a capital formation and wealth generation outside of the nation-state in this kind of global on-chain economy and I think why that's defensive is because we can build this kind of resilient independent decentralized systems that manage to kind of form their own capital, like without, for example, being taxed on it, also without the very heavy regulation, which is a big issue. There are only 16 designated contract makers in the United States. There are 16 license derivatives exchanges, but money cannot move globally at the speed of info in the traditional system. Here it can, and we can build these really powerful marketplaces that are global and accelerate human progress. I completely agree. I think markets are a very important piece of technology. I actually do believe that one of the reasons that we are able to accelerate defense, particularly with prediction markets, is because we're basically getting rid of any authority's ability to lie to us, right? Because we can delegate information and truth to markets. I think this is great. And actually, speaking about this, if you think about Futaki so far, as you said, we've had lots of messy details. I'd love to know what you think the key challenges in implementing Futaki will be for a world organizational decision making. Right. So some challenges, of course. Robin mentioned the matter of the autist in the boardroom. So there's always this risk that prediction markets and also decision markets are, in some sense, threatening to the power of insiders and some organizations that maybe are not making the most efficient decisions. So this is for sure a big challenge. So another challenge that is marching up, I'm going to try to connect the previous question from Martin to the combinatorial market idea from Robin, which is part of our basically our grand vision on fitarki, which is so the matter of noise When we're evaluating potentially a small or medium-sized proposal or decision in a company Many of those if there's a smaller medium, you're not gonna have a clear impact, visible impact on the share price, so that makes it really hard to use this directly to estimate, but if you've had a full network or a full tree where you can see, like you could have markets for share price or token price, but connected to these you also see the KPIs, markets for different KPIs, number of users, revenue, all the kinds of metrics important for the company. So you could actually see, okay, maybe I don't, there's a lot of noise, I don't see the impact of this decision on the token price, but I see the impact of this decision on the token price, but I see the impact of this decision on this minor KPI that has, but it's minor KPI associated with this major KPI, which isn't then related to the share price. So I think that's a, this is a big trend region for many decisions is actually have the markets, all the information connected to each other. Maybe to kind of reorient the question, what problems are you running into building and operating these things? Prediction markets, futakis, what are the hard problems, what are the solutions? I can say one hard problem, relatively hard problem is to just find traders that are comfortable going into those positions. And just to give you an example of, let's say we would do this for Ethereum. We would say, okay, conditional, we might go this roadmap or that roadmap and then we want to use this food mechanism, so essentially people can say, conditionally to this roadmap, I buy Ether and conditional to that I sell. But then to practically do that you need to believe that going back to this correctness of prices, I mean yes 6% is in a day but I mean we really, what is really the fair price of ether right now? And I think there could be a range of some people believe it should be 10,000 and some people would believe it could be less than 1,000. So if you are in the camp of it should be 10,000, then kind of probably in both decisions, you would say, well, I want to be long on Ether, I want to hold Ether and if you're in the camp of well this is really not worth it, it's just worth 10,000 and again in both sides, so you need that person that says well the price is pretty much right now and really that decision makes a marginal difference for me to hold the asset. And then let's say there's a really good proposal and you really like that, so you go long on that. Now you hold this conditional yes. Now the next day, some external completely unrelated event happens where the whole crypto market crashes by 50%. You hold this yes proposal and hold it under the price from yesterday, you still think the proposal is good, but now your incentives are for the proposal absolutely to not pass because then you would hold all the dollars and not the asset at a price that's now 50% or where the real asset is now 50% cheaper from what you bought it. So I'm not sure if anyone could follow, but those are kind of the challenges I see. So 20 years ago we had a burst of applications of prediction markets, including in corporations. So I think we saw typical failure modes there that are instructive about today. Typically if you went to work with a company about setting a prediction market, you would say, well, what are your most important issues and numbers you'd like to track, and let's set up markets on those, and they would usually say, well, that's a little sensitive. Let's do something a little safer, and they would pick safer topics, and then they'd get accurate estimates on them and satisfied users, but they go, yeah, but we didn't really care about those. So, you know, that would be one outcome. Or, you know, they would say, talk about something important, and then management couldn't resist having opinions that disagreed with the market, and the market gets proven right, and the management wrong, and they're just really mad and want to kill it. So, for example, the US government missile test agency had prediction markets on which tests would actually go forward. So they put up, they try to make a lot of tests, and a lot of them don't actually happen, because they have to coordinate a lot of different parts of the military, make a test happen, and if one part isn't there, the whole thing has to be scrapped, and a lot of money is lost. And so they wanted to know which tests would actually go through it and maybe save money by cutting back earlier but then it was more legible that they knew ahead of time that it wasn't going to work and that was a management problem so they didn't want that anymore they'd rather pretend they don't know how these things are failing yeah uh like i guess i'm gonna like talk about this from more like consumer product building perspective, less like the future side. And I think definitely like historically, it's absolutely true that marketplaces are like very hard like products to bootstrap and to activate like the atomic network effect. I think that like it's definitely like to incentivize like market makers or liquidity providers, especially for things like pop culture markets, right? I mean, you can always build the model for, for like elections or sports markets, or especially in limitless cases like the financial markets, right? You can use the volatility to understand how to price them. And so it can be attractive to institutional market makers because it's essentially like a short term retail retail option flow but at the same time you know you may look at Polymarket how much they spend per month to incentivize market makers to come like into their order book actually I spent a lot of time previously and like I spoke to Robin about it in Berkeley like kind of obsessing about incremental improvements in the AMM which we see like for example like paradigm just launched the PMAM, right? We're actually going to spin up a contract for it to see how much of an improvement is it. But I definitely think that challenge is like, how do you incentivize the market makers? Either you use cash or you use your token or what, but it's still a huge expense for bootstrapping like the initial network, yeah. But maybe over time it will be worth it. All right, brilliant.

https://streameth-develop.ams3.digitaloceanspaces.com/transcriptions/673595919dbb7a90e13b23f7.vtt

35 years ago I began outlining a vision of how betting markets could offer informed credibly-neutral estimates on far more disputed topics. I elaborated 25 years ago on how decision markets could support neutral governance, and 21 years ago on how combinatorial markets allow estimates on all possible combinations for existing topics. Now in the last year, we are seeing substantial crypto-based trials, especially re governance. In this talk, I’ll paint a picture of where all this could go.