Perfect. Thanks, Eugene. That was a great introduction, and thanks for having us sent. And and thanks for joining everyone who who made time to to drop in. Obviously, we know some of you, so hello to everyone who knows us. Quick introductions, and I can share screen. So I'm not your handle or Vaughn. Mackenzie Landell, one of the cofounders of Buzza, joined by my cofounder, Alex or Lejar, who is the technical side of the of the team on the I don't know what I would say. I'm the I'm the the the the words guy. I just say stuff, and he does all the hard work. And we're we've been working, as Eugene said, on governance problems for a while now. We we kind of kind of were a bit obsessed with it in back in 2021. We've just been working ever since on on the idea, and we we've we're, I guess, been deep in research since the work we did earlier this year with AAVE and starting to come out of the weeds now with with some ideas that we think are, I guess, could be useful to the community. We just had to share them. We'd, again, love feedback, love people to engage with with what we're working on. And, yeah, just please do do kind of, like, drop comments in as we go. I'll I'll start showing up. Hopefully, everybody can see this. Looks like it's sharing. Obviously, if you can't, just just yeah. Brilliant. Thank you. So as Eugene said, the title is crypto as he's moving beyond coin voting governance. I think we're really gonna focus on this moving beyond coin voting governance piece. Obviously, for anybody who has read been following Vitalik's blog will know that he has many times spoken about coin voting as being maybe not the most optimum way mechanism for governing DAOs and and protocols. So we really wanted to to tackle that idea head on and and really look at what other mechanisms existed. And so this is what this presentation will be about. The outline is we're gonna start off with with looking at that problem and kind of analyzing it and trying to explain our approach. Then we're going to talk about, obviously, Nathan, who's one of the organizers of MetaGov, wrote a paper about the limits of cryptoeconomics cryptoeconomics as a limitation on governance. So we kinda wanted to address that front because it was because we're literally taking cryptoeconomics and using it to solve governance problems. And then we wanna kind of walk through our approach, talk about these two ideas we introduced. One's called capture resistance, which obviously is not really ours, and then this idea of extractable value, which also isn't ours. We're going to kind of tie those things together. Then we talk about our approach for mechanism design and then go through some application solutions and examples. And with that, we'll kick off. Also, we're gonna be doing some handing back and forth to me and Alex. So just to just to expect that. So coin voting, what's the problem? We we had Vitali give us a really good outline of the problems, but kind of overall, when we think about token voting with respect to DAOs and and even with respect to to kind of l ones, This kind of idea of one token, one vote is great. It kind of you can see as a generalization of the majority rule that governs Bitcoin. Bitcoin's kind of like a consensus mechanism. But instead of using hashes as the resource, so more hashes, more more more voting power, you use tokens as the resource. Right? Seems to work, but the problem is that this means you get a concentration of power and a few large holders who kind of dictate majority decision. Again, maybe that's not a problem, but it definitely means that you could say DAOs or token governed DAOs or coin voting governed DAOs are effectively plutocracy. So they are captured by token holder majority. So this problem of capture comes up a lot. We we kind of see the reverse of this as being things concepts like credibly credible neutrality or legitimacy. Hardness, I think, has even been discussed. We want our DAOs to have certain properties, but it's actually very hard to get a definition of exactly what we want. And terms like credible neutrality and legitimacy are kind of notoriously difficult to measure. So as a team who are kind of focused on the idea of improving DAO governance with respect to addressing the problems with coin voting, we needed something more precise. And that kind of is where we were earlier this year and why we decided to go, into research to try and understand what that could be and to kind of create a much more precise, definition of framework. To just the other side of this, obviously, I think it was, yeah, a while ago now. It was 2022. I think it might have been earlier than that. Nathan published this paper, crypto economics as a limitation on governance. I think it currently lives as a as a medium post. But in the paper, Nathan basically says, hey. There's there's there's I mean, there's some words there on the right, but, really, the the idea is that cryptoeconomics is great, and everyone in crypto is obsessed with with using it. And we've obviously deployed lots of mechanisms, but there are some limitations. And and mostly, the problem being that if you only have an economic lens, then you can only really address economic problems. And the the really difficult, like, side of that is that you end up with a bunch of solutions or a bunch of problems that you can't address. And and for us to have a cohesive system, system, it needs to also address non economic problems. That was the idea. And we basically say, yes. We we agree, and we we actually actually really just want to kind of sit alongside that argument and say, well, we still have, let's say, underdeveloped or poorly optimized cryptoeconomic systems, and there's actually a lot of room for improvement in those systems even before we have to get to, I guess, noneconomic solutions. So that's really our position that we're kind of looking at what's been produced based on the systems we deployed and seeing what might be, still there to improve. So our approach, we first start off with a theory that allows us to kind of very easily explain why coin voting is undesirable. Something that hopefully does something more is is a bit more useful than something like legitimacy or credible neutrality? And then we try to generalize that. So we introduce this capture resistance as a desirable property for DAO DAOs, DAO systems, networks, governance, and then we we attempt to quantify that. So we propose this kind of idea of extractable value. Obviously, we're borrowing from the the MEV literature and saying extractable value is just a measurable quantity of capture resistance. So how much value can an agent, an individual agent extract from a network? And then we say we kind of provide a function for optimizing extractable value, and then we explore applications of that. So with that, I'm gonna hand over to Alex. Obviously, Alex, I will be driving, so Alex will will let me know when he wants me to to skip slides. Over to you, Alex.

Hey. Thanks, Alex. So, yeah, so crypto utilities, obviously, we're introducing these as kind of as a as a class of protocols and services. I mean, interestingly, this is a it's probably quite a fuzzy term at the moment. We're trying to give it some specificity, but the idea is that there are certain types of goods that may require a very high amount of capture resistance. What might those be? So one is services like stablecoins, oracles, and exchanges. So currently, we have stablecoins, oracles, and exchanges with some degree of capture resistance. Over time, we might find that we need more or less, And crypto utilities suggest, I guess, as a as a theory at least, could suggest that maybe maybe stablecoins, maybe like US pegged stablecoins may require a very high degree of of of capture resistance should the, you know, US authorities decide that any any kind of attempt to peg to the USD is is a problem and and maybe make that something that no organization can do without penalty. And so we might require something that's, you know, more more decentralized. It's even harder for anybody to extract any value from and thus make the make a stay make stablecoins much more viable, as an example. Same with oracles. If oracles become an attack surface, you may actually need to make it very, very difficult for anybody to capture any value from from from oracles and and kind of so on and so forth. Another idea is natural monopolies. So one example is and I think Vitalik actually gives this example very, very early on, I think, when he's writing on, I think, from his own blog about real name registries. So if you think about Worldcoin as being, like, maybe a fairly private, fairly shadowy organization that has unknown amounts of of, extractable value and is storing, like, biometric data, the question is where is it storing that data? How is it, storing that data? How might it make use of that data? And so we might actually you might actually want someone to compete with Worldcoin. Now the problem with the competing with Worldcoin is if we have a if we really want to have a set of unique names, then having two registries is absolutely worse than having one. So if we really want to have one, we need to really trust it. It may need to have incredibly high amounts of of capture resistance in order for it to work. Because if it if it if we need World World Coin, for example, to be a single the single provider, we really don't want it to have or anybody to have much ability to extract value from that from that network or from that system. And, obviously, them being on a monopoly gives them a real ability to do that. So here, we might want to have something that's more capture resistant. And finally, you know, even to touch back on the stable coin point, just things that are valuable but maybe aren't approved by every government in the world. So you may have examples of of goods and services that people need and find useful in crypto that any individual country may not want to be able to operate in their in their country. And we've seen this with with things like DAOs, especially DAOs in US, maybe are something that the the regulators don't want to have inside the inside the country and maybe have to be very decentralized in order to be able to operate, especially to issue and operate token token systems or DeFi. So moving on to solutions, there's a question as to what again, what with the framework, what can we produce with this kind of precision, what can we produce? Well, we can start to be fairly specific about the types of solutions to to to limiting or or kind of manipulating or altering extractable value to produce capture resistance. And so we've kind of broken these down into three classes of of solutions. So one is agent based, one is action based, and one is payoff based. And and and you can call them solutions. We can also call them techniques. So one of them is is is we're we're calling it filter aligned. So the idea is you want to a technique that allows you to filter for aligned agents. So if you assume a game of kind of a number of participants, you maybe just want to find agents who are already aligned. And the example I've given her in the picture is like, is like a biker gang where everybody wears the same colors and rides the same bikes. You know that just interest in in in motorcycles or particular brand of motorcycles means that you have a a bunch of aligned agents. And so you can you can kind of assume that you're you kind of have non zero sum payoffs of of kind of working together or or kind of serving those agents. And an example of this is maybe the values of the Ethereum ecosystem, which I guess lots of people share or, like, the the value the the kind of Huddl meme in in Bitcoin, which kind of forces everybody to to act in the same way. So there there there are types of or or, like, another example might be, like, CPU mining or GPU mining. So something that filters for a certain type of agent, which has kind of nonzero sum payoffs. The second one is, like, creating aligned users, so or creating aligned agents. So one is just subsidizing agents. So you pay them, and you kind of guarantee that they're they're never going to have a kind of a there's there's nonzero sum payoffs for them. They're gonna have positive payoffs no matter what. You could say proof of work does this by by kind of producing block rewards for users. And, yeah, the kind of idea here or the the I've used in the picture is bribing children with with sweets to get them to do things that are kind of maybe in their interest, like eat the greens, for for example. And then we call these agent based solutions because, really, their their job is to act on the agent by the filtering for for agents that you that you that have non zero sum payoffs or creating them. The second group class is action based solutions, so looking at the actions that can be taken. So if you think about a set of agents and the action space, the set of actions they can take, there's really two two approaches here. You could you could either reduce the actions or prohibit some actions. Very tricky to do in in lots of regards, especially given you you kind of don't know if there's kind of an open set of permission, a set of potential agents. But, for example, you you know that in Bitcoin, you can't, take certain actions because you have to publish blocks that, anyone can validate. So if you try to do something that's not valid according to the protocol, it's very easy for everybody to prove you're not following the protocol. An example I've used here is, like, erecting a kind of a a gap between fans at a stadium and the actual performer so that there's kind of a everybody can see, and there's kind of, like, a a a reduction of the the risk to the performer, which is good for everybody in the crowd, I guess. And another example of this is Rai. So Rai explicitly is a governance minimized, and the benefit of that is there's a certain set of actions that governance can't take, nobody can take. And for in a lot of regards, the the protocol is bricked. It can't be changed. Useful in some regards because as you can trust the protocol to act in certain ways, But, obviously, we we saw a problem with Rai where they their entire model is based based on the ETH. And so they they don't have a model for for state ETH or for liquid staking tokens, which which means that holding Rai is almost like or producing RI, it means you you have to to take a loss on the return on on LSTs. The another another solution is decentralization of power. So so, importantly, not all solutions are about decentralization. Decentralization is a technique. So this is where you take a high x v action, which is like block proposing, and you split it between a whole bunch of of users. So if there's maybe a 100 worth of value that you can take from a certain action, then you basically average it between a whole bunch of users or split it, distribute it between a whole bunch of agents. And then that action is actually not that it's not really worth someone taking taking that action because it has a lower it's it's better for them to cooperate because they get more value from doing that than than kind of taking the high x v action or percentage of it. And so a good example of this would be proof of stake for the reasons I just discussed. And this last set is payoff based solutions. Again, if you you think about the payoff an an agent can get from any particular action, one is you can counterbalance the payoffs. So you can look at payoffs of of multiple agents, and you just make sure they kind of cancel each other out to make sure they're nonzero sum. And one is and to the example here, these these these two fishes and one being obviously like a a feeder or cleaner fish when the trade off between the two is one is getting kind of, like, clean, the other one's getting food. And the proposal builder separation could be seen as this where you have two agents with slightly different trade offs, and you kind of have one proposing block, someone build building block, one building blocks, one proposing block so that the balance between them means that each one is trying to limit the amount of return the other one is getting or capturing of themselves. And the last one is kind of altering the payoff. So you can also, inside or outside the network, implement fines and subsidies. And we're gonna give you an example of one of these that we've designed, for Uniswap in a second called the cost of fee switch. And this idea of altering payoffs is quite simple. So for us, for a city, you may introduce bus lanes. So you say, hey. You can get through the city quicker if you use the bus lanes, which kind of encourages you to take buses rather than to drive your own car. So that's another class of solutions. Some examples, and I will run through this very quickly. So Uniswap cost of fee switch. So quite interesting that you can think about Uniswap as so this is a payoff based slash, you know, we're gonna alter payoffs. So we can alter the payoffs of a network from even from outside the network. So we we kind of imagine Unisoc governance as a as an entity, and we kind of say this this entity can extract a fee from its users by implementing the protocol fee switch. Obviously, this is something that is within the the action space of governance. And then to kind of introduce a a kind of a technique, we can deploy a contract that that kind of promises to match any fee. It just observes the fee switch, and it matches any fee imposed through a subsidy, which it will pay to users who move to a fork of the protocol. For now, just ignore where the subsidy comes from. But the idea is that for Uniswap, it means that activating the fee switch is unprofitable because any amount that you try to to capture, there's a benefit to you kind of change the the kind of you know, if I'm if I'm getting $10 per, you know, period as a user or as an LP, and then there's a fee switch which takes one unit of that every period. And then I know I can get two units somewhere else by kind of not having to pay the fee plus getting the subsidy of one, then I'm it's kind of more profitable with me to to switch away from from Uniswap. So that's the idea of these kind of outside the network payoffs, and, obviously, we've seen this before with vampire tax on Uniswap. And I guess maybe the interesting part here is this the the the interesting part is this acts as a limitation on the the potential high x v action in Uniswap, which is to maybe extract a fee. So if we think that this the fee is potentially too high or the the the x fee is too high, we can actually implement something from the outside, which almost acts, I think, in in, e economical game theory is called, like, a grim trigger. I if you do something bad, I will, I will punish you, and punish you maybe forever. So that's the idea. And then finally, there's a, there's something we've been working on. It's just as a as a concept, which is the idea of leveraging auctions. So where are auctions useful as a kind of a very favored mechanism in mechanism design to use auctions in to to kind of solve a governance problem? So where could they be useful? Now, admittedly, the the example is fairly complex, but we kind of think we can use auctions kind of to decentralize the power. So think about the actions that governance can take and then distributing them across a market, which is kind of gives it some decentralization. And the the thing that we would we would auction off are are the rights to pass proposals. So rather than having governance vote and saying, like, majority of token holders can pass proposals, we just say we're gonna use a market. So whatever the the highest bidder wins, they win the right to pass a proposal. They can pass or block the proposal. And then the questions we kind of ask ourselves are, can we use actions to decentralize governance power? So maybe yes. Can we use it to optimize resource allocation decisions If, if, auctions are very good at allocating the good to the, to the bidder with the highest valuation for them, then what we get is someone who, who, who's willing to pay the most, who who believes that they can, they get they're going to find the best use of the the governance rights whether to pass or or to obviously block the decision. Could they have some use there? Obviously, can we use this to reduce the extractable value that governance has? Because, technically, if you want to extract some value, you have to pay for it. And fund development. So it could be used as to actually just fund development on an ongoing basis rather than doing that in a from a central way from a from a treasury, which we could maybe even remove and replace with this mechanism. It could be fund ongoing development. And then finally, could we use it as a mechanism to capture kind of fees from this auction and use that to fund public goods on an ongoing basis? So just some things to think about. Obviously, we do have this example. I'm happy to walk people through a prototype if they think that's useful. And, that's, again, just one of the ideas we're worth, man. So, that's mostly it. Obviously, Alex and I are on on Twitter. You can find us there. We also have a website, forum, which we've just started. We're starting to kind of post some of our secondary research and literature review there. Some of our ideas will be there. We also have a Discord, which you're welcome to join, and, obviously, you can email us. So, yeah, thanks for listening, guys, and, obviously, open for questions.

Yeah. I I can I can provide a quick answer to that? So if you think about rage quit as an idea, I think it builds on this, like, concept that if anybody is is seen to be extracting too much value in a way that you disagree with, you can you can extract your you can just remove your assets. And I think that's it it acts as a limitation on extractable value. Right? But and so I think it it it's nice because it has, like it kind of reminds me of the concepts of subsidiarity. So you can it's it's up to you whether you can remove your kind of proportional assets from a from a kind of from a pool of common resources, and that acts as a check on people doing things that look like they might be an overreach or excessive in terms of in terms of extractable value. So it kind of acts to enforce enforce capture resistance. I guess my problem the the problem, I think, is that it also creates some instability because you you kind of can never really expect to have a a continuity in terms of the resources available to that group or to that network. And it it kind of ends up with with the users of the or, like, participants in that network almost being more PVP than acting as a as a common group because they they they always see the assets as being their assets, which they have kind of an individual claim to.

So

I think there's some trade offs with with the kind of rage quit, and I think that's probably why they it makes a lot of sense in some networks and maybe less in others where, you know, they they don't they feel like the resources have to be pulled and the individual claims are less valuable.

Hi, guys. It's

Yeah. I thought I thought it was me.

Okay. Thanks. I really appreciate the the survey of this. The thing that I think I like the most is the the consideration of these in in analog fashion where the tendency seems to say, let's, run a simulation of, like, a billion or trillion times and see how it works. And, I like the the considered approach to design. So so thank you. I'd love to know if you have any thoughts that the way you talked about protection, of of things and then oracles, specifically really was interesting to me. I think of, humans as an oracle, both individually, collectively as republics, maybe even holistically, you know, if you try and take a pulse of a species. You know, and protection of that oracle within the governance function has to be absolute. You know, like, we can't stop breathing for a period of time and then choose to start, you know, start breathing again for the the constant, right, that has to move through. How I guess, have you given any thought to kind of more to that protection function as you see it? Because I really like the way you disambiguate, like, the protection of the thing and then maybe, like, the projection of the thing, which is what we kinda think of as, like, money. Right? Like, where it's okay to use these sort of, like, abstract mechanisms because, you know, the consequences are less severe. Off mute.

Yeah. I am. Hey, Mel. It's interesting question. I don't know if you have a a direct answer, Alex, but I think that the maybe maybe it would be useful if you kind of gave me a an example of the kind of protection versus projection so I can respond.

May maybe it ties a bit to, like, the Mala question a little bit. I or or at least my my head kinda the question changed a bit while I was listening to that previous answer, so I'll I'll kinda hit it. But, you know, when you talk about the things that are, I guess, like, human identity, if we agree that one registry of who is human you know, who the humans are and that you you're one human and not projecting as two, it almost gets to this this other kind of basic corollary, I guess, where it's like, that protection of the human is may maybe part of the, you know, kind of the the function of governance. And this is maybe a bit unrelated to the broader survey that you're you're kinda giving. But I guess in my mind, there's a spectral sort of, consideration of what the protection needs to be. And then on the flip side, it seems, you're discussing the projection functions. Right? And maybe giving them some some other characteristics that are like, okay. Here. We're we're now making some distinctions between mechanism types, which ones are, you know, maybe more well suited to these types of things. And when you get into that sort of, I guess, approach, you know, I I agree. Like, I tend to agree that there are the things where there should be a monopoly. But maybe, like, maybe identity of humans shouldn't be that thing if that one thing could then, you know, cascade fail into alright. Well, we all put our trust in this one thing, and now that failed. Like and now we as a species are again kinda back in this, you know, zero sort of that recursive loop. I'd love I just I love that you're I love that you're doing it. Like, that's the thing. It's like, it kinda breaks my brain a little bit that you're starting to classify these things. And or it's it's an extension, right, of classifying mechanisms towards proper use and beneficial use to humans. But it kinda gets me back to that, like, what's the base? You know, what does that rest upon? And so, so, like, if you have any thoughts on that, awesome. Otherwise, I, I thought this might be a bit of just banter, but, it's a really good presentation, and I'm gonna, spend some time reviewing that. Thanks.

Thanks, Moe. I I actually feel like I do have a a response there. I don't know if it's a non answer, but it's certainly a response. I think, Alex and I spoke about this for quite a while. The the limitation and, actually, it's something that, is kind of hinted at in Nathan's piece, which is, like, we if we if we do have an economic lens, you know, we we it says, like, we need to envelope economics within a logical politics capable of seeing beyond economic metrics for human flourishing and the common good. So if we just have this economic lens, then we we actually lose some some we're we're losing some information, I guess. So if we were thinking about it, and then and then what, you know, the projection side of of what you're talking about is the is the only thing that we can see. Like, it's the kind of allegory of the cave, but, like, we can only see the shadows. We don't actually know what's casting them. And I think that's it's a limitation, I think, of of the of the approach. And I think that one one concept you have in in in in in economics, I guess, so in in at least in game theory, mechanism design is a side of transferable utility. So everything we talk about has to be something that is transferable. So it needs to be able to be converted into a into some maybe, like, a dollar amount or something that can be transferred from one agent to another, which again loses some some resolution. So I think that, ultimately, we can the our our kind of response to this was maybe these are these are complex systems. Humans are complex. Information problems are complex, but this is the best we can do. And as I said here, we're just trying to say, well, let's look at the systems we have and just see how we can improve them using the tools we have. And that's, you know, that's that's all we really seek to do. I don't think we can we can do I hope we can do much better than that, but but that's really the the kind of limitations that we're putting on ourselves.

I appreciate that. Thank you for speaking to it. Almost like you have to scope it by the limits to get to move in any direction. So I this is a really good a good survey of that. So I really

Right. Chezmo.

Yeah. In in in, at an extreme level, yes. That's exactly it.

One person, one vote?

So so when so I guess maybe the way of thinking about this is saying, I think that many people have considered one person, one vote, and also people have considered the downsides of those systems. Right? Like, that one person, one vote systems. But let's say that they are they're they're useful. What I would say is instead of considering one person, one vote, we're just trying to look at what does it say about what can we say about one person vote systems in using the framework of extractable value. Right? And why why might it be useful? So I think let me try and remember. I think the critique of one person on both systems in a digital setting is is, like, they're not is the civil resistance. Right? It's like the civil attacks. And so you have to spend some some money, some amount proving that someone is an individual person. And then you can say something about this concept of an individual person should have one vote equal to another individual person in a nation because both have arguably similar, upsides and downsides. So they they have kind of a, there's only really a certain amount of, value that an individual person can extract that's more than someone else from a nation, right, their their ability to inhabit some land. But in digital systems, it's completely different. So I guess, like, the idea of a person with regard to voting on what grants to fund, you may not want to give everybody the same amount of ability to, like, extract value. You may have very different types of agents in that system. So I guess that's that's probably why one person, one vote isn't a turnkey solution to to token governance because it's not always that the the system sees every individual person as equal. I'm actually sure

But I

I'm I'm answering all the questions. I don't know if Alex has an answer, but if not, I will I can jump in.

Yeah. Actually, yeah, it's a very relevant example.

Yeah. That that's very good question. We don't have an answer for that. It's a toy, the this proposal market idea, but collusion is obviously one of the the main problems. The solution would probably be you have to you have to have some kind of restricted set for one of them. And I guess this is maybe similar to the proposal builder questions where you need to make sure the proposals aren't builders and and vice versa. So I guess we can learn a lot from just watching how that's being developed. But, yeah, I think collusion's a massive problem, to be honest. It's yeah. If if you conclude, it's it's you so for example, you can be both sides of this market, so you can pay yourself to to to pass proposals. So it kind of breaks. But, yeah, it's a it's a toy. It's supposed to just explore mechanisms, which is very much this the kind of work we're trying to do at the moment is just explore as many mechanisms as possible and, try and figure out whether they are useful or not.

Alright. Thanks, Eugene. I appreciate you too. Thanks, everyone. Thanks, everyone.