Protocol Alignment: Governing like a Protocol by noturhandle | Devcon SEA

[Music] thank you everyone perfect um I am not your handle also known as Vaughn um going to be talking about protocol alignment governing like a protocol uh which sounds really really strange but trust me it will make sense in a second um first of all I'm going to start out by talking about why we need governance I think most people think they understand what governance is I'm pretty sure in the context of crypto they don't um so here's at least our picture of governance so um thinking about governing a piece of software is fairly strange so what I like to think about is governing something people kind of have a good model for which is a vending machine this is actually an incredible picture somewhere in Hokkaido uh has done the rounds on Instagram I love it um but you can think about blockchain protocol as being kind of like a vending machine it's fairly simple it defines some rules you can interact with and it does something every single time that you try and interact with it and if you think about the the um box the box is kind of like a set of rules um and they just specify how anybody can interact and the value prop at least from my perspective is that Protocols are great because they make the future more predictable nobody can come to a vending machine and do anything but put money in and get products out at at a predetermined price now governance the job of governance is to extend this we need a set of rules that specify how we can also make changes to that protocol if we could make any change we want that means the protocol would no longer be able to give us any certainty about the future the future would become less predictable the value prop is that protocol governance also makes the future more predictable if you have bad governance then the future becomes less predictable and now we have governance failure if you think about a vending machine a vending vending machine's job is to give you exactly what you want and if you go to a vending machine and no one's been able to update it right as in put more products in you'll find that the vending machine no longer does the job you expected to do the vending machine slowly starts to lose its value and again you no longer interact with the vending machine or protocol so we need protocol governance because otherwise our protocols would no longer get users you can think about the action space of all the things you could do in a protocol right or with a protocol or in life and then you can think about a set of actions being aligned with the goal of a protocol right updating the products in a vending machine is aligned with the goal of the vending machine punching a hole through the front of a vending machine not aligned so you can think about governance as being that ring it's the it's the rules you put in place to govern the set of aligned actions allowed to occur in a protocol so governing like a protocol from my perspective is really governing four buckets of actions if you think about protocol it's just being a set of functions that you can call right one for UNIS swap could be something like swap or add liquidity and then another function which is just how you update those functions and that's what we call the governance and the governance really governs protocol upgrades making changes to the code itself adding parameters or changing parameters which is changing parts or variables in the code um policy updates so anything that's like soft governance any rules and processes we have around the protocol itself and then of course resource allocation another way of thinking about governance is it's a set of State transitions between different versions of the protocol so you can see participants sending messages in they're voting they're putting in proposals and over time we're agreeing which versions of the governance to change so upgrading from V1 to V2 or indeed if we change a function having the first version of protocol and changing it to the second the third the fifth and so on now what I wanted to do is very quickly run through some ideas they're going to be fairly rapid fire we can tackle them in the questions if that's interesting um we've at butter have been looking at loads of different types of governance systems now one of the ones that we have already is token governance we like to vote on proposals we our tokens to provide some information about our preferences And We Gather those all up and we make decisions great this is a vote that took place very recently for tally we all know tally we love tally Tally wanted to put a proposal into Unis SW swap that basically ask for like some some resource to be allocated to them and this passed so when this passes we say like there's three 38.8 million uni voted for this we have to ask ourselves what happened and one of the things I want to get into your mind today is that the way that we vote is based on our preferences which is also based on our incentives and most people in the world are dangerously long themselves most people are self-interested they care about the maximum return to themselves so if you think about those people turning up to vote this is how you have to think about them and as I said a token holder vote is a token weighted average of stakeholder interests so that's what you're getting you're getting this kind of as an average a token weighted average some other mechanisms you could use include auctions you could for example auction off proposals I wouldn't encourage you to read the text I put on the slide I just put it there because I thought it was interesting but if you look at that I think most people understand how auctions work you have a seller they're selling something people are bidding for it so imagine bidding for a proposal we actually decided to design one of these a proposal Market um and uh turns out people didn't like the idea of bidding on proposals but you can imagine the idea that you're bidding to buy the right to pass a proposal and you can imagine all the ramifications of doing that um and come and speak to me if you'd like to Lear more about how we did that um another one is peer prediction this is where you have lots of people observe something and then decide whether they think it's aligned or not aligned with the goals um you could say that token voting is kind of similar to this but this could be any kind of mechanism you could design you can think about Uber ratings being a peer prediction mechanism and again we tried to do one of these things we ran it for element and it was uh it was interesting again we have some good results which we published online um and this is all to say there are loads of other ways of of actually getting people to do this role of deciding which which actions are aligned with the goals of a protocol one of the things that we've been really excited about recently is futaki or decision markets they're a mechanism that leverages speculative markets to make decisions I like to say you have to think about it as this you have four possible worlds in the future one where you make decision A B C or D what you can have people do is speculate on the value of you sorry speculate on the probability that you achieve the outcome you want in any of these worlds and you can use the probability to determine which world that you want to be in and you can use that to actually make a decision this works interestingly well um it's something that we're trying out we'll have some uh some trials actually going live very very soon maybe in the next couple of months um and um you could imagine a market for making a decision on who to hire as a uh as a prediction market so you can predict who would get the result you want which is the amount of Engagement with a particular proposal or with a particular piece of work and then you can sell the uh potential outcomes as tokens and people can bid on them and they'll give you the probability that each of these events will happen and then you can use that to actually make the decision on who to hire so a futak uh eli5 is if markets believe an outcome given decision a is more valuable than an outcome given decision B choose decision a and ideally decision a will be very aligned with the goal so what we like to do in governance is build a system where being long yourself is the same as being long a particular protocol that means you're aligned and is the same as being long with ideally an entire ecosystem this is an example of the same proposal which was the one that we put through a vote with Tally but instead we're now asking about how much each project so we have tally versus other potential places we could invest would grow voter retention which is maybe the metric we care about we have people bid on these token tokens and they produce either the uh they they produce an a proposal or sorry they produce a outcome that we care about and then we can make a decision uh I'm vaugh McKenzie from butter and uh please come and speak to us about fak and decision markets thank [Applause] you thank you and let's take some questions from the audience uh we can pass around the mic anyone has a question for this [Music] speaker oh we do have a question over here oh uh we do have a question here on the the first rle hello uh my question is how I I don't know if I is I'm going to say well how fi works on on those absolutely um so a DA could decide that it cares about something such as sequencer Revenue so for example and it could then decide to ask a question um he then decide to ask a question um about whether a particular decision say if they choose to invest in a particular protocol or if they choose to um roll out a particular program or make some decision if that's more likely to increase sequence of Revenue and then you can compare the probabilities and you can make a decision based on which the markets believe is most likely