Panel I The Ethereum Governance Process with Christine Kim | Pragma New York 2023

[Music] I want to bring on an introduce Christine from Galaxy to talk about the ethereum governance process in a nutshell a lot of you have kind of asked us questions about how does the protocol get better what do the devs talk about what is a core Dev how do we vote on things what Zoom meeting do I need to join all those questions that we usually get most of them will be answered today and uh and doing that will be Christine so please give her a massive Round of Applause hello thank you so much for the introduction um I'm really excited to be here and talking to you guys about ethereum governance a quick disclaimer before I start the presentation um all of what I'm about to say is just for informational purposes not for investment advice or legal advice or anything like that um yes so the agenda today is that I'm going to be talking about the people involved in ethereum's governance the forums on which these these decisions are made prior upgrades in ethereum's history and some ones upcoming like Cancun denb um Prague Electra we'll see how much content we can get through before this half hour is up and to start I really want to begin this discussion by talking about the Bitcoin governance Triumph ret because if this is a topic that you're interested in which I hope if you're not already by the end of this presentation you will be this is a very kind of Infamous um diagram Infamous image that has shaped the discourse around governance of decentralized permissionless protocols so on bitcoin the governance of ethereum the way that the Bitcoin protocol is upgraded is primarily um governed by these three distinct groups of people you've got developers Miners and nodes developers pretty self-explanatory they are you know the people who have a very strong knowledge of the Bitcoin code base they are the people who are um technically minded and know how to upgrade the code can propose different changes to the Bitcoin protocol um and develop the clients so the the software that people run to connect to the Bitcoin blockchain um so it's like a it's it's like a group of people that have a very distinct specialized skill set um miners very similarly have a very specialized distinct sort of characteristics miners are people who have a lot of money to spend on machines and can spend a ton of electricity you know devoted to mining the Bitcoin blockchain um not everyone can be a minor not you can't just like spin up your computer and be like oh I want to mine on the Bitcoin blockchain no it's actually very very the difficulty the mining difficulty of of Bitcoin right now is extremely high so miners it's mainly like professional miners as a company or if you're going to be a mining pool you know um developing software these are all very specialized skill sets there's not a ton of overlap between developers protocol Developers for Bitcoin and Miners and then you get to the third group of people nodes um that I think is very special it in like the governance conversation around Bitcoin because if you think about it developers run Bitcoin nodes miners run Bitcoin nodes so when we're talking about nodes who are we really talking about here and what characteristics define them the way that I Define it is that nodes are basically the users who are not developers and who are not miners who run Bitcoin software who run Bitcoin clients so you could think about exchanges custodians that're all running Bitcoin nodes for different purposes but the most important type of of people in this node bucket are independent users just people who don't have a financial incentive to run Bitcoin nodes but simply want to connect to the Bitcoin blockchain want to verify the code being run on bitcoin want to contribute to the decentralization of Bitcoin the ethos of Bitcoin and so to really ensure that this Bitcoin this Bitcoin yeah governance trium it works Bitcoin developers have worked very hard to make running Bitcoin nodes extremely easy that's one of the things that was Central to the block size Wars um that Infamous for between Bitcoin and Bitcoin cash um run having bigger blocks would make it more difficult create a higher barrier to entry for users to run nodes um and that's why you know Bitcoin is designed the way that it is there are certain parts of its protocol that uh Bitcoin developers are very hesitant about changing um so you can see how the governance of Bitcoin coin really does also shape the protocol of of Bitcoin of the way that Bitcoin is developed and prevents certain changes from getting through um so yeah so that's kind of Bitcoin governance in a nutshell and it's a good starting point to contrast and to highlight how ethereum governance is extremely different and how the this that that triangle that setup doesn't can't really be applied to ethereum um for several reasons ethereum as a technology is very different from Bitcoin um it's a lot more complex and ethereum the ethos is to try and make participating in ethereum governance as a developer as a validator as a node operator as easy to to participate in as possible and it's and it's really hard kind of to to define the characteristics that distinguish these three groups if you try and Define it by you know what we just talked about the Bitcoin governance Triumph RIT so let me dive a little bit deeper into that and explain a little bit more in detail why it's really hard so developers on ethereum you have an app player which Bitcoin doesn't really have um and so as much as protocol developers do Define you know how the ethereum protocol gets changed a a lot of those decisions are informed by application developers by people who are developing defi apps for people who are developing nfts they're the ones who are interacting with the smart contract language solidity they're the ones suggesting certain op codes different improvements to the evm and so you don't have just this specialized set of people who are very familiar with the ethereum protocol um like the execution layer the consensus layer but you've got a ton of people who are just using ethereum as an application developer a smart contract coder that's also influencing decisions heavily so like the the range and the scope of voices is is much wider um and then of course you don't have miners on ethereum anymore you have validators um and if you think about the whole reason why we mve to proof a stake it's to open up the amount of people who could participate in security right you're not you don't have to have a major business or like have a ton of capital to participate in staking anyone can be a validator is you know supposedly the ethos um but in reality a lot of the people who are participating in governance in this way by being a validator are actually people who just have some eat lying around they don't necessarily have to have 32 eat maybe they just have two eat in their pocket and they want to earn some rewards on it um so when you think about validators and you think about it kind of the way it practically shakes down you can Define it and differentiate it by people who are stinking I just have some eth lying around that they want yield on and validator note operators so the people who are operating um operating running ethereum client software for the purposes of for business purposes to try and make a business out of staking to take people's eth stake it for them um and fulfill the responsibilities of being a validator um so we go now to the final bucket the nodes and and on bitcoin of course we we were we Define it very clearly as the people who are running um nodes independently not because they're a Bitcoin developer not because they're a minor um but simply for the purposes of of connecting to the Bitcoin blockchains to you know um contribute to the ethos of decentralization and we got to ask ourselves you know are those primarily the people if we think about ethereum people who are not developers who are not validator node operators is that group of people a very large group of people in ethereum and I'd argue no not really most of the people that are running ethereum nodes are the validator node operators or a lot of people who are running nodes are doing it for a business as a business to encourage users to connect to their nodes because running a full node on ethereum is considerably harder and requires more computational resources so you have the blockchain infrastructure providers people like infira who are running and operating nodes on us behalf um so when we're thinking about who are the who's the third arm if there had to be one I think it's it's not as easy to Define as just like independent independent users who are running ethereum software you have many businesses that have kind of capitalized on ensuring that it's easy for users to connect to ethereum to stake on ethereum without running their own software and um yeah and have and and be able to do that service for them um yes so this diagram kind of is trying to show that governance on ethereum is not easy to Define in that trilemma sort of way the next couple slides in this next section of the presentation is the way that I Define governance on ethereum and the four key actors that I think are the most important in ethereum's governance and names that as you kind of cover or as you kind of like continue to follow protocol developments on ethereum you'll find these names and these groups of people and their interests popping up um frequently um so yes first and foremost we have the ethereum foundation um the ethereum foundation is plays a major role in governance um and for different reasons over time initially when the foundation was created ated um the foundation was allocated a very large portion of ethereum Supply so of course that makes this organization extremely important and it had the responsibility because ethereum launched without being like a fully fleshed idea there were many different Milestones that ethereum had to reach on its road map proof of stake being one of them that the foundation was meant to help steuart um to ensure that the that development road map got met um so they had this money in the beginning and so that's why the ethereum foundation was very important it housed it was like the core development House of etherium and you can actually see that same model replicated in many different protocols right so there's like the core op Labs there's is there a foundation for arbitrum probably um is there a foundation for Avalanche is there like a core foundation for salana um you know when you when you launch with like a very ambitious development role Map There's is usually like a central entity that's you know trying to pursue it um but over the years things have changed ethereum has grown and um a the and the the foundation's role has has been to kind of um try and reduce its impact on the ecosystem and um make sure that the the stewardship of the ecosystem is is shared by many different people but to this day the ethereum foundation still plays a really big role in funding a lot of the research that goes into informing different uh development and protocol um decisions that get made and um the etherum foundation continues to to House of course um maintain the most important execution layer client or the most popular guest um and you know organize things like the ACD calls which we'll get into in a couple slides um so so yeah so for for many reasons I think the ethereum fun Foundation has been continues to be a very important player in ethereum governance but for different reasons because now it doesn't have as much eth Supply as it used to um and so now they're important for other reasons than just you know they're a huge bag holder which they're not anymore um and actually have a report on that if you're very interested in knowing about um supply distribution over time there's a report on galaxy.com small plug um if you're interested in looking at more data around that so yes the first key player I think that I want to highlight for AUM governance is AUM Foundation keep an eye um keep pied up and then client teams obviously um the people who are building the software for users to connect to ethereum um there's many and um there's actually a great website client diversity.com I think that shows how popular each of these clients are um and they are the developers who who maintain and build the software and one quick thing that I want to to mention on this slide is when we're talking about the people who build ethereum software they're not it's client teams and the term ethereum core developers is not synonymous um you might hear the term ethereum core developers thrown out a lot and the definition of that term has also changed over the years um but C ethereum Core developers in my mind at least defines people who are working on the protocol of ethereum not only as a developer for a client team but perhaps as a researcher maybe they're helping out with testing upgrades maybe they're EIP editors um so even if they're not part of you know one of these client teams um I think you could still attach the term ethereum core developers but that term and the definition of ethereum core developers is a very hotly debated subject in the ethereum community so it's not that important but you might hear it around a lot in discussions about ethereum governance um so I thought I'd just make a note there um but clientes are very easy to Define and they're really the people who who are building either of these protocols oh there's a new one coming out reth shout out to re they're building trying to build you know a new ethereum client um yes and then we get into the third group so who are the people who are actually running that software that those nine clientes are building right um primarily it's the validator node operators not the stakers again um it's you know the likes of people like Lio rocket pool coinbase Kraken you got validator node operators these are the people who are running the software and I really wish that there was better data to show the types of people running nodes the number of nodes over time um because this chart is from ether nodes and there's really only a handful of sites that are are showing this data and the accuracy around this data is not that great as you can even see on the on the text there that number is just from a a survey a survey a random survey that was done by the ethereum cad herders um they're a project management group um in the ethereum ecosystem but we don't have time to get into them right now though they are they are doing good work um so yes so validator node operators and when we think about validator node operators one thing that I want to highlight about them is the likelihood of validator node operators potentially saying no to an upgrade so let's just say the centin on the prior side and the Ean Foundation was like hey here's a new release like go forth and upgrade um the likelihood of validator node operators resisting that change I think is getting it's getting the chances of that is getting lower over time in my mind for three reasons the first is when we think about the consistency of upgrades that happen on ethereum the norm and and the idea when I run a node is that on ethereum I'm probably going to have to update it in 6 months um that kind of conditioning so to speak I think does make the upgrading process a lot easier and the push back to upgrades um kind of decrease over time and then you've got of course the increasing um application layer you know you've got tons of applications now that would be irreparably damaged in value if you were to try and resist an upgrade and potentially create a permanent chain split things like l2s you can't have L2 that wouldn't be impacted by a permanent chain split that L2 would have to choose which is the real ethereum um so for you know reasons of increasing liquidity and and you know frag the concerns the risks of fragmentation of liquidity but also different types of applications that would be permanently damaged um I think there's there's lots of of good reasons to think about why it would it's becoming increasingly more hard the more valuable the ab layer of ethereum gets to to resist an upgrade or to you know rock the boat and cause a permanent chain split um which might not be a bad thing and then the third reason of course is that who are these validator node operators a lot of them are businesses and one of the worst things that you can do for your business is jeopardize the network on which your business is built um so you know for for many reasons I think these validator note operators their their incentive and their kind of default mode is to resist any contentious any kind of upgrades code changes that might um split the network apart um so that's kind of important to keep in mind in terms of like what's motivating these actors of course it's not going to inform um how well we'll get into it with certain examples of upgrades in the future but um it's important to understand what's motivating these players and it's important to understand how governance might um be how contentious governance decisions might be accepted by different groups how it might be accepted by the ethereum foundation versus client teams versus validator node operators so the last and the fourth um team that I want to highlight um group in ethereum governance that I want to highlight that I think is very important and not talked about enough is DAP developers dap developers are a very important group group that informs decisions on ethereum protocol governance these are four examples um Where Da developers have just been the key focus of these upgrades EIP 1153 um was primarily you know advocated by Unis swap labs and op op optimism Labs um and it's because you know this upgrade would save their users a lot in gas fees um ERC 4337 not primarily pushed by dap develop Vel opers but the whole idea of it is to service and to increase the adoption of daps on ethereum by making it easier for users to interact with daps on uh to make it easier to interact with daps yeah um and is you know attracting a lot of interest from daps and wallets and other service providers on top of ethereum and then you got EIP 7516 which is recently included into the Cancun denet upgrade um and one of the reasons why it was was because uh developer on top of ethereum was advocating for it right the layer two rollups that were that we're soon going to start to see you know blobs on ethereum those developers who are building L2 rollups were like hey ethereum developers we need this because if you're going to have blobs and if we're going to be able to use those blobs to you know service our our users cheaper we're going to need a way to price it and to um better more easily price the user transactions on l2s then you got to give us this um so another a great example of where applications D developers um influenced governance influenced the protocol development of ethereum and then finally self-destruct another really great one protocol developers are like we need to get rid of this okay we this soft cou is not safe um but it took them forever to get to this stage of like let's actually deprecate it because they were thinking about the app developers certain dabs that would be impacted by the removal of this op code and how do you ensure that certain daps that are using that off code can still you know you know won't be affected or if they are going to be affected have a work around to ensure that their application keeps moving and chugging along so I actually think dap developers um have a big voice in the ethereum governance system um yes so those are the four now let's talk a little bit about the forums I'm just going to keep an eye on time um let's talk about the forums that um are very important to ethereum governance and where the decisions get made where discussions between these four groups really happen the first and foremost that's most important is probably the ACD calls there that's where a lot of the decisions on about ethereum get made um or at least communicated on these calls they happen every week now and they um are frequently attended by EF authors client teams ethereum foundations researchers um eum Foundation researchers and the cat herters um and it's open for anyone to join um these these calls are live streamed on YouTube um but a lot of what gets talked on ACD calls are informed by what's discussed on online forums like eth magicians and eth research um so you know people who are presenting eips talking about what protocol changes should get made on ethereum their first you know put up on these forums and then you know they trickle down to the ACD calls Discord another very important forum for more day-to-day coordination I wouldn't say a lot of like long-term objective thinking and planning and decision-making but Discord is a really great place where developers do like real time how is this upgrade going or like real time debugging um and it has a ton of different chat rooms where people organize for calls that are not ACD calls like calls for specifically for EIP 4844 specifically for I don't know veral tree development um and fun fact if you guys know or have used gter before the purpose of Discord um ethereum developers used to coordinate around getter so what they do on Discord they used to do on getter um but Discord was just a better way to manage multiple chat rooms and then GitHub is the last forum and um is really I would say if a decision gets made on ACD and you want to confirm that the decision like the execution of that decision happen it happens on GitHub so like if certain pull requests issues decisions about what goes into an upgrade what doesn't all of that is recorded as kind of like a final step on GitHub and GitHub does host like the ethereum specifications right like if you were interested in building an ethereum client if you're interested in knowing how ethereum works this is kind of like the the you're getting it straight from the horse's mouth um and there are talks about moving away from a system like GitHub um because GitHub the website is of course centralized right so like GitHub as ay could shut down access to their website in certain parts of the world um but that's why developers are encouraged to host localized copies of what they're working on um a lot of processes and decisions on GitHub of course are replicated and duplicated in terms of Records on Discord all these other forums um so yeah governance forums for people to know okay so past upgrades um many different upgrades have happened on ethereum I'm not going to go through all of these but I do want to highlight how many there have been and highlight the fact that a lot of what I've been talking about has changed over the years so when I talk about validator node operators and the fact that these are the people that are running nodes that's a very recent phenomena for most of the upgrades it was miners like that everything that I was talking about in terms of like the interests of validator node operators that's pretty pretty new stuff because before we were as if you were interested in governance we were talking about what are the interests of Miners and how much power do they have over the governance process ethereum Foundation like I said their importance and their role has changed drastically from the beginning of ethereum to now so with this slide I just want to highlight that with every single upgrade that has happened on ethereum governance on ethereum has changed um and governance on ethereum is likely to continue to change as the protocol of ethereum gets changed imagine what happens if you know ethereum does achieve its dream by being a completely like da layer most of the application activity happens on l2s imagine the governance how governance might change like maybe another fifth group that starts to grow in popularity and importance and size is really the l2s um they're like L2 foundations but anyways that's getting ahead of myself so let's end off this presentation and talk in the last couple minutes about the future the future upgrades Cancun denb it's it's coming guys G probably not likely for this year but um possibly for next year who knows um obviously it's a little too early to say um because we're still in the midst of developers are still in the midst of testing the upgrade but the reason why I um was a little I am a little hesitant about saying that kakun D might happen this year is because of the addition the last minute addition of 2 a two eips that are the ones underlined and we already talked about the blob based fee why that was included um and the role of course that um L2 developers really had in pushing that through I want to talk very quickly about EIP 7514 um it's to address the issue of validator set size growth important issue and when the decision would was made there was some concern among the ethereum community shown on Twitter and other forums that um actually we didn't talk about Twitter but anyways that um that that look you know the the decision was made very h that um I don't agree that you know this is as pressing an issue as you sh as you saying maybe we shouldn't include EIP 7154 but EIP 75154 the solution that it gives to the issue of validator set size growth is a short-term solution there are more there are other proposals that are trying to address that issue with more radical changes to validator economics and network monetary policy and so I one of the things I really want to do right now is just to make a call out to people who are interested in ethereum governance and to just ethereum stakeholders in general that this conversation around how do we address validator set size growth is not one that ends with e75 7514 and that it's important for ethereum stakeholders to consider what are some of the other proposals that are being considered to address this issue um because it is going to have major ramifications on the netor work and we can't have another like I didn't know this was an issue or like you know kind of like after the fact that decision has made made um being surprised at kind of the results of of how governance was done and how consensus came to be um so EIP 7514 kind of a unique a very recent I shouldn't say unique it's a very recent example of where I think governance happened primarily by developers but there was such an opportunity and there still remains a really big opportunity for other people to enter into this discussion and to you know make their voices heard about how validator economics and network monetary policy um should change moving forward um so yeah so Cancun denb coming up Prague Electra is the name of the upgrade supposed name of the upgrade for the for the next one after Cancun denb and there's a question you know what who which upgrade is going to be in the driver seat of prag and Electra um it's not going to be like all of these all at once because a very recent Trend now in the way upgrades are are created on ethereum is by having just one big code change and several smaller ones so for Cancun denb the biggest code change going into canun de denb the one requiring the most amount of testing is EIP 4844 Shanghai the biggest one was obviously the state deth withdrawals merge it was the transition to brip stake um so prag Electro who's going to be in the driver's seat there there's so many different priorities that people are going to be lobbying for is it really to address the issue of validator set size growth is that the Cod change that needs to be the main one in PR Electra is it the major overhaul to the evm with the big eof upgrade is it a change to the way that ethereum stores state with veral trees who knows it's going to be a contentious heated topic and for it to come out to the outcome that I think will um reflect the most like robust and resilient governance system that ethereum has we need the voices of many different stakeholders to participate um so making another shout out for PR Electra and for the discussions around that um that I think people should really participate in and um make their voices Hur in um and I didn't get really into the technical like the increasing complexity of ethereum as ethereum upgrades more and more but one thing to note as I kind of started off this slide about about timing is that the more upgrades we do and the more complex ethereum becomes the longer it's going to take to test these upgrades because there's a higher technical risk there's a higher amount of of probability for bugs the more complex your codebase gets which means testing is going to take longer and so um as you know we think about how upgrades might get slower um for good reason um you know each one is going to have hold you know more and more importance because of how much effort and time and cost goes into it um anyways so thank you thank you so much for listening and uh yeah if you have any questions I'll be just out there