Maximal Decentralization On All Axes

Transcript

okay howdy uh title of my talk is uh decentralize everything I'm Evan Van Ness you can follow me on the twitters at Evan underscore van underscore Naz so I wanted a big thank you to my opening act vitalik buterin I really got the crowd warmed up really really thankful for a great uh warm-up Act um for those of you who can't see on the live screen on the live stream uh there's a there's a there's a few die hards that I'm I'm thankful to all of you for for sticking around um vitalik probably said half the stuff I'm gonna say but much better so uh yeah so

um I think this talk is really like a a an exhortation about decentralization um so you know it is a goal uh it's a goal that we should all have I think it is a goal that makes the ethereum community unique but it is also a self-interested goal and I think people really especially in Bull markets tend to lose sight of the fact that uh decentralization is not only uh something that we do because we care although we do and we we should but also but we should care because it is in our own self-interest um so you know as I said just said like why decentralize like um number go up um uh it really is a thing because we

can ship faster I mean not only is it the whole point of the entire space like we are not a centralized SQL chain um we believe in maximal decentralization at all levels um if that's news to you I don't know you probably belong at the sequel on a conference or something um we believe in maximal decentralization at all levels and I think it's like something that is always worth repeating uh at any Community Conference because it often gets lost in Market ups and downs um so I have to be careful in what I say here because I would say this is not a universal view and uh some people can

get a little bit uh pedantic about certain things that I may or may not say in the slide but you know look it come at it from a theoretical level like when we when ethereum staking is live um uh the incentives are that if no client has more than a third of the network then the network cannot lose finality from one single client book right so if you think about that it may it lets us ship much much faster if we know that one client can have a single bug and there's no loss of finality so running clients allows us to move faster and I think you know we've really seen

it already right with testing the merge we are now at like 15 successful tests we're at the point where the only problems that we're really having are basically configuration issues because we're having tests so often that clients you know understandably you know it's just a weekly test it's just one more thing to do and you know they make a small mistake because because it's you know it's just the weekly test and that's an interesting sort of uh thing because it actually means that we're sort of like randomly you know stress testing different things right because uh you know even when config issues and happen and you know things that definitely would not happen on mainnet like the network has still been extremely extremely robust but I mean staking uh you know for like

the last two years I have been tweeting and it was uh sometimes felt like okay I just got my volume turned up I thought I was talking loud enough but yeah wouldn't be the first time this happens every year at ECC and apparently I'm I don't project enough um staking you know for for a while it felt like a little bit of a lonely chaotic Crusade for better staking diversity um you know prism is a fantastic client I have a good relationship with them um it with you know my uh you know me and some other people like fizz uh saying you know you shouldn't be using prism you should switch to minority clients but really it was in their own self-interest too right like nobody

wants to have the fate of the ethereum network on you know the code that you wrote no matter how you know how much work you put into software there is always the danger of bugs so um you know we have improved a lot on that front I think prism which was up at a you know at a super majority level not long ago by by any means uh is now down to about 40 percent of the network it's a little bit tough to know for sure with exact detail um but you know Michael out of out of Sigma Prime has done some pretty good work around block Printing and we have a relatively good estimate that we feel good about that you know prism is down to 40 percent and actually Sigma Prime

is now at around a third and the the ideal is that nobody is above a third so you know I I you know it's it's become a thing where we no longer even necessarily say that you know you should reuse Lighthouse either um you should run not even non-minority clients or sorry not not majority clients or not plurality clients but you should run clients that are under a third um that's the staking layer um obviously we have a bigger challenge at the execution layer uh Geth has been around for a long time it is an impressively hardened piece of software it is quite robust uh it is you know that they they optimize for that

um but you know that also means that right now like they do have the the fate of the network on their shoulders so um from an from an execution layer standard like we have a long way to go uh for better client diversity um and you know it's it's fair to point out that like the other clients the other execution layer or clients um all have some things that they're doing different that are interesting um whether it is you know better traces better things for developers the um um whether you like Java because you're an Enterprise whether you like nevermind because they do some of the stuff I said whether you like Aragon because you really want to run an archive node um

whatever it is um you know we have a some some really clients that have that have features that the gath doesn't have um of course Geth is also very hard-end and it is now used by like a crazy amount of ethereum competitors and Roll-Ups right so um it is uh there are trade-offs but you know especially if you are staking uh you should want to use an execution layer minority client um just because of the way like I said that the spec works okay so I sometimes ask people what is ethereum's biggest long-term problem and uh you know it's a discussion question

right so the the answers can change over the years my answer for a while um and I've never said this publicly and uh no no I might get some flag for doing so but I don't know who cares is that it's not technical um you know vitalik just talked about all the the technical questions I would actually posit that the biggest long-term problem is that um from the outside vitalik is seen as the leader of ethereum and that is on a governance or political level or technical level much less true right now than it has ever been before but in the past um except maybe at the beginning because

in the beginning right there were three different clients they were counterweights this was on purpose because we were trying to be more decentralized than Bitcoin we didn't want to get into a place where we felt like Bitcoin had gotten into where we felt like blackstream had a veto so there were three different clients uh you know the python client led by vitalik um Geth led by Jeff and of course parody which was not called parody then but uh led by led by Gavin um you know and you can overstate this in in a lot of ways too but um I think especially from the outside a lot of people really do not expect uh that like vitalik is as little in charge uh as he

is and of course that is some somewhat by Design because vitalik doesn't want power and I mean I don't think there's anyone in the world who wishes he could have less Fame and and power than uh than vitalik um but uh there are practical reasons why this really matters um if you want Community coordination um like things around the the um like reducing the client diversity or improving the client diversity at the at the staking level like I was just talking about um yeah I mean I felt like I was uh going uphill for for for years um and um you know in some ways we probably could have solved that if vitalik had

tweeted out about it a lot more um but you know on the other hand like it's it's also about the ethereum commits to decentralizing and doing the things that we're supposed to be doing to do our you know to be decentially maximally decentralized um you know a massive like problem that we have now is like people staking with exchanges and of course I totally understand why many people do it I'm not going to get into those reasons I think it's obvious um um again like I would say this is an exhortation because I feel like the people that that stuck around probably know why um a lot of this is but you know there's a spectrum you know it's from homestaking by yourself you know uh running your own validator which you

know is really not a big technical commitment you know I they're really there are plenty of guides out there if you're willing to like be careful and run a a validator especially if you're you know just doing one or you know whatever it is it's really like the technical knowledge is um less less than you might think and certainly like the devops commitment is also less than you might think I think there's been so much fun over the years about how hard it is to run an ethereum node uh that people were like a little bit uh okay careful about over committing to how hard it was to stake but yeah so you know then there's like decentralized taking pools like rock pool there's some others um then there's Lido which you

know I don't even call a staking pool I would say it's sort of like a tokenization service with governance on top of staking pools um you know all of these one thing I would definitely say is um one of the reasons why it matters so much is that uh probably 50 percent of steak is in AWS right now which in my opinion is insane and if I had my mistaked Eve and a staking pool that was running on AWS I would feel like it was in danger so I would not be doing that um it is shocking how many of the exchanges and staking pools instead of being responsible and running it on

their own servers or computers are putting it in AWS you know if we ever have any sort of like regulatory environment or you just AWS has a massive you know outage across multiple regions those people are going to lose money and it is um you know going to be a black eye for ethereum um yeah don't stick with exchanges commit to a bit doing better um I think one other thing that is like you know good is like switching away from systematically important software and committing to doing so in the future like solidity compiler doesn't really belong in this list but as a community I

mean obviously like you know committing to doing a an evm compiler with a language is a hard long-term problem um that you know I would not ask anyone to do on their own takes you know years and millions of funding but like we should recognize that like ethereum can do better here and I you know in some ways we are um with Faye with Viper and there are challenges there as well because of you know developer tooling you don't want to have too many um RPC calls like you know in the past we were certainly were a lot more centralized on one way to read the chain um and uh I think it's great that you know there is like the whole Bitcoin meme about inferior goes down ethereum goes down um you know that's not true that's like

that's just chain raids but um you know it's certainly better if like you're using other things other than just things that run on uh well first the same the same thing like in fira or the same you know things that run on abs uh you know while it's like uh it's sort of by the way real props to the solidity team like for never having had a security incident and same with metamask like it is it is it is unbelievable the hate that both of them have gotten in the community over the years um and yet like again they have by far gotten the number one thing right which is neither one of them have been hacked um so that is I mean that's a really impressive thing but it goes back to my point earlier about shipping speed

um when you do have everything relying on you you have to be extremely careful take it slow um you know not even make breaking changes which might be fine technically but what your users um you know might might screw up um you know fun one I just threw in there is like the BLS Library Mommy just uh came out with the new BLS Library um so that is faster than blst so I think we'll see some of the the staking client switch to that um and again this is also sort of you know this is kind of a laundry list of things but Roll-Ups people often ask me what roll-up they should Deploy on and uh and they and what I I why I often

find is what they really want to know is which roll-up is going to win because they don't want to be on the roll-up that doesn't win and I always tell them like they're thinking about it the exact wrong way which is like the ethereum community should be committed to not no roll up winning um you know whether it's a diversity between optimistic and ZK Roll-Ups whether it's a diversity within those two like we should have multiple bets in optimistic and in ZK Roll-Ups and that is really important and so like right now I'd say if you needed to play on something you pick an optimistic roll up like optimism or arbitrum and and you deploy um you know if you need ZK for whatever reason like you stick around

um you know long term that is that is not imminent probably although they like to say so this week we'll see um so this is just a really quick action item list like you know it's basically what I said um don't say quickly don't stick with exchanges stick with minority clients run your own nodes whether you're staking or not um you know the the leaders of ethereum somebody asked somebody the other day said um like nobody understands that you know like Danny is uh more important for the next two years of ethereum than vitalik um and uh I think it's like that's that's very obvious to me but I think that's probably not well known so I think like uh one thing that we can do

is literally elevating some of the very senior researchers in ethereum um like like Danny who um really is like doing all the coordination Don crud who you know is dunk sharding or tank shouting depending on how you want to say it and of course Drake Justin who came up with BLS signatures for ethereum and switching away and committing to switching away in the future for systemically important software I'm pretty sure I'm done now I am thank you [Applause]