5 Foundations of Multichain Governance Harmonizing Decentralized Ecosystems【DAO TOKYO 2024】

Transcript

[Applause] hello everyone um thanks for coming back from lunch I heard it was delicious but we were stuck in the speaker room so we missed it um okay well let's start off with some self intros I guess we can start l first of anthering hello yeah so anthering from E storage happen in this space um since 2018 always doing like infra stuff um before that I'm doing like the tech traditional tech companies comp Valley and also like financing New York do we need to do a little bit for the project okay um so EST storage it's a storage layer for Ean

that is saying we're layer to Native ipfs for eum we can be later three for any later two as well so that's why we're working with a lot of later two instead of like competing because most of ler to trying to um get more TPS we're just helping them to store more stuff on chin um like much lower like cost yeah hi uh I'm Jacob I a rust developer in the blockchain space since about 2017 I recently joined matter Labs which is they're the developers of ZK sync which is a zero knowledge uh layer 2 on ethereum that tries to fulfill the

values of the ZK creto hey guys my name is Terry U headed strategy at eclipse eclipse is building eclipse is bringing salana onto ethereum we're building the first salana virtual machine layer 2 on top of ethereum great to meet youall cool um so today's panel is about governance now we have one live chain which is ZK sync I believe uh eclipse is kind of live you can tell us more about that and if storage is still in testet but I'd like to hear um how everyone is approaching governance I guess in the case of ZK s what's live and uh in the case of yall what will be live or how you're thinking about it sir sure all right so uh as of a

couple of weeks ago now uh ZK say governances live through something called ZK nation which is our governance platform uh it has uh three it's three prong approach there's the uh token delegates there is the security Council and there are the Guardians so um token delegates of course uh using the ZK sync token can uh submit and vote on proposals and then the two other councils are made up of of individual members that have uh veto or emergency Powers I can go um one of the key flaws of nl2 is liveness in the case of a

single sequencer failure right so say for example your arbitrum optimism and you have a single sequencer and some reason you know um like there's an explosion in the region that their AWS box is in their sequencer goes down then the chain is down uh for as long as a a 24-hour I believe period before where someone can open up the settlement window uh or the settlement contract for permissionless proposals I think that is the biggest kind of weakness of l2s today and so where we're thinking about is okay let's have a decentralized sequencer set number one uh but let's have it such that we can prevent collusion by having a mechanism of governance where you can vote out existing sequencers from the sequencer set and replace it with a governance

designated entity that's like number one that we're thinking about and of course like second of all we will have for the time being guard rails around upgradability of the bridge uh with something um similar to you guys in which case is is a security Council so those are kind of like the two prong governance items that we're thinking about right now yeah um so e storage is a little bit different like I mentioned before so we're not like a typical L2 um aim for the more TPS we're like helping Ean is like ler two and also helping other later two as um L three so in that case so there definitely will be a lot of like multi chain stuff um So currently we're still on the public test net so we're designing that but for sure so we

will consider like you know how to gather like multichain governance into one um we have been very active for the op governance and uh I think it will be great to take a lot of like their lessons and definitely they're doing great um since today's topic is about like governance so one product we have been developing it's I think it's worth to mention here is web3 protocol decentralized HTTP um so the idea of that is right now um all the the apps although they're like decentralized you know as like a app but their front end their website everything is centralized so basically they have to e conquer like DNS infura those kind of like

centralized component so what we do is helping them to De decentralize everything so basically they can just um store their like front the end in the smart contract um so in that way so even their website can through the governance so if they want to update any like you know um the code or change any part so that process hat you through like the governance as well in that case it's really hard to hack because a lot of like hack come from the DNS hack um and that part we already support a lot of like L2 and L1 yeah so for for that that's something we want to explore or later like how to governance through like the decentralized web

hosting um very interesting um I'm curious okay so in your case Terry I guess it's mostly aspirational right you're not live yet yeah mostly aspiration right security councel is life that's great um but I'd still love to hear I mean I assume you guys have pretty clear U design in mind at this point um is there anything that you would still take from any other L2 like some feature of their governance that you think is really smart and you wish you had implemented but maybe you didn't have time for and I'd love to hear from the others as well uh nothing particular comes to mind right like I think I think generally like token governance like I think

arbitrum like something that comes to mind is what I don't want to do which is I think arbitrum um Dow Capital allocation based on token governance voting was a large mistake because the way arbitrum da works is basically arbitrum DA has a huge treasury of all of its ARB and on top of that all sequence of Revenue per year goes back to the arbitrum Dow all the Dao treasury is controlled by um people who coin vote uh delegates can propose uh ways this treasury can use their capital and then these are ratified via token holder voting and so recently what happened was um like a couple of the gaming delegates got together and and proposed a $200 million arbitrum gaming fund right

arbitrum is neither you know a really great place or technically suitable for gaming uh it is not really a place where there has been a lot of gaming Innovation with the exception of treasure da who has kind of spearheaded this movement and so uh it is questionable whether that is a good use of capital or not it is kind of hard to tell at the moment you know gaming has a very long investment cycle and it takes a very long time for it to pay off and you know if it can bring a million 2 million users onto arbitrum orbit that is actually a huge payoff but it is unclear whether it was a good choice or not and uh it was something that the token holders were able to pass through so that's something like an example where I don't think it's a good idea to

put Capital control in the hands of token holders Jacob that's a really good question I think that governance is really difficult to get right we as we've seen from reality yeah practical experience um Capital allocation is definitely a tough one projects that I've worked with before have had uh many issues with uh you know giving out grants or uh something like that to projects that end up not following through uh which is it's difficult to see the future and I

think people are really concerned about their bags and so that's going to be the one of the issues that gets probably most of the negative press and therefore most of the eyeballs is going to be misallocation misuse of capital um although we haven't had like you know the concrete design um the the plan for everything for governance but we do have some small triy for the grants for example like how to um distribute like the grant because usually um you know the team decide like who will get like how much and this time we are actually working with some developer communities um so what we do is we have like a small like one month

like campaign stuff um so during that campaign so they help us to um recruit like developers study what's the web protocol the decentralized HTTP I mention how does that work and uh they're helping them to build something and then uh letting like the developers holders U that Community holders and us so together to decide like who will get how much something like that yeah um so I know both of you guys uh mentioned security councils right now when I hear that I think oh we can't trust token holders we can't really trust token holders right

and now you just mentioned you also don't think token holders should really be trusted with the bags either so I do wonder what do you trust token holders to do so this is where I think um this is where I think like replacing the sequencer uh and or you know potentially replacing one or few of the sequencers is a good idea because and this is kind of a nebulous thought we have at Eclipse as well but if you issue a token on L2 I assume zks is issued on the L2 not on L1 right so like arbitrum is is also issued on the L2 then all the funds that are locked and issued and canonical on the L2 if there is a liveness issue with the sequencer or censorship with the sequencer they become invalid on the L1 right versus

like an l one asset that you lock through the canonical Bridge uh if there is an issue you can just withdraw from the canonical Bridge right so if you have an asset that's issued on the L2 it's not on the canonical bridge and therefore on the L1 if the L2 goes down then they're gone so if you issue your native token on the L2 right then those people that hold the native token on the L2 have the vested interest in the correct continuation of the liveness and safety guarantees of the sequencer operation right and so giving them control over that is in a way incentive alignment versus giving them control over the entire treasury which can be used in you know various different ways to benefit various different folks who have various different bags in the ecosystem is incentive misalignment so

that's where like my thought is going is what do the token holders of the L2 really care about and what is critical to their success uh to me it is you know something that they should then have control over there it's a genuinely difficult problem I'm not sure that that um I'm fully on board with the statement that I don't think I said that I don't trust the token holders with the bags um I said it's a very difficult problem um and as we saw like recently with uh optimism sometimes the security Council

has to step in and perform very essential services to protect the network it is definitely an issue with centralization of power um because as you mentioned if there is a security Council of 12 individuals who have like ultimate veto power or something like that then really um they they kind of do have this uh centralized power where they can just say yeah you know what nope I don't I don't trust the uh token holders with this proposal and they have that power but um it's this constant push and and pull between you know with centralization of power you have um

people who connect quickly and decisively and sometimes in the case of uh you know if there's a network issue that is uh potentially a very valuable and important way that the security Council can step in and act to protect the L2 um but then conversely you look at that and you say well now again we have more power and Trust centralized in these individuals which if they uh act poorly or against the interests of the network or something like that then now you're in trouble in a different way so it's a difficult push and pull and I guess to clarify like there's not a single implementation of a security Council in L2 today that is controlled by the token holders they're mostly

people who are reputable kind of folks inside crypto and or reputable folks inside the ecosystem and their kind of very narrow role is to be able to either pause uh withdrawals or pause the bridge is my understanding those are the two things that you can do yes and when you say reputable people though that is uh another way of saying just trust right usually we say don't trust verify but in this case it's they're reputable it's a argument based on Authority it's they have this reputation they've acted this way in the past we're trusting them to act this way again in the future um and that's difficult to square with a lot of the

values that we typically like to espouse in crypto it's definitely a a guard rail that l2s have to it's like a like a diaper where they have to take it off one day you know to get potty trained I think that would be the ideal but again it's it's really tough to to get to that point Fair because it's impossible to have like secure code uh that is just like by definition secure right yeah um I think people who are developers have probably we heard of code rot right code in many cases is a liability we um as developers you know uh we see more code more complexity is often uh a liability rather than you know a

resource especially you know if you want to deploy an upgrade or some new attack Vector comes out you know it's constant state of of development yeah um I think terara just mention like you know um if we put like all the things in L2 and uh if something happens so how can we just withdraw like the F or something because I think a lot of L2 have such issue because it's too centralized very limited noes and uh yeah so that's why I think like if by require you have to submit your data into like a d right if um by like um for a44 or like Celestial or like whatever um and uh yeah I think that's that's not something about like the governance

that's something about like if something happen on L2 whether you can just with store your money because on L1 we knows you have you can do that but on l2s um so what's the design for that yeah and uh what we do is actually helping all the da to maintain the data forever just in case if something happens we still have all the history data and then we can just withraw the money uh withraw all your fund yeah well it would be more like um imagine the sequencer goes down and it stops posting batches to the DA layer then you're like okay all our funds on the L2 are stuck on if it's go if it goes to the L1 bridge then you can take it out using

forth withdraw forced withdrawals but if it's not then you need a way to make sure you can elect a new sequencer which is where governance can come in useful you seem very concerned with liveness uh is that because you work with Solana VM uh so we we won't suffer the same um so a lot of the salana uh issues back in I would say like January 2023 there were a bit of consensus bugs uh they have been patched if it does happen again uh yeah like that would be of concern but just because we have a means to elect a new sequencer it wouldn't necessarily be able to resolve consensus bugs in the

fundamental Salon a client actually yeah it's It's Tricky right because you're in a lot of ways using someone else's code right and they kind of uh solve those problems for themselves but then if you want to be Salan equivalent then you're going to have to kind of yeah you you put yourself up for a very hard job basically it seems to me well so we've built the the we built the L2 in a very modular way such that anytime the gve client which is the salana client pushes an update just instantly we can update our our uh our our L2 execution engine as well so like two weeks ago actually salana had a critical vulner unability uh where they needed to push out an update to like the

uh the SF backend so uh like part of the virtual machine that kind of the salonic code compiles down to uh yeah we got that done in like 30 minutes just like right after the push the patch was pushed very cool that's really impressive and I really respect the people who like you said are doing the hard task because the people who avoid the hard tasks are not the people who are going to make the changes so Props to you thanks cool um so I guess the way we've been talking so far it sounds like elus need governance they need a security Council they need all these things uh and yet obviously we're all building on ethereum

which doesn't have any governance right so is there a reason why it's not possible to have an ungoverned L2 is that ever going to happen what do you think uh what do you think about that ethereum def definitively has governance um there is a means for which right there's a bunch of hard Forks like every year you get a hard fork and if you want to have a hard Fork you need some way for people to decide you know what to include in this update how do you communicate this with you know the validator clients that are building actually the implementation of the ethereum software and how do you you know in the past it used to be that you also needed to coordinate with miners but now you also need to coordinate with like lsts and all these like large organizations that have stake in

ethereum all need to be coordinated if you want to push out a significant update right and so um like there is a kind of social process in ethereum it happens via these like core developer calls where you know ideas get uh inseminated and then they become disseminated in the research forums and then people comment and then you know they become I actually don't know how things get included in uh in hard Forks like I actually don't know maybe I mean it depends on the hard Fork but like how do you decide like oh this is going to be get in pectra this is going to become get in like sh Shai or how do you that's not something that I can speak on with a great deal of uh experience so yeah fair so there's a method in ethereum in which like things get included in updates and in a sense

the L2 is very similar right the L2 is a dynamic um Network where things change and there needs to be a way for our ideas to generate uh and then get disseminated in the network and then get implemented inside the actual execution engine of the L2 now whether that's in the settlement layer the bridge the withdrawals or the actual execution environment uh that's still um means that you need some sort of governance process what I'm trying to say is like eth has governance process and therefore l2s can continuously to have continue to have a governance process even Bitcoin has a governance process right I think ethereum and and like you mentioned Bitcoin while they do have governance they also have the advantage of being significantly uh older more mature larger ecosystems than uh any of the l2s we're talking about here um and when you're talking about a much smaller

ecosystem um especially when you have a governance token um you have have to worry about you know like there are there are more there's a larger set of malicious actors that would have a capability to like uh buy a large amount of the governance token for example and start submitting like these bogus proposals or or even actively malicious ones so um when your ecosystem is larger or bigger or older um I think the while ethereum definitely does still have governance sometimes as you mentioned it it doesn't necessarily look it it's not staring you straight in the face so much I guess it has governance on what the direction of the

network goes in so far as if you have a large percentage of stake you can decide to run a different version of ethereum or like you know you can decide to ignore a hard fork and you know go toward another path and if as long as you get enough of the stake in like the governance pool then you know you can kind of like Fork ethereum quote unquote are people still worried about Lio I I don't think it's ever going to happen but like you could it's not like you know vitalic says oh here's here's P Das now everyone vote do you want it in the next hard Fork it's just that you know people decide to run it on their client software right yeah um I think both of you already just mention a lot of part I think for isan definitely has like

governance right because we know there's like core def team we know there's like is foundation to decide where the grand goes and there's like even for the meeting side like the ad that count or something they're like just a different like organization doing like different stuff um like you said it just there's no like um you know proposal saying like we want to do this feature everyone vote whether we want to do that I think it's because it's it's biging love and it cannot just let all the individual to decide like the important things otherwise um because we know the valuation is so high a lot of people who um invite who who has like you know B on that will be hurt for that um otherwise for the L2 which is you know has much

small history for that I think they're just trying to do something different to see whether this work like op it definitely is like more transparent compar with like eerin I think they intend to do that to see like whether that works but still I guess for the future wise so they will have a lot of like proposal um for um the community saying like we want to try this one so what do you guys think but I don't think they give like all the power to see okay everyone has the token to to vote for that right it's just like they give like more transparency um I guess for l2s and protocols like us who are less mature um still you need a lot of like guidance at

the very beginning and then so you need a lot of involvement from the community to build your consensus so Ean um threw a lot of time to build their consensus already yeah I guess doing governance it's like a process of building consensus yeah um so what do you guys think is the role of uh token holders uh com in um relation to uh sequencer fees right should token holders in l2s receive sequencer fees I mean we've seen the amount of tax that l2s pay to L1 go down significantly right so we eum is not

really receiving as much uh taxes as it used to before 4844 so where those sequencer fees going to go in the future and how do you guys uh think about uh that income right in ethereum we had a very clear system we burned the E it benefits everyone in l2s where's where's the money going I think sequenc refuse should go to daps um daps are where fees are originated people want to transact with daps dap bring in execution revenue for the L2 sequencer fees and execution Revenue should therefore go to deps now if there's any excess that you can charge because there are network effects of being on your L2 in particular then it's a question of I think it's then a

question of like legality right like can you actually distribute that that to token holders or not then of course they they should belong to the Token holders if the token holders are owners of the protocol but then you know like that is a question of legality I'm not sure I have a great answer to that question what what are you currently doing with sequencer fees uh I as I mentioned I just joined matter so I don't know that part it's all right it's all right it's all right it's all right they're not significant anyway at the moment so I don't think it's an issue oh we don't have sequencer because we're like storage notes um for people who provide that but I actually agree with Terry I think partial of like a sequencer fee uh need to go to the dab

and dab can decide whether they want to get back to their users because the user who are using their dabs and to you know just create gas or like SE of fees right um but they can decide it that part yeah yeah so I I think that's a great point right as someone who builds on top of l2s and not building it all to myself I I think yeah we should get the money right however um I'm curious um how do you think of uh attribution because having been on the other side of that um you know recently Optimus and did the retroactive public goods funding and uh I mean for those who are kind of in the in the middle of that attribution was a like it's very hard to tell like

when you say adap right if everybody's using usdc does that mean usdc should get fees or should it be this contract or that contract you know a call an ethereum call goes through a lot of different contracts so who exactly should get those fees so perfect attribution is impossible voting attribution should never be done what blast does is say in a in a contract call you hop through multiple contracts uh and you pay a total priority fee it gives the priority fee in excess of the base fee prata depending on how much gas your contract uses yes you can say this incentivizes people to raise um like make it inefficient to use uh like can we stop

this thing oh okay yeah and we can say oh yeah that incentivizes me to make horribly gas inefficient contracts but if that's the case then either number one the cost is borne down to your users in which case they will stop using your protocol and move on to something else that is more gas efficient or number two you are actually routed through in a composable protocol and because of that someone will use something else that is more gas efficient or number three you actually have the ability to charge someone that much more margin because people are sticky and want to stick with your product so no matter what it's actually fine yeah of course we want to encourage people to make things very very gas efficient but in crypto where everything's open source and you know

there's an incredibly competitive Dynamic uh on daps and infrastructure as well I think this will resolve over time so I think like the blast method of doing prata depending on how much gas you use is actually not a bad method and in fact the top daps on blast have received up to I believe 200 e in uh in in gas rebates uh because of sequencer free Revenue share and that's very significant yeah I wasn't aware of that that's very cool um in the case of optimism um they just did a really bad job so that sounds much better come to eclipse and that's we don't do salana I'm sorry um but yeah um I guess U how do you guys well you you okay

sorry um cool well we have one minute left so I'm going to give it uh to y'all to uh share some closing remarks on whatever you like um okay yeah so um so the last I still want people to learn a little bit about like web 3 protocol the decentralized HTTP um cuz that one I I feel like we need the real decentralization from the bottom to the end so right now all the DFS um are still have like two centralized component so if you're like the app or if you're just like website if you want to decentral your website just come to me and talk to me

thanks so um matter laabs is a value driven company and you can see all of those values written out in ZK creto uh it's not long if you have um if you've read the cipher Punk Manifesto I think it's probably about the same length maybe a little bit shorter I would encourage you to go read that um I think Poss possibly the most relevant point of that document to this discussion is the right of users to exit the protocol with low fees so we've been talking about governance we've been talking about um any uh issues with power or something like that uh

centralization and one of the things that ZK sync is really focused on is making sure that users can exit the protocol low fees anytime if they don't just if they don't agree with what what's going on so that's my conclusion fun fact the bonsai tree over there is $70,000 um eclipse is building salana on eth if you are an eth maxi and you've always been jealous of salana now's your chance to CH play around with something that's a that's bringing you know salana onto ethereum um as an L2 if you are someone who loves salana and you love transacting on salana now is your chance to try something where you can pay gas fees in E and you can have access to e liquidity uh but it feels like salana and hopefully one day it be faster than

Sol right now it's about a fifth the cost but hopefully we'll be able to bring it this speed uh over faster than Sal as well amazing thank you guys thanks for moderating [Music] [Applause]