Reversing the innovation flow: having rollups lead the innovation in the Ethereum ecosystem

Transcript

[Music] hi guys after a bit of uh artistic unwind that we had in during the previous call we are getting back to the technology and let's greet and meet Alias tazartes who is going to answer a very important and interesting question can rollups lead the innovation in the ethereum ecosystem welcome hey guys can you hear me well all good nice um so we're going going to talk about um the governance in the ethereum ecosystem and precisely also can how can rollups lead this movement as opposed to the approach that today is on ether L1

where you have a more um top down approach or L1 Le approach so ether has this interesting property of being or aiming to be the world computer and when usually you can Define layers in like two words you probably can't with a THM you have this funny tweet where the person was saying ether is so complex you probably you probably um cannot just summarize it or just do it in in a few words so etherum aims to be the world computer and as such you would probably want to upgrade it as you go as you learn and finally aify it but uh this world computer it's going to undergo a

lot of transformation until it can be stable and can be completely um rified so today we're going to see what is the government's flow of the therm ecosystem the limits of this current flow the multiplication of the clients and uh we're going to see that this challenges the the status quo in governance the main risk which is evm fragmentation and the way for all um let's just start by defining together etherum it's just the community run technology that powers the cryptocurrency ether and the thousands of decentralized application the world computer and what is governance it's just the systems in place to allow decision to be made so usually I don't

know you can think about executive team board of directors shareholders project manager and ethereum is um a bit more complex um body so if it's Community run who governs uh if no one no one owns thether how decision made and I thought this this tweet was funny because Peter is the tech lead of gu and danad is the lead researchers one of the lead researchers of ethereum and they were talking about how a decision is made or is deemed relevant and Dad quite um sassily was saying the decision is his but it's not so far from the truth so what is actually do actually who knows here how a decision is made on ethereum like a protocol change okay who has

heard about the EIP word how who has heard about the rip word word okay interesting so direct CH came up with this very simple um model it doesn't have to be so um concrete but it's just the idea that etherum is a is characterized by set of values they in turn get um transform into a Clear Vision of what you would ideally want ether to be then this this is in turn created into a road map and then this road map is implemented by clients so if we wanted to list The Ether Valu I mean they're in the air they're not always named but you would get oops okay I'm standing in front of the I'll go there

you would get decentralization transparency openness censorship resistance sustainability and so on and so the vision is you would ideally want a secure permissionless chain where transactions are fast they're cheap they're secure maybe private and then user accounts so Identity or even credentials is quite easy to manage so social recovery Safeguard all of this is also so like present in the spirit THM so the values we get the vision and then we get the road map which is bit more chunky so this is this idea that okay if we want 10K TPS Subs scent how do we do it and it's so it looks like etherum did the the approach of going for rollups

instead of full shouting for now and then um some heavy um emphasis on ZK so this road map is basically you would want rollups uh you would want decentralization so low requirements for your healthware and then lots of ZK ZK everywhere aggregation privacy all of it um so we got values you have a vision now you have a technical road map so how do you actually enforce it in the real life is if you were a country how do you actually pass the laws and who is enforcing it and this is down to the execution clients team so you have to think very concrete example of this is there is a hard Fork every now and then

so every six months or one year in ethereum this is how they call the ability to upgrade the protocol and um they have they have things that they want to put in it so next hard Fork is going to be Petra prag and Electra and they think they think to themselves okay guys we need one way to improve ux on ethereum for EAS so for user accounts so they have this idea in mind and then from there they're going to choose an EIP so etherum Improvement proposal and stick it in there so you have to think of uh of slots like uh every now and then there's a timing there's a good timing a tempo for a good Improvement and then the Improvement is actually a specification that you write so this year they wanted to improve ux because they think that uh the the current ux of

EAS so user account so metam mask is not so good so there is a this need to to insert the EIP so how do you do a EIP you have an idea so you think okay transaction should be cheaper you in you transform this idea into concrete code or ideas for the specification this is called the EIP then you need to Champion it so you need to push it and then the core devs of the etherum so G neine bzu Aron reth ejs um they have a voice and they're going to debate in a Discord call it's quite a it's not so formal as a voting for a president but they would Gather in a call or a zoom and they will say okay guys I think this EIP really Nails the the goal of changing this so current

example is a EIP 772 they want um this achieves better ux we don't have to go into precise deal but this is how it works so you have the vision the values the vision the road map and then some precise law or like EAP will fit into this road map and enforce what the road map is about so eum wins this is a Flawless process this is uh very well constructed and what is the problem first problem is you're making changes that take time and need governance um um like uh you need everyone to buy in but you don't know if this will last so one example is that um there was a EIP to add a compression

function on etherum and actually no one used it ever so you would get this this has been used a thousand times since it was inserted uh like almost eight years ago and there's Bill like millions and millions of transaction so this is really a like a loss of resources loss of time and it's so weird like I don't know it's it's it feels it feels a bit it feels a bit like you would probably want to try something out before putting it and trining it in your protocol so in my mind I would like even for countries it's weird that we can pass laws without ever trying them I read that some countries want to try to have laws be passed in specific cities they try them for two years and if it works they put it for the whole country I mean this feels like we're dealing with common

sense or um first principles but it's really not easy because when it comes to governance and code and the etherum prot everything is a bit more complicated the other way that this process is not perfect is EIP 3074 we don't have to go into very strong details but you have to imagine that this EIP was imagined four years ago and it was scheduled for inclusion this like a a month a month from now so it took a few uh like some time to mature in people's mind they submitted it and it got scheduled for inclusion four years after so what is the chance of it not being relevant anymore well quite High because it got scheduled for inclusion by the corev team but the teams that are working not on the

protocol but are working on product so you have to think smart wallet teams they said wait guys this EIP is really against our vision for the for the etherum account abstraction uh experience so when you when you hear the word account abstraction or inm you have to think the difference between metamask which is which is eoa and account abstraction you have to think more about a a wallet on your phone you sign with your face ID you never have to to interact with 12 um 12w seat phas this is more like Web Two experience and like social graph uh privacy like all those things baked in or at least a place for it and so after this e got voted in so it was supposed to get coded and get included in the test net it was actually superseded by

7702 so at the same time as the cevs voted yes for this the community pushed back and they have to they had to go back on it so it's it's it's it's a bit weird ideally you would want the push back to come before first of all second of all you would need to have good numbers on why the push back is Justified or why the law is Justified like everyone is is is thinking out of um suppositions or just uh they're they're they're comparing very abstract uh eips and they're they're comparing the ups and down uh expost so they're thinking six months in advance one year in advance what will this do um and because ethereum adopted the

rollup Centric road map now you have a bunch of rollups which are faster to to adopt to like adapt faster also to um be able to change stuff and they represent a good um a good field for people to try out new ideas um and this Fields a bit like the this idea of having cities that can try new laws before the whole country adopts them so so this is why some of the people in rollups came up with this idea it's not ether Improvement proposal anymore it's rollup Improvement proposal and it's more like a bottomup approach so the good example is the cave team they're working on account obstruction

and they wanted to they wanted to imp they wanted to add to Ether the ability to verify a face ID signature or like secure Enclave so to make it cheaper and and more available to anyone so they proposed the EIP 72 to it's hard to say all the eips if you go to a zoom call from allar devs you will hear them like use all of the numbers all the time they're like oh yeah this is a 5678 no no it's a 7 771 and it it's really you cannot understand anything it feels like everyone is reciting Wii codes but it's actually low proposals uh so they made this into a RP and while this has not been adopted on Main net so you will not

be able to verify easily face ID signatures on Main net L1 you can actually uh verify face ID signatures on more than four rollups so I don't want to make mistakes but I think polygon has it probably Z casing Kakarot B it's on base seola and I will need to check for linear but the main idea is that a rollup is inherently more flexible than the main net and this is this can be good and then you can probably imagine then the merge or uh dun sharding which was the abil the the a main main change for the etherum protocol they're not so they're not so adapted adapted for rollups because uh they're like Hefty and they way to

rethink the infrastructure of ethereum but for smaller things like pre compiles erc's transaction type uh AA infrastructure it makes sense that you would try them out on rollups before uh trying to implement them on mainnet so can devs do something I mean this this feel like from the beginning of this talk we're dealing with common sense or at least first principles and people in theum are very very practical and smart and they often go for like path of if not least resistance the path that makes sense so they try to do things that make sense and so if it's so simple why is no one doing it why is no rollup pushing pushing for a for like very brave or at least uh spearheading

uh Innovation there's actually one unknown um limitation which is that so who here knows a bit about the the client team you have to think of the clients team they are kind of the guardian or the they safekeep the software so when you think about ethereum you have to think that there is software behind and there are teams that implement it it so happens that ether has many uh clients developed so there's G bzu so G is a team is supported by EF bzu is pretty much for those who are Dev is just you write the eum protocol in goang then you write it in Java then you write it in Rust and then you have many clients and then those team they are also part of the governance prot the the

governance decision it just so happens that the GU team is dominant in terms of Mind share or market share or tvl and so if you look at it um so if you know the rollups name so arbitrum optimism scroll those rollups are based on Forks of gu what does that mean in practice it means that in practice if they want to innovate on the protocol level they have to write code on the GU client that will just Implement their idea so I don't know they want to increase uh they want to have they want to add a new type of a signer they have to write the code on on on gu and in practice what this means is

that they need to either make their own implementation further away from um the GU implementation so they're going in unch territories or it means that they have to ask the get Dev to do it for them the get Dev not only do they have limited bandwidth they're also very conservative which makes sense and the devs that uh forked get so imagine the Arbitron team they want to implement something new they're going to be extremely careful because they have hundreds of billions on their chain or maybe dozens I I don't remember the T CH but it wouldn't shock me if it's 100 billion and any change is EXT extremely risky it's not like they have client diversity or they don't have

the leeway so rollup team don't want to Fork away from gu G doesn't want to implement the innovation no innovation and this is changing actually now that you have more rollups uh clients so more rollup stack you can innovate once more you don't have to rely on someone else to say whether or not you can innovate so three example of this is EF so ether Foundation sorry there's a lot of acronyms etherum Foundation created their own specification so the they wrote etherum in a very simple python uh repo that is open source so if you want to say okay from now on I don't know etherum um I'm going to change this uh protocol

thing what I'm going to make gas 10 times cheaper why not you you're rollup and you want uh users to have a subsidy so you make some operations cheaper you go there you for the repo you try out you can experiment very easily and it's quite easy it's in Python it's very fast in one day you can Implement an EIP play around with it ask some people to have a look at it so it's becoming very much accessible re is a implementation of etherum protocol in Rust and they released many things that make it very easy for people to experiment as well so alphanet is a rest based Oru optimistic rollup framework and they can extend the evm very easily as well so you have to think that if you like to Tinker around this like reduces the the friction and

the hardship by like one or two orders of magnitude and Kaka is a good example it's my Shilling moment uh it's a highly it's a it's a zkm written in Cairo but because it's so because Cairo is so high level Cairo is a ZK DSL it's uh this allows the Kakarot to be a ZK evm as opposed to the two above uh clients which are just EVMS um this also means that because there are new client uh implementation you can play around and you're not limited to just one implementation so new actors are less dependent on get so if you're on arbitr and you want to create a new transaction type you can play around with the new implementations but the main risk and I

think this is what I still don't have an answer about is evm is strong because of the mode of the network effect everyone knows etherum and if you develop on solidity the strength is that everyone knows what you're doing and you can choose a chain kind of like a commodity or not even a commodity you have a trade-off so if you're fine with a somewhat reduced reduced security but very fast transaction and maybe low price High liquidity uh strong Community you might go on base if you want strong security you stay on L1 there's so many trade-offs and this is what makes evm like the the leader of the Mind share of the current market because the base layer is not opinionated and the all of

the tools are reusable so let's take a real example after RP 7212 so you're Dev and you want in your daap and you want to be able to use face ID this is quite quite normal maybe you want to have a checkout in your ticketing service so you write your solidity contract and you put in the solid contract Now call the face ID verifier you don't know which chain supports it or not and you you cannot know at the time of uh deployment so maybe you choose a chain that supports it but this is is like the first step in evm fragmentation and this might continue so my M main worry and the thing I most certainly don't have an answer about is how to limit as much as

possible evm fragmentation and how like evm fragmentation really goes against the client diversity um atos of e themm and also the Moe my main concern is more about the mode so evm ecosystem mode this idea that everyone knows the same language everyone can deploy on any chain there is no vendor loock in there is um Global governance this go down as fragmentation increases and so I genuinely don't have a like a solution governance is really hard I've joined I've started to join roll calls andar dev's calls so the governance bodies of ethereum and you it's almost a full-time job to follow what is happening it's a second full-time Jo job to understand what is

being discussed so every single area is a rabbit hole that could take a whole life so so there is a rabbit hole about how to design good mols how to safeguard decentralization this would take probably one year of engineer to even get up to speed on how to understand these things and this is one example out of 100 then there is new way to store data so if you were at the fireside chat with vitalic he mentioned hexar Patricia merri and then everyone's like yeah sure okay he by on the top of his mind made calculation on the maximum size of the proof size of proving that you are included in the slot of this tree but this is not accessible to most people the veral tree discussion this has already taken two years it will take two

more and you expect rollups team who have limited bandwidth to jump in and express a opinion is quite hard so even though in first principle you would want the governance of eum to start in rollups because they are essentially the closest to the users and they also can they are also the closest to the devs that are building daps as opposed to building protocol um this is quite hard so my view and it's quite utopian is that most improvements should start on rollups actually um I hope that this works out for one two reasons one is e THM needs to aify and because it needs to aify I

think it's fine to let rollups innovate and whatever Innovation survives gets adopted on L1 like this also feels good um um also rollups have they have like small community and they have good feedback loops so I think this also approaches and one thing maybe doesn't apply so much but there's a v v Tre upgrade and maybe I'm wrong but I heard the Matt from the alev that said that um this will take 18 months of L1 uh cor devs kind of um this is a major upgrade so kind of the merge and he I think that if L1 devs are busy they they will have stuff to do so it will leave

us the rollups kind of the range to innovate so it's not so bad it's like uh they we have a saying in French I don't know if it translate but is while the cat is not here the mice dance so this is what I think and then my second belief is that evm fragmentation will must be limited so I actually talked to YF today and I don't know so there are some challenges so if you have RPS it means that rollups innovate then you have this this question what is a rollup and can you actually form a cohesive body of rollups that form like a union like are you thinking European Union confederated states like what is this so if a rip gets adopted how do you force base sepolia to Bas sorry base to adopt

it do you like how can you do it in L1 it's more simple because there is the spec and you just you have to go through the cves and they are kind of the last Gatekeepers of what gets into etherum in interestingly I don't know if you guys realize but eum is different from other places or Dow or I don't know but it's not because you hold the E that you can have one vote in what happens on eth it's quite not if you hold one eth you hold one eth and you trust that the core devs who build the client have good philosophical views on the world they have good talent of devs they have good um way to create an EIP and they have good ways to communicate why the iip

should be merged you have to think that because you are ethereum holder you believe in the values of ethereum because you believe in the values you can see that the people who like push the vision are good you can look at the road map and participate so it's not gate kept like you can enter and participate to any c whenever but theend at the end of the day is the devs that vote so this is it might look like a democracy and I don't mind that it's not but it's not and so where rips face the same problem like is this a democracy probably not RP as it stands is uh I can Kakarot can go and say okay I have a new idea um I have a new transaction type

where I can batch many transactions uh I will I will create R and um I will uh adopt it on my chain unilaterally uh everyone agrees everyone says no I say I don't I don't care it still happens like no one can force you not to adopt RP or adopted there are two safeguards for RPS one is that you would get Buy in from L1 that from the time you push a RP the L1 should not do its own governance uh piece so a EIP that clashes with it because this would be pretty bad so you have to think RP is has at least the

guarantee that L1 main net or eim will not go against it um second guarantee is um if the RP has Buy in from other chains um you also know that if they they want adopt the same goal with different means so you have to think of I don't know uh Native account abstraction on zync and stocket does not respect any standard way of doing it and 4337 which is the etherum account abstraction respects a certain standard so the second buy from RPS from rollups is that you should respect a standard and if you agree with some certain governance decision at the rollup level you should fall in line with that

specific implementation or not implementation but at least the spirit of it apis and so on I think in a few years this my uh conclusion is that um rollups will be ZK and uh you will because of the Pro system and also like standardization of maybe processes I I I don't know but I see it as um everything everything is narrow because every governance is on L1 now governance design space is huge because it's all on the rollups and it's going to narrow down later on in a few years when um the ZK like infrastructure of rollups and Main that converges to what is called like I don't know this type of

type one ZK VM or even just ZK VMS in general so prove everything zik snack everything I think this is very PR present in the his ecosystem let's it thanks a lot thank you Alice uh we still have a couple of minutes for quick question uh if any so maybe I missed this because um maybe I think I'd come in after you had started but what's the current stage of kakarat and um and the company and your protocol so kakar it's a it's a for those who don't know it's a ZK VM based

on the stocket stack and it's on testet but for example uh when uh EF when um the EIP for the EIP 374 the one I talked about in the beginning was adopted uh on L1 so scheduled for adoption we actually implemented it instantly on the test net and I think that was a good example of um one a very cool thing where rollups can actually kind of front L1 for things that are scheduled for inclusion which is very nice because you can try some updates really fast but it's also the first example of a fumble we fumbled because we adopted the EIP the first week it was in scheduled for inclusion and then it got cancelled so this is one

example of what you don't want to do when you roll up uh or you have to be careful about is this very the Fine Line you're working on when it comes to governance you want to be reasonably sure that what you're doing will either end up end up on L1 or will not be clashed by L1 and you also want all the rollups to follow you so this is this kind of tension you have to play maybe there is incentives may maybe there is not maybe there is this foro idea if one rollup adopts the new killer feature everyone will follow so there is these tensions that you have to play around and um one thing I will say is we need the processes probably in um ether L1 governance to become more if not transparent more smooth or at least have

more input and also for rollups we need this push for uh yeah streamline and so on thank you yeah I just want to quickly ask based on your experience and acute knowledge with the different rollups and governance what do you currently find to be the best method of incentivizing and rewarding developers and activity um for different chains so for a chain to incentivize devs so devs that build smart contracts I genu it's a very hard

question I genuinely don't know I I found that for me my favorite thing is um is a ENT entrepreneurship programs so like entrepreneurship Studios buy rollups I think this is when you're rollup I will say this you want to maximize the number of teams that build on you so entrepreneurs that build businesses and from there you can have it because you have M share so many devs participate to agaton maybe your token is expensive maybe they Deen spending on your chain but at the end of the day you want people to come and build um so for this probably a mixture of a smart grants program but you don't need to write checks like so many of them retr funding entrepreneurship program and um good air

drops thank you thank you Alis thank you very much thankk you [Music]