📺 The zkEVM is Here | Alex Gluchowski of zkSync

foreign hey bankless nation welcome to another state of the nation episode we've got a fantastic episode here for you look something we thought they said was years away apparently has just come to ethereum

this is a key cornerstone in the the the efforts to scale ethereum into the future and that is a z zk evm compatible roll up we're going to be talking about that today david who do we have on the show today to speak about that yeah we have alex the co-founder ceo of matter labs who is the team behind zk sync and zk evms uh what we're talking about today or zk roll ups uh especially in the last year of 2021 of ethereum uh esp

that video of vitalik talking about how the internet of money needs to cost five cents has been a source of comedy for everyone who pointing out how high ethereum's gas fees are well when vitalik made that comment about how the internet of money needs to be five cents he's talking about zk roll ups he's talking about the zk evm uh and as of very recently the zk sync team the matter labs team has been able to put a zk evm on a test net which means we have a zk evm ready to go we just got tested out first

and so we're going to we're going to ask alex about all of these questions that we have as what's left in this roadmap what can people do now when is this going to be ready for users and of course the obvious and most fun question when token and just a few more conversations of that nature yeah it's a it's an incredibly important subject particularly this year this is the the year it's a um you know proof year for ethereum scalability approach we're also going to be talking about probably toward the end of the episode maybe at the front too about ukraine understand many of the um the matter

labs team is actually located in ukraine alex himself was born in ukraine i think crypto has a story to tell and a role to play there so we'll be talking about that david before we get into the episode though we've got a special psa from our friends at opolis all right opolis is awesome like straight up awesome if you want to become a self-sovereign web 3 worker you got to check out opolis what does it do david oblige is a dow that helps people work for other dows it's a dow

that supports a dows and so a lot of people would love to work in web3 would love to be a self-sovereign worker but we still haven't figured out things like healthcare and benefits and payroll and taxes and all that stuff and opolis is bridging the world of traditional employee benefits to people's that work for a dow so it's like a health it's like a employee infrastructure collective and so we all get to share our resources that we need in order to live a self-sovereign life when it comes to our employment but also still get all the

benefits that we need to make sure that our risk our personal risk our health care risk family risk all that stuff is taken care of so you can actually sign if you sign up for opolis before i believe march 1st uh which is you know you got some time but not too much time you actually get a bonus of a thousand work tokens which is the opolis dow token and a thousand bank tokens out of the bankless dow uh oh may first excuse me first even more time uh so if you are trying to get into the world of web 3 if you are listening to ryan and i yell at you every single weekly roll up about getting a job uh

the opolis might make this life easier this transition easier for you yeah totally i mean i i resisted becoming a um a you know self-employed individual for a number of years just because i got health care insurance health insurance through my my corporate employer and i couldn't let that go but opolis provides an awesome off-ramp to that they give you health insurance take care of your benefits all the messy stuff so go check that out guys all right david um gotta start with the question i always ask you any state of the nation episodes which is this what is the state of the

nation today around the state of the nations we are getting synced we're talking with the zk sync and we're getting synced with zk sync i really wanted to do like a uh that's sync not sunk right yes not sunk i really wanted to do like a zip metaphor because for some reason zk sync and zk's always feels like very zippy to me but i couldn't figure out yeah fast yeah but i couldn't figure out how to roll that into a metaphor but i'll just talk about how we're getting synced with alex from zk sync today awesome we definitely want to get in sync on zk

sync and that is coming up for you we're going to get to alex in just a moment but before we do we want to talk about the sponsors that made this episode possible slingshot is a decentralized trading platform that combines the performance and ease of a centralized exchange with the openness and transparency of defy slingshot aggregates liquidity from all of d5 in order to find the best price on thousands of crypto assets every token on slingshot comes with a price chart and trade logs to give you insights into the market's activity in real time slingshot is available on polygon arbor trump and optimism saving you from the

high gas feeds and low transaction speeds of the ethereum l1 there are no fees to trade on slingshot and any positive slippage is given to the users trading on slingshot is a social experience you can even set your chat avatar to your favorite nft or soon a slingshot 2099 nft avatar once you bridge your assets to polygon arbitrary optimism go to app.slingshot.financetotrade and use the chatbox to share your trades with others and find other tokens to ape into the brave browser is the user first browser for the web3 internet with built-in privacy and ad blocking to keep

you in charge of your digital footprint inside the brave browser you'll find the brave wallet the first secure crypto wallet built natively inside of a web 3 crypto browser web 3 is freedom from big tech and wall street more control and better privacy but there's a weak point in web 3 your crypto wallet the brave wall is different brave wall is built natively inside the brave browser no extension required which gives the brave wallet an extra level of security versus other wallets with the brave wallet you can buy store send and swap your crypto assets and you can even manage your nfts and connect to other wallets and defy

apps all from the security of the best privacy browser on the market whether you're new to crypto or a seasoned pro it's time to switch to the brave wallet download brave at brave.com banklist and click the wallet icon to get started arbitram is an ethereum scaling solution that's going to completely change how we use defy and nfts over 250 projects have already deployed on arbitram and arbitrom's defy and nft ecosystems are growing rapidly arbitrary increases ethereum speed by orders of magnitude for a fraction of the cost of the average gas fee when interacting with arbitram you can get the performance of

a centralized exchange while tapping into ethereum's level of decentralization and security if you're a developer who wants low gas fees and instant transactions for your users visit developer.offchainlabs.com to get started building your application on arbitrom if you're a user keep an eye out for your favorite defy apps or nft projects building on arbitrarily many of your favorite apps are already live with many more coming over soon you can find these apps at portal.arbitrum.one and you can bridge your assets over to arbitrom using bridge.arbitrum.io in order to experience defy and nfts the

way it was always meant to be fast cheap and friction free hey guys we are back with alex glicovsky he's the co-founder of matter labs matter labs is the team behind zk sync that's what we're going to talk about today the zk roll up that is pioneering a zk evm just went to testnet want to hear all about that matter labs is really uh special i think for bankless and for those who are decentralization maximalists because they've been uncompromising on

many things including open source values this is also why we're excited to bring alex on bankless alex it's great to have you on bankless iran thank you david it's an honor uh you know alex we want to start because we understand that um a large part of the team or part of the team of matter labs is actually located it has been located in ukraine uh how's everyone doing give us an update is everyone safe what's what's happening yeah so obviously everyone is uh in a

lot of stress uh because we have a really large part of it i mean in both russia and ukraine and we were lucky we managed to evacuate most of our ukrainian team in safe places mostly abroad uh before that happened because we are a security uh first company like the securities is the the paramount concern that matter of labs is kissing and uh we live by the motto uh only the paranoid survive and the rumors appear that the situation might escalate we just thought

we can take risks like the history is full of black swans and uh if um things happen you know that it's going to be really really difficult to do anything but it's really cheap to get a team on a few weeks vacation abroad so we decided we should go for this even if the probability is very low unfortunately though a lot of team members have family members back in ukraine and some of them are trying to evacuate them so yeah there is what is just going on it sounds like a very stressful time we

we are glad that um you guys were able to make some provisions and keep everyone safe during this time as well alex if you're okay with it we want to talk a bit more about ukraine understand you're born in ukraine you have some personal experience there also in the world of crypto of course you are here for reasons of values alignment with with what we're doing in the space and decentralization so i think we want to talk more about that maybe toward the end of the episode would you be okay with that absolutely we'll be having it okay awesome well let's do that uh first we

want to get into some of the recent events and news from a technology perspective which is this and this is very big this happened last week i believe the zk evm on testnet okay this is something i think the the crypto community thought was um years away in fact as recently as a year ago when we had vitalik on the podcast we talked about the future of roll-ups and scalability and he was very optimistic about

zk roll ups in particular zk evm roll ups uh evm stands for the ethereum virtual machine of course which allows all the programmatic d5 magic that people experience on ethereum um but he thought at the time that would be years away like years plural you know maybe three to five and here we are one year later and you guys have just released an evm zk roll up on test net was this a you know surprise development tell us about what was actually released

and how you were able to deliver it maybe ahead of what a lot of people uh thought was going to be the schedule for this thing uh absolutely it was not just with alex prediction but actually an opinion of a lot of adult crypto cryptographers in the world that this will take a lot longer to deliver because of inefficiencies of zero knowledge proofs for general computation their knowledge proofs are uh you know you have to convert your program into specialized arithmetization form like essentially in some mathematical representation of this

program and it only works uh well if your program is is fixed and well defined but with generic computation we have loops we have recursion we have all the turing complete flexibility we can build any programs uh and it just doesn't convert well like you have a lot of overhead when you're trying to represent this flexibility in terms of pure mathematical function you can imagine like huge polynomial mathematical function so uh and everyone thought well this has to

stay so we just have to find better ways to uh uh you know to be more efficient with proving with overcoming this massive overhead through maybe more efficient hardware maybe more efficient proof systems and so on uh but what we did uh was we just realized there can be uh there are shortcuts to make and they will compensate for this huge inefficiency in in interesting ways so what we realized is that during complete part

of the program uh is only gonna take maybe one percent of the entire computational power that we need while most computations will be spent on like heavy specialized operations uh which include hashes storage access and so on and so on so what we did we built a hybrid product which combines this flexible part with the rigid part and the rigid part is well optimized and we knew how to do this and

there was a number of uh breakthroughs also in the crypto cryptography around pros so this combination gave us this breakthrough which allowed to deliver uh technology much quicker so alex let me uh regurgitate that to make sure that i'm following because uh as much as i like to think that i understand zero knowledge cryptography uh it's not really my strong suit and i would imagine that's probably true for the listeners so you talked about two components of a zk evm that um uh are

really needed to have a fully fledged fully expressive roll-up that has all the composability that we all enjoy uh in d5 really the thing that really makes d5 power powerful is its composability and making sure that we retain composability in a zk evm on a zk roll up is a tough challenge and you talked about how there's the expressive flexible side and then there's this side the other side of things that are i think just always the same something like signatures or

things about a rollup that just like don't change but are necessary and rigid and always there for all the rollups and so you've been able to compartmentalize these things and the things that don't change you've been able to extremely extremely optimize and then that's been able to for you to uh separate or compartmentalize away from the flexible things and allowing you to maybe tackle these two problems separately simultaneously is that a fair way to put this uh yeah you can think of it as imagine a

computer which can which has a cpu which can execute any program and then you have some specialized micro schemes for digital process like a let's say a gpu for specifically for graphic processing it can do one thing but it can do it really really well and much more efficient than the cpu and cpu is very generic can execute any program arbitrarily but it's it's slower so this is the kind of analogy we have here and for for as far as composibility is is concerned

if we can express an arbitrary program an arbitrary during complete program then we can do anything evm can do so we including compostability compressibility it's not some unique special case so what what what's really important also in this regard is that we should be able to execute programs of arbitrary length and it's it's a little bit hard with snarks because they have a limited proving capacity but we use recursion which was one of the

breakthroughs i mentioned where we can combine we can have a lot of um circuits of limited size but we compile them together in a line and then we recursively prove them uh with multiple layers until we get one single snark which attests to the validity of the entire execution of the block so with this particular design of the zk evm would you say that this is uh like the end state of um or the end goal of a zk evm or do you

think that further designs and further iterations will actually change the the basic fundamental primitive or do you think that this is the the design structure that we're going to go with for the foreseeable future i think the design structure the uh the architecture the big picture is going to remain roughly the same there are a few things which i expect to change so first of all i have to clarify that when we say zkvm it doesn't really mean that the we execute evm byte code

what we execute is the virtual machine byte code which is optimized for zero knowledge pros and we have a compiler from solidity and from viper which and and from like uh in intermediate called like any any other languages uh going through llvm which is a very um optimized very well known compiler platform um compiling into this zkvm bytecode and then we execute this efficient bytecode uh so llvm actually allows us in the future to uh to

integrate um other like most modern languages from rust to c plus class to golang to python whatever all of them have front ends for llvm and then we we can reuse those front ends and then use our back end to produce the the resulting byte code so if people will find other languages uh to be more useful for specific tasks in the future then they will be using those um and generally i i also think that in the in a further future we will switch from

pure snacks based on plonk to uh to something fully transparent uh that doesn't require a trusted setup but generally the architecture i expected to remain roughly the same well opening up the design space for developers to uh to make it life easier for developers is of course crucial for building out ecosystems so we're going to definitely talk about that later in the show one last question before we get into some uh back to some of the basic stuff about the test net is this zk syncs breakthrough if breakthrough is the right word is this

zk syncs breakthrough or is this uh just kind of shared knowledge like is is this your guys's victory here with like what you the design pattern that you've discovered or how is this uh design structure being shared around just the greater zk roll up ecosystem uh i think it's a general approach it's it lies on the surface so we independently came to it uh from uh guys from from stucco for example but we are building on the shoulders of giants we we are using the pattern of tinyram

which was uh introduced uh actually by the founders of of stockwell uh a few uh years ago uh but then uh we we're using this like we invented this approach of combining the uh the heavy and and the flexible parts together i think independently from from other teams but i can confirm because i don't know what their decision process was okay alex so uh tell us uh what is here today in the here and now with the zk sync 2.0 public test net release which

is a z key zk evm on test net so what is this uh the the subtitle is the zke evm has arrived the first evm compatible zk roll up on ethereum's test net so what does this look like in practice how mature is it and then you know after this the question we're about to ask is uh when mainnet but first tell us what's here now uh well what you have now is a separate testnet where you can deploy smart contracts written in solidity

and you can interact with them using web3 api which you normally use with other ethereum labs so it looks very similar to other l2s and and sidechains that people are currently using for scaling just that it's based on zta vm technology so how mature it is is it we the this announcement is about the opening the testing to the public it's been running in private for quite some time we had teams building on it

and we were collecting feedback uh and fixing problems so we wanted to release something to the public that is more or less mature and people are not having children's diseases so alex what are the goals of this test net are you trying to get developers to like try and break it or are you just allowing them to get a feel for what it's like to build on nck evm what are you looking to get out of this test net it's well which is offering people to to start building because the interface won't change the interface will only be

extended by certain things uh which which are clearly documented in in our docs uh but the basics will just work so we wanted to show that you can take your existing code and deploy it without modifications or with only minor modifications and it's really important that you can reuse your code bases across l1 and different l2s and you you don't have to worry you have to rewrite everything completely from scratch you can rely on existing security patterns which you

understand well you can use code bases that have been heavily audited uh it will all just work and yeah we intend to release the the mainnet as soon as uh the testing shows mature and you know like everything around uh the uh full specification is fully implemented so uh alex can we talk about that for a minute so say i'm an uh a defy app developer and i have something i've i've wired up in solidity it's working on the evm right now maybe that same uh

solidity app i've now ported that to like polygon for example their you know their proof-of-stake network maybe i've i'm trying to port that to you arbitram and optimism as well if i take that code and i try to deploy it on the sync test net how much modification is is really required there because the understanding is there there's like degrees of the modification that uh that is required and could you contrast that with um

maybe uh what something like starkware requires a a developer to do to port their application absolutely so the degree to which you have to make modifications depends on what your program is using uh if you are if you have just generic code most likely you will not need to change anything we support most things most uh primitives including creating contracts create create two uh you know like

hashes all the ethereum hashes which by the way translate to the same you know we actually like if you if you use hedgehog it's gonna produce ketchup hash if your shot of five six it's gonna use uh shot of five six for not replacing them under the hood as previously uh was thought of that we will need to um so uh there are a few things that aren't just not supported yet so if you rely on them you will either have to wait or you will have to certain that circumvent though those parts of code replace them with

something different those are uh some certain cryptographic primitives so we don't support pairings yet we don't support uh mo mode operations uh although the compiler will soon uh also uh bridge that functionality and um we don't support native if sending and like and receiving so you have to wrap if as a um your c20 token for now uh the part um

yeah so apart from that everything should work uh and whatever is not working is clearly documented hey could you contrast that with um starkware so so far a lot of that sounds similar to some of the optimistic roll up style evm chains that we've seen but how does that compare with uh starkware i mean i know i understand they have a different you know programming language they have a completely different programming language which you have to use from from like you have to actually rewrite your code from scratch in that language

um they have a transpiler but i i'm not sure to what degree it's uh it's functional and i my understanding is that it will produce chirocode which you will have to then manually review and adjust you cannot just like keep your code in solidity but maybe i'm wrong here but in any case the input for stacker compiler is called in cairo whereas for us it's evm code and what's really important we support the same web3 api as all the other depths you

most likely will not have to change your interface you will be able to rely on the interface and you will be able to integrate services that or like have integration of services that already implement this web free api uh like uh ether scan and the graph and and you know like a bunch of tools that work with evm um today on mainnet and then on different l2s uh and they will just work out of box which which is not the case for uh cairo as far as i want to just reorient to the

uh the user's perspective here so are are you saying that it could come to pass that your your favorite defy application say an ave a compound you know a maker uniswap you know same interface you're using on the existing ethereum main net uh with with high gas fees of course and then one day you have the option with the exact same interface to use zk sync instead and it's all you know evm zk under the hood and suddenly your gas feeds drop from like dollars to cents

and from a user perspective you know it's it's the same kind of ux experience that i've always received except now there's there's far lower gas fees is that the sort of thing that we can expect i mean same block explorer same wallet same user interface just lower gas fees or is it going to be a bit more uh incremental and clunky than that no it's it's actually the the exact same user experience so a big part of that is that we allow people to reuse wallets so any ethereum wallets works natively uh

and the the keys are actually kept in the wallet so you like when you sign a transaction you you you just sign it from the wallet uh so that that's really important so yes this allows for the migration of the entire ecosystem uh on like you know on us so you you extract the ecosystem uh uh by default you don't have to write all of that from scratch i think uh arjun is maybe one example of this with people are they using they're using zk sync i'm not sure if they're using the zk evm under the hood

but um you know arjun is kind of a wallet wrapper that's that's generally been on ethereum for most of its life and now it's it's starting to spin up wallets on zk sync is that an example of this are they using the same infrastructure technology here uh well arjun is a really interesting uh example of uh so argent is on the mission to allow mainstream adoption for ethereum in general because they solve the problem of security and usability you don't have to mess with your seed phrase which is hard for me

for most people like they they don't understand security of it they you know they they have never been hooked in their history of art uh so right now uh we integrated argent with the kissim version one where people can make transfers can make trades and can interact with existing uh d5 contracts on maintenance ethereum through batches with liquidity providers uh but it's also gonna be uh used on version two and for the users the

experience will remain the same you will just have more functionality because you will be able to interact with contracts one or two also directly uh so what when i say the experience is going to be the same it's actually not true because the experience is going to be better so there are a lot of things which will be possible on layer 2 which are currently limited in ethereum for example account abstraction so it thinks arjun for example will work much better on zika sync than it works on ethereum

it's going to be a lot cheaper and easier to maintain the code um you will not need real layers people will be able to interact directly with uh with the peer-to-peer network to propagate transactions uh people will actually be able to pay a native gas not in in justin eath but in any token they they transact in uh so it's it's gonna open a lot of possibilities and and also for for other um you know use cases other dapps that

uh that will do interesting things with account abstraction with meta transactions etc this is why a lot of people in the ethereum ecosystem think that the zk evm zk roll ups are going to unlock the golden age of ux for defy which is something that we've just been missing basically since the entirety of defy's existence and alex um one of the properties about optimistic roll ups is that they they they both both arbitrage and optimism just boast about how you can just take your code

and immediately deploy it on an optimistic rollup and that's one of the benefits of optimistic rollups is that it's just a very low touch transition and that's what i'm hearing coming out of you in zk sync is that it doesn't actually take all that much work to deploy your code on zk sync which is just phenomenal because we already have so much of d5 built out on the ethereum l1 as soon as just being able to copy and paste it and put it on a zke rollup uh is oh it kind of almost sounds too good to be true um but i what i do know uh that i want to

check with you is that while we do have this benefit of just being able to copy and paste and put it on zk sync the developer tooling i think still is going to be something that the zk evm ecosystem needs to build out because the developer tooling for building d5 on the ethereum l1 has been extremely robust because that's what we've been having that's what we've had for the last six years is this a an important conversation when it comes to building out the zk evm ecosystem

how much rewriting and just like rebuilding of the developer tooling needs to be done to truly enable developers to unlock their full potential for building out divi defy on a zk evm yeah this is a great question so we we have to split developer tooling into two categories one is the tooling around the code itself and the other one is is general tooling around smart contracts deployment and so on so the second category is completely covered by our support of web3 api so