Super Scaling with StarkWare

Transcript

hey bankless nation it is time for another estate of the nation super excited about today's episode this is super scaling with starkware we have two the two starkware co-founders on we have ellie and yuri of starkware we're going to get to that conversation in just a minute but david this is the time to bring on starkware i mean this is uh zk rollups rollups in general are uh are starting to really take off i think in a big way starkware is rolling out some crucial tech why is now the time to have starkware on to tell us about the scalability strategy for ethereum and blockchains in general yeah i recently tweeted out that layer 2 is here it's just not evenly distributed yet and some components about ethereum are super

scalable dui dx is doing more volume than coinbase immutable it just released it's a gaming powered gaming focus nft platform and it's already like throwing out the the transaction throughputs the scalability that we all want to see on ethereum is being built in these different corners yet that is the first version of starkware's products dark x stark net is also coming soon all these conversations are are happening all at once and definitely in the moment of where like crypto twitter just absolutely blew up over ethereum scalability it's a fantastic time to bring on starkware to talk about how scalable zk roll ups can actually be when put on top of a decentralized permissionless settlement layer that is ethereum yeah i feel like

uh it then the knowledge isn't evenly distributed either right it's like we're doing our part but a lot of this technology is uh is hard to understand and so part of what bank was trying to do is break this down into mental models we put together the uh the modular modular blockchain podcast you wrote an article about that david would refer listeners to that if you want to get a high level overview but we're now on a journey towards like unpacking these zk scalability type solutions and starkware is definitely one of the leading ones one of the ones we are most excited about in the space so gonna be a fascinating conversation stay tuned to for that also some quick announcements we dropped a podcast with visa visa is going all in on crypto in the crypto space other fintechs are following that

was an episode we just dropped on monday david any uh sneak peek on that app uh not only visa but also anchorage which is really the emblem of the d5 mullet you got visa in the front you got anchorage in the back what is it like to completely redefine what it means to have fintech fintech based on crypto rails not central banking rails so all of that and more in the podcast that came out yesterday on banquets there you go d5 mullet thesis right so the mental models just keep flowing out of bankless guys this is uh this is a crucial stuff for you also we need to give a shout out to opolis they're doing some cool stuff these days they are sponsoring this message i tell you david i haven't i've heard i've heard this so often like somebody telling me i'd quit my

corporate job my boring corporate job in a heartbeat but i can't because i'd lose my health care this is of course the us i know other countries don't have these problems but this is how they keep you a wage slave this is how they keep you a member of uh the corporations right you might want to go where your heart is go towards a dow go where the opportunity is but you lose benefits opolis is an organization that services dows that helps solve this they do payroll and they do benefits for daos and for self-sovereign workers we think the future is moving from the gig economy to the ownership economy towards self-sovereign workers and opolis provides provides benefits for those anything else you want to say about

opolis david yeah this is really the infrastructure that we need to go fully sovereign and not only when you sign up for opolis can you start to have health care with your other dow workers right these are all the same health care albos is like a co-op you guys all get to like choose your health care together we all get to be insured together and also rates go down as the organization grows up gets larger but if you sign up for opus you get a thousand work tokens and a thousand bank tokens if you sign up by the end of this year uh and so the work tokens is the tokens of the opelus network again it's just just like every other dow if you use it you become an owner in it uh and so opus uh making sure that all of the infrastructure

needed to be workers in the dow decentralized world are there so that we can actually sustain this ecosystem going forward guys there's a link in the show notes if you want to get plugged into that bankless dot cc slash opolis david i want to start with the question i ask you in every state of the nation what is the state of the nation today sir i've done this one before but i'm doing it again today we are building and the starkware team has been super hard at work building out zk tech for everyone on ethereum to use as we go through a weekend of drama and fud uh the only answer to really any of that shenanigans is to just keep on building uh and that is exactly what we are doing here on the state of the nation we are building today i want to

find out from the starkware team if they are abandoning ethereum my guess is their answer is no but we will get into that in just a minute before we do we want to thank the sponsors that made this episode possible arbitram is an ethereum scaling solution that's going to completely change how we use d5 and now it's live with over a hundred projects deployed gas fees on the ethereum l1 sucks too many people want to use ethereum and it doesn't have enough capacity for all of us and that's why teams like arbitrage have been hard at work developing layer two solutions that makes transactions on ethereum cheap and instant arbitrary increases ethereum's throughput by orders of magnitude at a fraction of the cost of what we are used to paying when interacting with arbitrary you can get

the performance of a centralized exchange while tapping into ethereum's level of security and decentralization that's why people are calling this ethereum's broadband moment where we get to add performance onto decentralization and security if you're a developer and you want to save on gas costs and overall make a better experience for your users go to developer.offchainlabs.com to get started building on arbitrum if you're a user keep an eye out for your favorite defy apps building on arbitram many defy applications that are on the ethereum layer one are migrating over to layer twos like arbitrary and some are even skipping over layer ones and deploying directly on layer twos there are so many apps coming online to arbitrarium so go to bridge.aubertrum.io and start bridging over your ether or any of the tokens

listed and start having a d5 experience that you've always wanted living a bankless life requires taking control over your own private keys not your keys not your crypto that's why so many in the bankless nation already have their ledger hardware wallet which makes proper private key management a breeze but the ledger ecosystem is much more than just a secure hardware wallet ledger is the combination of the ledger hardware wallet and the ledger live app and if you're used to seeing all of your crypto services and favorite defy apps all in one spot ledger live is where you want to be not only does ledger let you buy your crypto assets straight from the app but it also hooks into all of the defy apps and services that you're used to using ledger live you can stake your ethan lido swap on dexes like paraswap

or display your nfts with rainbow you can also use wallet connect inside of ledger live to connect to all the other defy apps that keep coming online defy never stops growing and the ledger live app grows alongside with it so click the link in the show notes to see all of the defy apps that ledger live has and stay tuned as more apps come online and if you don't have a ledger hardware wallet what are you even waiting for go to ledger.com grab a ledger download ledger live and get all of your defy apps all in one space guys we are back with the starkware team pleased to introduce you to ellie from starkware co-founder of starkware also yuri co-founder of starkware gentlemen great to have you on bayclass how you

doing today hi guys thanks for having us well we are super excited to uh to dig into everything that starkware is doing in this space and we're excited because um this is really this is really kind of the scalability answer i think to crypto and to blockchains and and you guys are are really delivering it and ramping up the delivery of it so we're going to talk about starkx dark net everything that you're delivering but first we want to congratulate you on this massive series uh c fundraise so there's a 50 million dollar fundraise a 2 billion valuation congrats on that how does it feel to be on the other side of this fundraise well it's uh you know it's uh the the

series c was done with a sequoia they've been our investors and close collaborators since our series a going back to 2018 so we've known mike and the sequoia team well and that made for a very simple and uh easy process so regardless we're happy to to have this sort of uh milestone behind us i want to say that it's better to be on this side than on the you know site before that of course of course you guys there are so many different things that we want to dive into with uh everything with starkwear but first uh you guys have a lot of things that are named stark something and so at the start of this podcast we want to go in and actually label and define these things because

we're going to be referring to them more and more throughout the rest of the podcast so there's starkwear there's starkx there's starknet can you guys define and delineate between these three things for us sure so first of all starks start starts were co-invented by uh two of my co-founders uh michael bryaptev who was our uh chief architect uh and got his phd under ellie uh before uh co-founding starkware with us and with alessandro chiasmus they're co-inventors of starks that's the zero knowledge protocol that we're using in our software stick uh starkx was our first product and this is a permissioned uh standalone

customizable scaling engine that runs over ethereum and this powers immutable x and surveyor and dydx and diversify and suitable power seller as well starknet is our permissionless decentralized ck roll-up and the alpha has been live on a public test net since mid-june and is going live on mainnet ethereum before the end of this month okay so just to reiterate that starkx are these multiple scaling chains that are app-specific right so dydx is using an app-specific zk technology uh platform which we call starkx and also

diversify is also using this and sober is also using this uh and immutable uh but then starknet is a maybe a little bit more what users might be familiar with when it comes to like optimism or arbitrarium one where there's one canonical chain that is permissionless to build on and everyone can build on that that is not let yet live but has been in alpha since since early june and maybe we have some announcements coming soon about its uh release date maybe is all of that correct yeah yeah yeah yeah so as as as i mentioned it's been on a public test and since mid-june it's coming to humanity ethereum before the end of this month i should just uh just one my minor

correction uh the stark x instances are not chains they are essentially uh applications services scaling services uh powered by the starkex cloud service running over ethereum okay let's unpack that a little bit they're not a blockchain so where does it how does it why is it not a blockchain where the differences lie and how does it actually achieve scale [Music] elliot do you wanna um so um i think what defines a a blockchain i mean there are several things it's not there's no formal definition but it's sort of uh this um uh it's very permissionless it's very

you know everyone can come submit transactions and uh you know it's something that basically the public sort of uh runs and uses and um star kicks um is is as you said the line of products that each one of them is sort of a back end that services one particular customer that is another business that is customer facing so dydx it's an exchange that faces customers and it uses a technology stack that is starkx it basically uses the zk start technology in order to compress computation and achieve computational integrity without any trust uh but for trust and math and trust in in the

blockchain so it's sort of an interface and a settlement layer that's very specific to a customer um that's starch so i i think it would be a stretch to call it a blockchain uh it is more an interface in a settlement layer and sort of something that boosts up the um capacity of some other blockchain and in all of these cases it's ethereum a compression service if you will yeah all right that's okay a compression service right so you guys talked about uh said the word cloud who is actually operating the compression service is that you guys yes for start x it is correct okay net it will eventually be a fully decentralized network where anyone could

function as the sequencer improver on the network okay hopefully including you and ryan stark is starcx is this technology that you guys have in your geyser's own cloud that that you guys run and you guys allow people like dydx or so rare or immutable to tap into that service yes okay and then start net is that same sort of service but more streamlined and just made more available for everyone and rather than have it being on one single cloud it's is it is stark net an actual blockchain uh it's closer to it and there's another distinction so starkx is sort of crafted

towards a very small number of uh special purpose use cases that are extremely important for instance uh dealing with uh a massive amount of uh payments you know very high tps like for a payment processor another example is a massive amount of trades or perpetual uh swaps and positions or minting of nfts so it caters to uh things where you have very high demand but for very specific functionalities now start connect in addition to being permissionless um it is also a platform that is turing complete and universal very much like ethereum so it means that instead of us saying okay you can either

you know you can do one of four or five things that our customers really want like you know minting trading transferring blah blah blah basically it says okay anyone can write any kind of logic for any purpose they want it could be generative art it could be you know compressing many votes it could be some game design it could be a whole number of things that we don't actually know what they are we basically um you know the invented and then designed and offered tooling for writing um any smart contract deploying it submitting transactions to it so instead of limited functionality for these things that are in high demand it is this general purpose uh framework so it's uh closer to a

blockchain it's not quite a blockchain because it's a layer two right so i think blockchains are more layer one so just like optimism and arbitrarium i don't think they're described as blockchains they are layer twos we are also a proud layer too i wanted to add to that a very important feature that starkness has had for a while uh and will have on mainnet uh by the end of this month is composability and this is something that a lot of dapp developers care deeply for this allows for you know for this money lego or nft lego or or a smart contract lego to thrive and have this amazingly accelerated and open development framework that that developers enjoy so

much this this is super cool and i'm surprised as we're going through these names that tony stark you know wasn't involved somehow because this is super advanced tech uh but i'm sure you get that joke all the time um can we talk a little bit about the uh stark x first because uh i wanna wanna just cover that and scalability now really with starkwear is kind of the conversation then i would definitely wanna spend the bulk of the conversation talking about start net because that is the next thing that is coming as you said you're you're releasing uh to maine at the end of this month so i'm we're close to the end of this month we are we are breathless with anticipation here but let's talk about starkx first because there is this impression in the space we just saw it last weekend okay

literally last weekend i don't know if you guys you know hang out on twitter at all but the twitter world world was a buzz with this idea that uh ethereum is not scaling today that there is no scalability on ethereum today hence the need for all of these other alternative layer one solutions and you know there was a lot of things i wanted to say but one of the things i wanted to say was of course it's scaling today right look at what stark x is doing look at what dydx is doing look at what diversify is doing look at what so rare is doing immutable x scaling using this compression layer this this layer too on ethereum today now app specific for now

but later uh this can be more generalized in in something like a stark net but can you talk about that a little bit because i don't think people understand or appreciate the the numbers here and what's actually being done across say these four the big four stark x applications today dydx diversify so rare immutable do you guys have any numbers here or any sort of metrics on um how uh what's happening how this is scaling so far yuri why don't you start sure um you know dydx would be i think one sort of remarkable case study that we could point to um the transactions that dydex used to do on layer one uh each transaction composing of two trades

uh a couple of markets only uh those were around 250 000 gas maybe 280 000 gas per transaction the transactions they're doing today on starkx touch many more assets they they have the ability now you can put up collateral against many more positions and so it's a far more complicated transaction we estimate that equivalent transaction on layer one to it will have consumed about six hundred thousand gas per transaction those transactions today in production on star kicks consume sub 500 guests not 500 000 guests sub 500 gas per transaction wow now we're

talking about this this isn't a demo and this isn't uh like a nice sort of case study that we ran for an hour that brought down this is day in day out the past 24 hours were at 10 billion i believe trading volume for a dydx at the end of another one of their fabulous epochs uh cumulative to date over 230 billion bias and boy uh dollar settled on ethereum uh star x in total over 55 million transactions dydx is probably around half that in total so a very very dramatic scaling today already on ethereum i we our estimate at 600 000

gas that's about 300 000 gas per trade that's roughly uh 3 tps would have consumed all the ethereum's layer 1 capacity okay we're doing on a very regular basis 15 times that on this on on star kicks yeah yeah this is just uh show what we're talking about so this is the tps the effective tps that we're seeing i hope you see my screen share this is the effective tps over the past 24 hours so you see it's like on average i'd say on the dydx system so this is like i'd say on average a tps of seven uh with peaks at uh you know 13 tps and

again as we just mentioned the effective tps if all of the ethereum would have been consumed for just uh settling these these trades uh ethereum wouldn't wouldn't be able to to settle this many it can handle a tps of i guess one of them yeah so i i mean at its peak basically dydx is uh another ethereum right oh no it's it's multiple it's a multiple experience wow wow that's just one so that's just one of our systems and it's consuming on a regular basis you know over the past 24 hours we haven't looked at the numbers but it's gonna be less than one percent of ethereum's gas while doing something like five to ten sorry 5x or whatever 2 to 5x the capacity of

all of ethereum and consuming one percent of the gas yes and i want to get to so rare and immutable in a second too it's like but but but what we just said i mean that is scalability right this is why i feel like there's such a narrative mismatch in in in the market in crypto twitter these days and it's uh it's a very bizarre world we're in right now just just to recap there are multiple ethereums perhaps two to five ethereums worth of capacity that is being settled on ethereum by dydx and what you guys said is less than one percent of ethereum's capacity so two to five ethereums being compressed down to just one percent of ethereum but i i want to emphasize that we're not

supporting two to five because that's what we can support we're supporting two to five because that's the demand right right we can support substantially substantially more multiples of that easily today and so when i said uh that would probably be another i don't know how many but like another five to ten ethereums uh you know we meant it i forgot how many million nfts over the past uh half year probably again that would have consulted i think about about 25 million nfts are we talking about so rare or immutable now or both of them both of them both of them cross both of them okay give us those stats again so we don't miss about about i think maybe 24 25 million nfts minted to date wow you ever take a few million here

that's a few million so what's the start of this show and i said that at the beginning of the show that scalabilities here is just not evenly distributed yet this is what i'm talking about dydx hyper scalability it's just uh people are expecting like this thing to be happening like ev everywhere equally across the board but no that we are getting hyper scalability uh like you know sub one penny transactions uh in very specific uh examples and these are applications yeah and these yeah and i think that starcx is bringing well i wanna stress this these are production systems so both surveyor and immutable are working in uh valydia mode where they chose to have an off chain data solution starkx employed as a star

off chain data solution uh because they wanted reasonable gas cost per per minute uh we are minting batches of 600 000 nfts at sub 10 gas per minute once again 10 gas not 10 000 gas wow that's incredible yeah and i think bankless listeners will be familiar with what a volidium is because we've talked about that we did an episode on modular blockchains that kind of talked about uh and touched on all of these differences but it's basically modular blockchain if you kind of break it up consensus data and then execution right you're talking about valydium uses um not ethereum for its data layer uses ethereum for its consensus layer but it uses something

else for its data layer and a volition of course is when a chain head kind of has the choice an application that's the choice um i want to ask about this because this so this is a question incredible extreme scalability super scalability as we said is happening right now are there any trade-offs here refresh us on this with we're still camping on starkx right now but what are the trade-offs here with starkx from us you know decentralization perspective ellie what would you say yeah there are actually many many trade-offs that are behind the scenes i'll just mention a few of them so if you want latency to be smaller meaning

so what what would what would entail so right now we're basically taking a very large number of transactions and you need a very long time for them to accumulate and then you're producing one single proof for that so you need also a lot of computation to get that thing so what you get is you're increasing latency which is bad but you're decreasing and minimizing amortized gas cost per transaction because the larger the batch that's basically what the magic of starks give you the larger the batch the lower the amortized cost per transaction will be okay so one very clear trade-off is that as we grow the batch sizes all things considered latency grows but amortized

gas cost goes down another thing is that um as you um yeah i think that's probably the most important trade-off there are other internal trade-offs in our system for instance if we wanted to keep if you kept the same batch size and latency in everything but you actually wanted the uh on-chain proof to be smaller you would pay with off-chain computation because there are various mathematical parameters you can play with there and that's another trade-off that we have at our disposal you know buy a bigger machine have even more latency

and even for a fixed batch size you can decrease that so i there there's an obvious trade-off in the context of data availability if you want to operate in roll-up mode your data is on chain it is no doubt more secure it is no doubt more expensive uh the beauty of the volition solution which is coming online in in a matter of weeks um is that this will uh this decision will be handed off from the app to the user and so the user can decide on a transaction by transaction basis if they care enough for security to pay that premium on chain data will it literally just be

like a drop down menu where it says here's my transaction i want it stored on chain or i want it stored off chain it'll be just that simple and here are the costs right and it'll have two different costs associated with that yeah that's pretty that's pretty fantastic indeed ellie can we go over that latency trade off again so it's like i kind of understand that but like what is the trade-off that an individual user is making with latency right it's like if i have if there's higher latency how does that affect the i guess security position for an individual user does that just mean things get settled to ethereum less frequently so there's a longer window of time where they're potentially vulnerable

that's a terrific question so uh currently on starkx this is very much abstracted away and the risk of this window is taken by our customers you know dydx diversify immutable and so on now what does it mean it means that uh you as a user and and i'm assuming that many of your listeners have used some of these systems the ideas so they probably noticed that they have instant finality and even if they want to basically go you know retrieve all their funds on chain they get it very very quickly now this does not mean that the relevant stark proofs happen with the same finality but what is happening is

something like this dydx which gets these orders and settles them and matches them uh and calculates the positions knows that once it settles something or did something it basically now goes into this uh stack or queue where which is controlled by dydx and uh you know it will take some time for a batch to reach enough size for it to be convenient to put a proof and then for the proof to be generated and um and accepted on chain and currently for dydx for instance with its very large proofs this this is a number of hours but the user doesn't uh experience any of that what's happening is that dydx is essentially assuming whatever risks are

associated with that and i think they made a wise choice the risks are relatively small because um they know that um you know that something that has been settled will appear in a proof and that approve will be accepted on chain and basically they know what the next state of the so it's a little bit like i think in traditional trading a lot of you know settlement is t plus three but a lot of exchanges will will sort of assume the risk for three days with a case of uh uidx and other customers it's a small number of hours and again it's in their control for instance dydx could decide that it wants to close a batch every whatever two minutes and then the proof would probably be

much quicker to generate and would go on chain much quicker so maybe the latency from dydx's side would not be hours but minutes but then the cost would be that the amortized gas cost per transaction would be higher because the amortization is over a smaller number so most of our customers have chosen to basically crank the knob to maximal batch size minimal amortized gas cost but also maximal latency which is a wise choice switching but when you when a bunch of more users come and deploy more transactions at a higher rate you guys can actually uh increase the rate at which they settle to ethereum

without also increasing the per user gas cost is that correct for sure and we can do that thanks to uh a remarkable technology called sharp which is which stands for shared prover and this is a technology that's in production today serving at the moment so rare and immutable and diversified and hopefully dydx will be added soon to this system and sharp basically relies on the fact that all of these systems run on cairo are a zero knowledge proof uh programming language and a cairo program can describe any logic any application logic that one wants to implement it can also implement a sequence of application

logic uh and that sequence can be immutable's logic followed by service logic followed by diversifies and so on and so forth so you can bundle all these transactions and all this these completely disparate applications into a single proof okay and now this proof can be verified once on ethereum meaning all these applications can amortize their gas verification costs across sort of the cairo ecosystem now imagine if ryan is launching a new game okay and it has just a few users he's just started out not terribly sort of big yet from day one he's benefiting from the marginal cost of the entire cairo

ecosystem the entire sharp ecosystem so that that is a huge cost saver for low bandwidth applications that's fascinating because i create such a virtuous cycle because now here i am with my my my small game here and like i'm incented to bring david's game in too because i wanted him to share in the cost uh for example thing and it also breeds some network effects too i would assume where everyone wants to kind of camp out in the in the same location to spread these costs um we would assume for sure yeah yeah so okay can i just zoom out for a minute because i feel like part part of the narrative around ethereum about scaling these other things is like um let's talk a little bit about your approach to scalability because

i think this is fundamental for people to understand it's it's it's very math heavy okay we call this a you know a compression technology and we did an episode with um justin drake uh called um moon math cryptography and basically the the whole thesis of the the episode was hey these blockchains are trust machines and the way to scale them up to get a you know a giant leap from one generation of the computer to the next is is actually it's not just creating bigger blocks and you know bigger block sizes it's actually cryptography right that's the technology that scales these trust machines right and so to me it seems like you know stark and

zero knowledge this is a big ad advance in terms of cryptography technology and the key to all of these unlocks that you're talking about so not asking you to get into the you know the moon math or the specifics that would be probably definitely beyond david and i and beyond many in the audience but can you just talk about this philosophy because this is a different path to scalability than i think we're seeing with a lot of alternative chains which are increasing blockchain size uh like uh creating more complexity and and difficulty in running validators you guys are scale scaling with math and cryptography can you talk about that for a minute uh ellie what would you say about that

yeah so um um i completely agree with justin that blockchains are about being a trust machine and i would add to that that the the reason we trust them currently is that um everyone and anyone can be uh participate in validating that all transactions are good and they can do so using very um simple off-the-shelf computational means so just just here's here's a bad way to scale what if we just uh 10x uh whatever the uh gas you know the gas limit of a block on ethereum or

increase the block size on bitcoin so what would happen is that there are machines that could process 10x more but it's probably not going to be my laptop the one that i'm using to talk to you and very quickly you'll just have a very small system of very big computers that process very high tps some so-called blockchains are already doing that and you know whether they are decentralized and trustworthy is is open to debate certainly the conventional normie world has been doing this for a very long time i'm sure that banks and you know credit card companies have very big data centers but blockchains don't want to do that they want to

continue and maintain inclusive accountability that allows each and every one of us to use their laptop to check that everything's okay so what are we going to do well there's this amazing math that started in the 1980s um and we're not going to describe how it works but to quote from one of the very earliest papers in this field i'm probably misquoting the exact thing i can find it later but it says that using that this technology again that was invented already in the 1980s even a single weak computer can monitor and track a computation done by a herd of super computing computers even if they

are run on faulty hardware by malicious parties so it's a very cryptography is this amazing powerful technology that allows even the weakest of computational devices to monitor and know with very high certainty that a huge computer run by an evil government even has been executing with integrity and this is completely mind-boggling so you know the the little man can basically check and know that some huge corporation monopoly or government is operating correctly because of the power of cryptography now this was already known some parts of it were known in the 1980s what was not

known was how to make this like really efficient and tractable and that's i think where a lot of the inventions coming from from starkware really help first of all making it practical on today's computers and more recently making it available to programmers i just read today someone that wrote you know here's my experience of working on startmit and the very first sentence is you do not need to know anything about the math of starks it's all been abstracted away you just have a programming language and you write your code and you automat you know you press a button and so you just program as a developer no math no nothing and you get this uh

very low gas cost uh and the integrity offered by some magic that is behind the scenes it's incredible and it got by the way i just got to say it's like cryptography is just what a gift to humanity right of all of the technologies that we can unlock at this juncture in humanity we get this freedom technology that we've just unlocked what a gift to humanity it is so one part before we get into the conversation about stark neck which we definitely want to go to next next um it's it's really fantastic that you guys have abstracted away all this complex technology and just allowed people to build on with his coding language and apis the other side of that spectrum is like well if you guys are just making it super easy you guys are centralizing all

of the development onto your guys's cloud server so there's this one cloud server that's doing all of this central processing for these many many systems so what happens in the case where the police force of the united of the world come and knock on your door and says unplug the server what happened what happens in that case well if it's of course if it happens before we decentralize which would be in a small number of months then yes uh you know maybe we'll just uh release everything before we uh go out with the police to some undisclosed location but the the plan and the hope in our you know we certainly believe this will be the case is that within a very small number of months it won't be

uh starkware operating these things and the code will be available and many many operators including uh you and ryan will be uh running it just like with ethereum so kind of in that way where we talked about having the drop down menu of where you deploy your data might you have a drop down menus like oh i want starkware to run my server but or i i want to run my server like you guys probably would do both well like ethereum like uh you know you don't get to choose your miner there's some consensus thing that pick the next so something like that you'll send it to some network peer-to-peer network and there'll be some uh some mechanisms by which a decentralized network of approvers and sequencers will pick it up and it will be accepted you don't choose

which one but that's i think that's probably even better you're not you're not burdened with that you just send your transactions to the network ethereum style yeah wait so is this now what you guys are talking about right now is is that stark net or is that something to okay exactly okay so we've actually gotten into this dark net okay can you guys talk about the evolution from stark starkx into starknet how is that transition going to happen yeah and i'm actually as you're speaking i'm actually going to bring up this um this diagram that i've seen in a lot of your documents that kind of illustrates it and uh yeah so for for people who can't see this on the podcast there's three steps and uh over the top there's 2021 and 2022 and this is kind of the starkware roadmap as i see it

and uh the starknet roadmap to be more specific the first step is planets that's a single operator single app roll-up the second step is constellations single operators multi-app roll-ups this these these two steps are both in 2021 and then the third step in 2022 is the universe which is decentralized operators moving from single operators to decentralized and multi-app roll-ups as well so with that with that context too uh yeah to david's question so you know planets is essentially stark x right these are single operators so rare immutable x diversify dydx they're each running a single app roll

up okay and that's been in production three of these systems went live in in in april the numbers we quoted earlier you know exceeding uh 50 million transactions etcetera this is across all these uh systems constellations is essentially what's coming in a few days time and that is a single operator like in the context of the startnet system initially much like the optimistic role of playbook optimism arbitram we're following in the footsteps of our very talented colleagues in the row up space starting with a single node operated by us it will be a multi-app roll up this is this goes back to the composability point

that i made earlier so folks would be able to deploy their smart contracts on this invoke one another smart contracts and you would have a single roll up with all these applications running on top of it and universe which is coming next year is the fully decentralized vision where the operator namely the sequencer improver functions are fully decentralized and all this bliss is shared across multiple applications sharing estates invoking one another etc so on the step two the constellations is this where immutable and dytx will actually find themselves on the same roll-up so that that is a separate question of of if and when the starcx instances currently running choose to port and

move over and run over stark nets okay uh but for stark for constellations we're talking about for example uh teams like uh zigzag which is building a decentralized exchange uh that announced that they're going to be launching in a matter of days on starknet and influence.east a wonderfully talented uh team building a game these will all deploy on starknet and coexist there and be able to invoke one another so guys david okay so the conversation that we've been having in this space the space has been having and i know a number of people by that are confused by this point

the zk roll up bulls think that zk roll up evm is powered zk roll ups are going to be here in six months the optimistic roll-up balls think that the development and like infrastructure needed to make zkevm rollups isn't going to be here for three to five years what's going on like what's true so unfortunate unfortunately vitalik published a post in early 2021 this year uh announcing that the curl ups are the long-term solution um that they're coming in three to five years uh as as we apologized already a few weeks ago uh we're sort of ahead of schedule so so this tooling is here

now in fact it's been on a public testnet since uh the middle of june supports general computation supports composability there's still a ton of work to be uh to be done and it's being done as we speak if folks hop on the starting discord they'll see a very thriving ecosystem of tooling and infrastructure that's being built in a decentralized fashion by dozens of developers outside of starkware of course starkware itself is contributing a significant uh amount of talent and resources into this but all this tooling is being built on a week-by-week basis and making tremendous progress in that respect so we are able today to support essentially any application that wishes to build on top of starknight and in

fact what we're already seeing is people doing stuff that is just completely out of bounds on layer one so to give a couple of examples uh guilty gyoza just a brilliant programmer who's doing all these sort of physics simulations on starknet now some some people said you know what is that good for i think that's a sort of almost a silly question in the sense that the question is not what is it good for the question is why aren't sort of people registering the fact that any computation any computation that one chooses to do is now suddenly within the breach of blockchain developers whereas when we were limited to layover player one

uh it was a very sort of resource-constrained environment so so yuri i just wanna just highlight the answer to that question is like you're saying it's here now you're saying general purpose zk roll ups are here now ahead of schedule and that last phrase ahead of schedule it's like like that that's not something we hear very often in crypto okay folks making those statements consult with us now that's fine you know making it's well known that making predictions about the future is very hard and if you're trying to predict something about

a field of expertise it would be prudent to go and ask some of the experts what they think so you know you have two of the experts uh it's not a prediction you know we have uh how many four systems running many of them for more than a year we just saw the numbers um now with general computation on ethereum scaling it as you mentioned you know two or three ethereums each one of them today not three to five years not six months this is something that happened in the past now next week sometime um starting at alpha it's already in test net but general computation same

sort of scale um available to everyone next time you want to make predictions about zk roll ups you know you can know you know where to find us dm us check with us before making the prediction and uh yeah this is great look this is incredible you won't find a single person listening who will be disappointed that you just said that okay we've just been um we feel like with crypto in in terms of scalability we've been disappointed too many times right which is like when when when teams talk about dates and the shipping it always seems to slip slide now you guys are shipping yeah so you guys are shipping you have it so can we talk about uh what then is

shipping next week right like what exactly so understand it's a general purpose version of starknet are there any applications that are launching is it going to be permissionless at launch so can any team decide to start building on top of this or is there going to be a phased rollout some sort of white list uh yeah give us some more details on what should be next week we'll link a post uh with a lot of disclaimers on like what is not in there so there's a lot of things that were very um meticulous in pointing out you know we don't want to make false statements so but we'll share there's a lot that will

be missing initially next week account abstraction won't be there be rc 20 tokens there won't be any uh it will be very hard to deal with fees a lot of things will be missing but there will be applications and in terms of um deploying um anyone can submit a contract initially for computational purposes so that you know the system might be buggy and not running as well as we want we will do a white listing process just like arbitrarium and optimism and other um l2s to ensure that uh you know our system is not crashing and this will be

removed later on um [Music] i'll find the link i i just want i want to just mention some of the uh of the folks who are building and and some of these will launch with the alpha some in the uh weeks following the starting at alpha launch i want to reiterate what ellie said this is called alpha for a reason uh and we we strongly urge everyone to sort of proceed with caution that you know when our training wheels are on uh we're all sort of figuring this out together so i know we we we want sort of people to fully appreciate that fact uh but some of the sort of the ecosystem that's being built around starknit so we

have the nethermind team we're building the warp transpiler and the voyager block explorer open zeppelin were defining the standard contracts for starknet last time they defined that that was for this blockchain called ethereum the aragon and equilibrium teams were building full nodes figments were building api services arjun tour building uh tooling and a wallet the tourist wallet ledger or building an application to go live soon uh the shard labs team building tooling the uk uh front and scheme multiple [Music] whole bunch of applications a whole

bunch of applications so just to name a few snapshot announced uh uh recently that stark vote we're building stark vote together so this is to make voting for daos uh once again sort of open to the general public and nod and exercise for the wealthy whales uh there are multiple games being built first and foremost there's sort of the wonderful prama was really one of the very first folks to start building on starknet building dope wars and another surprise uh sort of a game that will be announced very soon the proof of humanity team guilty gyoza we mentioned the influence team were built moving their uh

wonderfully uh uh interesting game from the layer one ethereum brick sacrament the starting dot dot js framework by sean hans exchange we mentioned the qsar team maker has a core unit uh moving sort of maker to or deploying maker on on stark net and hopefully one day soon minting die on starknet and rumor has it uh just following crypto twitter there's some that some core depths of ave are looking into uh building an avi version on startnet so as you can see there's an awful lot going on so i can't sort of give you exact dates of each one of these things but

we're talking weeks here yeah this is these things to start rolling out this is fantastic and i'm wondering if you could kind of project forward six minutes for us right it's like getting getting to a user's head right so all of those apps sound wonderful if they're deployed on um starknet that would be fantastic also other things that are important is wallet support right it's like you know the metamask of the world you mentioned argent which is fantastic um layer two sorry uh fiat on ramps as well are going to be super important so if you project this forward you know six months to maybe even like nine months to 12 months do you think all of those things will be in place like this will feel like basically a fast ethereum type of experience or do you think it's going to realistically take a

little bit longer i'm i'm an optimist i think uh in six and definitely 12 months all of that and much more the main thing that look back and and be surprised at is like oh you know some new classes of applications that cannot live on l1 because of the gas crunch there and suddenly there's some new class of stuff that that uh you know we can't predict right now it's going to be there um the main thing that will happen within um the next six months and certainly nine months and 12 months is a lot of standardization and a lot of uh ecosystem growth will be a whole bunch of

standards for various things you know account abstraction fee abstraction um interactions bridges erc20s erc721s and so on and people will be using them but much more importantly will be either past the decentralization point at universe where there will be multiple provers running permissionlessly and will probably also have uh decentralized volidium for data availability and volition uh initially we'll start in roll up mode and we'll probably already be on route to various decentralized governance mechanisms and all they entail so all of that will be uh

basically our our goal is you know starcraft currently is you know a company that is running all the start provers for these production systems and within six to nine to 12 months you'll see that dramatically changing and and we will not be running the majority of the provers and sequencers and data ability providers well ellie that gets right into the part of the conversation i want to go next which is that long term horizon for what starkware is with as it relates to the rest of the ecosystem and zk tech is actually a platform agnostic tech and so we actually can talk about this tech outside of ethereum

those are the conversations i want to get to next but first we have to talk about some of these fantastic sponsors that make the show possible matcha everyone's favorite decks aggregator has just launched an open beta for gasless trading so if you're trading more than five thousand dollars in common eath and wrapped bitcoin pairs then your gas fees on matcha are free and that's why you should be using matcha matcha routes your orders across all the various d5 exchanges on ethereum polygon binance smart chain and gives you the best possible price without any trading fees or unnecessary slippage matra has smart order routing that splits your orders across multiple liquidity sources if matcha sees that that it gets you better pricing trading on matcha is super easy because it pulls the liquidity for me

into a single and easy to use platform and has even saved me multiple times from accidentally picking the wrong decks to trade on and getting a bad price matcha also allows you to make limit orders on chain so you can set and forget your defy trades and they will go through automatically while you're away so when you're making a trade head over to matcha.xyz bankless connect your wallet and start getting some of the best prices and most liquidity when you trade your crypto assets alchemix is one of the coolest new defy apps on the scene it introduces self-paying loans allowing you to spend and save at the same time deposit the die stablecoin into the alchemix vault in order to get an advance on the interest it generates borrow up to 50 of the total amount of

your deposited die in the form of al usd stablecoin here's the craziest part the loan pays itself back and you cannot be liquidated unlock your assets potential in the ultimate defy savings account and brand new to alcmix is the eth fault where you can deposit eth into the application borrow al eth against your deposits while having your advance gradually paid back over time v2 is rapidly approaching which will allow for even more collateral types plus a variety of yield strategies to choose from harness the power of alchemix at alchemix.fi that's a-l-c-h-e-m-i-x dot f-i follow alchemix on twitter at alcomixfi and join the discord to keep

up to date with alcoholics v2 and to get involved in governance all right guys we are back with the starkware team coming into the second half of the show just tying off some loose ends with with some of the things that are uh not yet talked about on the show and the first one i'll talk about want to talk about is the long-term prospects for starkware the company uh because you guys are trying to decentralize your own product which kind of means you're trying to decentralize yourselves so what's the long-term future for you guys what's the long-term future for starkwear i want to take a swing of that first you know i would mention that we've we've decentralized successfully with our kids uh so [Laughter] so i wanted far say that the number one objective is that keep on having as much fun

and uh meaningful fun and uh keep on uh expanding this amazingly talented and capable team that will grow into an ecosystem that's uh objective number one um but as as david correctly pointed before the break uh zk stark technology is much bigger than just a solution for um you know scaling ethereum uh initially of course it can and will be adopted by uh we think pretty much all blockchains that require scale it's sort of just like

all over the world you know you'll see fiber optics everywhere because they you know scale bandwidth really well and you'll see uh you know cell to not write cellular phone technology everywhere because it's like scales you know communication or makes it easier you will see zika starks on all blockchains that still exist and and we start where would like to help that happen once we can put our focus right you can only do one thing at a time uh but beyond that we think that this technology that allows even let's go back to the cryptography that allows even simple citizens to uh sort of hold and check the strongest of monopolies and governments this is something that we all need

today so we think it will permeate from that into the larger world now that's about the technology so there'll be a whole bunch there is already a whole bunch of other teams that are driving this forward with starks and we're very proud to see them and help them you know there's the winterfell and and myden effort that is is now incorporated into polygon and we very much support this effort there are some other teams that we know are dabbling with starks but have not gone you know public with that information and we support them as well um so this will be there will be in multiple ones now for starkwear we hope like you know what are we gonna do after this decentralization

so like two very possible routes after like let's say decentralizing one is it's very plausible that some of those uh conventional um institutions and corporations and so on might want something that is uh their standalone version of stark technology and we can help with that and the second thing is hardware uh as more and more proverbs and more scale is demanded uh we might uh want to try and help with that so there's at least two opportunities ellie i want to ask you a question here about decentralization so some other i guess uh layer two's answer to this partially uh and even layer ones is

about you know community ownership as well right some sort of like tokenization some sort of form of community ownership you know whether that takes the form of a governance token or something like that how important is is community ownership to your roadmap to decentralization what would you say about that very important but uh on the specific question of token uh we we hold to our position of no comment gotcha that makes sense um can you talk a little bit about like you mentioned this with ethereum right so and you also mentioned that the this technology is um agnostic as well um but it seems like as as david and i

have talked about this it seems like those chains that have embraced this more modular design this modular thesis rather than you know maybe the contrast is a monolithic where you're doing execution and consensus and data availability all in the same platform the more modular chains will be probably first to pick up this technology and ethereum is definitely with a rollup centric roadmap definitely adopted that near as maybe another one tezos is maybe a a third um do you think that the the more modular chains will be first to adopt this sort of technology and i guess do you have any thoughts or comments on where this might leave the rest of the blockchain space like the more monolithic designs do you think

they eventually capitulate and sort of become rollup centric and more modular design oriented any longer term thoughts here i it's it's this is something that's hard for me to predict uh for two reasons one is that we our focus is just very limited on this one thing in stark net so it's very hard to uh i mean we did get approached a bunch of times by various l ones you know asking if we want to deploy and we gave the same we need to focus answer and then i would say if i if i had to give an answer i would say just like i said with like fiber optics or cellular technology starks are going

to be anything that still lives in uh you know five years in this blockchain will will want to use just like you know any communication network in a modern country today probably has some fiber optics so i would say any blockchain that is massively used would also want to use zk starks because they they just offer you this asymmetry right having one huge computer do a whole lot of work and everyone else trusting the outcome even if that node doesn't have to be trusted that's a very powerful technology that that i think all blockchains are gonna want one subject we haven't touched upon uh yet is is privacy are there any privacy

ties in terms of what you guys are doing well i mean we we are starkware and all of it is like stark and we're sort of omitting the zk which is part of you know the whole line of stuff that we invented um and we even did a little bit there's some open source stuff that is already uh you know there's a signature post quantum secure signature that we developed that's also open source and um um so yes for sure at some point we will add privacy and zk to that's another thing about focusing right we there's probably demand for that we just can't uh

you know uh defocus right now so so but yeah there's tremendous opportunity and demand there the scale with the privacy is uh you're right we sort of neglected it recently just for focusing purposes hey david you're muted thank you one thing i want to ask about is uh the economic sustainability of starkware the business like what's the actual business model there how do you guys make money sure so at the moment starkx uh these are commercial agreements that we have with our customers that generate revenue in stark net we hope to be one of the participants in this open network and provide services

to the network and hopefully generate revenue by doing so as i mentioned there are other venues to consider beyond that for example hardware we're building applications for the many normies we expect will show up and want to benefit from this beautiful decentralized trustless world so you know those are all directions that we're actively exploring and uh and working on right so with the the commercial agreements with the stark x chains um like immutable dydx uh you you're if you're making money off these things it must be coming off of the transaction fees of the network right uh i don't understand how it would be any anything else how how how are these agreements

made with things like teams like dydx and immutable and how does the can you just illustrate the the value flow from the people that transact on these things to how it actually becomes economically sustainable for starkware sure so as you know is just going to say going back to the dydx numbers if if uh dydex transaction on layer one if such a layer one existed that could support their uh their existing throughput if that's at six hundred thousand gas and and on our system they end up spending less than 500 guests well there's a huge pie here uh and and you know everyone could be made uh exceedingly happy the customer dydx startwear

so um you know there's there's a lot of value to be had is it just as simple as uh say i'm an uh an immutable user i'm on nemo or dydx and then i make a transaction the system charges me a dollar to transact on that uh on that layer two and then that split up between you you and dydx are you an immutable uh i'll give an example d5 pooling is this beautiful service that seller is going live with uh powered by starkx and d5 pulling caters to the long tail of d5 participants who are priced out of the market on ethereum layer one today okay they want to transact on ave on compound whatever it is layer one transactions

are absurdly expensive okay we're talking about people who want to put a few hundred or a few thousand dollars to work and end up spending tens maybe hundreds of dollars in transaction fees per transaction defy pooling basically allows them to move their assets two layers to pull them together like carpooling services and those carpool services then initiate transactions single layer one transactions that go from that layer two system onto layer one so [Music] they pull together a hundred users or a thousand users that all go combined in a trustless fashion say to an uh initiated of a transaction so a single layer one transaction now is split a car

so they can pay this service uh they spend far less on gas ethereum is far less burden with all these transactions ave is getting all this action that would have been sort of just you know left out of the market um sellers making money hopefully starkware is making money everyone is better off for it the market is more efficient right right so pie is bigger the pie is actually bigger yeah right yeah abs absolutely the cool thing about all these roll-ups it adds just economic activity to the the greater pie right growing the pie so the idea for starkwear is that you know if if we can 2x the pi uh

you take like a five percent of the the extra growth or something like that that's the business model oh well 2kx the pi but yeah [Laughter] is there an actual like percentage that you guys charge or how do you guys actually structure your fees the the the the fees are sort of commercial agreements that what we put in place with our customers and uh that handles sort of the the uh sort of the revenue side of things yeah is there each each individual agreement with each individual chain is its own agreement or do are they kind of just like similar agreements similar structures no no they're actually they're actually very similar in structure i felt like that point actually okay i want to say something about actually uh like a potential uh you know business uh model

just around fees on stark net for for and now not for starkware just for anyone running operators so um the argument goes like this um you know currently take the uidx as an example right we are consuming much less than one percent of the network but uh offering scale that is like roughly 5x or 10x more than that so there's like a 1000x factor or 1000x reduction in cost between doing a transaction on l1 and doing it on our l2 so now what if um the fee charged by such an l2 would be let's say one-tenth the fee of an l1 so if you run the

numbers first of all users will be extremely happy because they're paying just one-tenth but this leaves huge profit margins for the uh network right and if you take into account that for instance on ethereum today the daily fees are around like 60 million dollars and you're taking uh you know you're consuming one percent of the network but you're charging let's say uh in total 10x that while doing a thousand x more everyone's happy and the actual you know uh revenue per day for the operators is something around if i got the numbers right like six million dollars per day

just in fees and most of it will be profit and the users are happy now i'm not saying this will be the main or only business model and it's not something you know we we explored in depth but i'm saying just by the fact that there's this huge amount of fees being spent today on ethereum along with the amazing you know 1000x scale factor that that arrow l2 will bring starting next week there's a very lucrative business model that sits just even there fantastic that's a that's a fantastic color and guys i want to thank you for your time in helping us unpack what is this really really complicated technology in very easily to digestible ways i'm going to zoom all the way back

out for this last question and something that really just fascinates me personally is like the the relationships between humans and code right there's there's a there's a link here and crypto as me and ryan often say crypto is a political revolution it's not left or right but it is anti-authoritarian right uh and so when you guys are looking at at starkware and looking at the code that you build do you see any sort of like values alignment like what how does starkware what you guys are building at starkware align with the values that you guys have sure i so i you know i can i can share my personal perspective so i you know so looking back over the past few decades technology is just technology right it

can do all sorts of of things good good things and bad things and often the same technology can be put to good use and malicious use um i think that a lot of the things that we're building are inherently uh for the better in the sense that they are inherently supportive of shifting the power of balance away from centralized entities into the hands of individuals in all sorts of very meaningful ways that ellie alluded to earlier and uh and that is something we care for deeply and and i think it's something that a lot of people all over the world are concerned about you know with these past few years and they should be concerned about and

having the technology that can support those values and do so at scale meaning not as a thought experiment not as a little sort of exercise in independence but something that can actually serve their daily needs yet protect those rights and and privileges is something we care for deeply i want to add two two different thoughts one i think um the values that uh we have at stark where i think the first things that jump to mind that i think distinguishes us are like uh meritocracy and uh excellence um you know we have like audit teams looking at our code and

saying you know we haven't seen many code bases that are you mentioned code like that are as professional and clean even though this is like very very moon math stuff and very easy to get wrong so i think we have um i mean our whole team right it's it's uh uh you know both oohy and i don't really code but uh fortunately for all of you guys i don't think the audits would have looked the same way had we dabbled in that but um like this very uh we're very proud of this um you know excellence professionalism and meritocracy when it comes to code but

also you know to math and science and innovation the second thing i would say is that uh we do we are having fun but we do recognize that through this um you know opportunity offered by inventing a technology that we believe will transform the world we understand that there's immense responsibility on our collective shoulders we can try to get it right for humanity and we could mess it up and i think so far we've really really tried to focus on making decisions that are correct not only for our own personal short-term

goals for instance i'll give one example we founded the company in 2017. everyone was icoing everyone said guys you know just do an ico at the time we raised 6 million people were raising you know 10x and 100x on that with technology or with promises that were not as you know impressive as ours so we didn't do that because we said it it's not right for uh for you know the state of the technology and we went first for building very specific applications handcrafted like starcx is very limited it does like a few things um

so we're like treading with a lot of responsibility and recognition that we are lucky to have this power to uh you know unleash this technological revolution and we hope will uh make decisions that are good for all of humanity but uh you know who knows ellie yuri it's been an absolute pleasure i think that's the perfect way to end it tremendous amount of responsibility on our shoulders being early to this movement we completely agree it's part of the the bankless nation it's part of the bankless journey we appreciate so much what you guys are doing in the scalability space and we are

massively excited about stark net and uh what you're delivering both next week and in the weeks to come so thanks for joining us today thank you guys it'll be a fun ride absolutely bankless listeners um i hope you enjoyed this i hope you learned a bit more about one of the the premiere layer two technologies coming i certainly learned it was even i knew it was close but it's even closer than i thought the end of this podcast so once again another bankless podcast i am bullish some action items for you if you are a builder if you're interested in developing on starknet and this technology make sure you check out the show notes join the the starkware discord service where you'll find a lot

of information about this and fellow builders who you can build alongside with also we'll include a link to the post that was mentioned earlier about the start net alpha release and i expect to see some posts next week about the mainnet release too very excited about that risks and disclaimers of course none of this has been financial advice maybe this is scalability advice though eth is risky defy is risky crypto is risky you could definitely lose what you put in but we are headed west this is the frontier it's not for everyone but we're glad you're with us on the bankless journey thanks a lot hey we hope you enjoyed the video if you did head over to bankless hq right now to develop your crypto investing skills and learn how to free yourself from

banks and gain your financial independence we recommend joining our daily newsletter podcast and community as a bankless premium subscriber to get the most out of your bankless experience you'll get access to our market analysis our alpha leaks and exclusive content and even the bankless token for airdrops raffles and unlocks if you're interested in crypto the bankless community is where you want to be click the link in the description to become a bankless premium subscriber today also don't forget to subscribe to the channel for in-depth interviews with industry leaders ask me anythings and weekly rollups where we summarize the week in crypto and other fantastic content thanks everyone for watching and being on the journey as we build out the

bankless nation