Scaling with STARKs - TOKEN2049 London 2022

so happy to be here I would love to tell you a bit about what we're doing at starkware so uh what I don't need to tell you is that blockchains have a scaling problem and what I do want to make sure in that context to sort of to emphasize is that validity Roll-Ups or as we uh uh often hear others refer to them zero knowledge Roll-Ups are the future and the purpose of this talk is to emphasize the point that they're not only in the future uh they're in fact the present so uh what do I want to talk to you about I want to talk to you about Stark X which is our first offering it's been on the market for over two years this is a SAS service that we offer scaling as a service essentially uh the what's under the hood there and what we sort of were forced to develop under the hood there which is Cairo our uh programming language that allows us to prove any business logic with a stark proof and finally I want to tell you about Stark Nets which is uh the open permissionless network Layer Two that we're building out using Cairo to settle on ethereum and help it scale so let's start with uh Stark X it's the leading layer 2 on ethereum it's been live since June of 2020. uh let me share some stats so cumulative Trading uh Stark X is settled just over three quarters of a trillion dollars over the past two years over 250 million transactions sold on ethereum I can't stress enough you know if you were to ask me a year or two ago I would not have imagined these kinds of numbers and these basically all fit into an increasing level of comfort with a software stack and how battle hardened is right this is all software money is represented as software as code to the extent there's the hack to the extent there's a vulnerability it will be exploited uh we have a half a billion dollars in total value locked and another uh one particular anecdote worth mentioning is that we've minted uh just under 80 million nfts on Stark X to date by far more than any other uh player in the ecosystem these numbers number of transactions settled total value settled number of nfts far exceed all other layer twos combined you see here at the bottom some of the uh notable names that are scaling using uh uh Starks and these are wonderful partners that we're very proud of of having worked with over the past two years um here's some uh recent numbers this is from mid-october uh and this chart shows uh you can look here at the the cyan is the ethereum plot and the blue is Stark X and what you see here is that multiple weeks over the past six months Stark X has processed has settled more ethereum transactions than ethereum itself okay these have been processed as the layer two and settled on ethereum and this particular week a couple of weeks ago uh 12 million transactions processed by Stark X which is uh dramatically more than ethereum itself the beautiful thing is that this all relies on ethereum security and ethereum resources all these proofs are verified on ethereum the verification consumed well under one percent of ethereum's resources so we use essentially one over a thousand right a thousands of ethereum's resources and in the process are able to process uh way more transactions than ethereum itself this is these numbers basically manifest the ability to scale uh using Starks let me share some more anecdotes um one measure that I think is very interesting in this context is for various transactions how much gas those transactions consumed on ethereum layer one and how much those Now consume in gas using Stark x d y d x for example we're settling transactions on layer 1 ethereum for anywhere between a quarter of a million gas to 300 000 guests they've been settling tens of millions of transactions on Stark X at 480 guests so 700x Improvement in the gas per transaction so rare and immutable have been minting nfts for 200 000 gas per mint they're minting now on Stark X at anywhere between 10 and 35 gas per minute so 20 000 X Improvement and when I try and sort of Drive the points home I often say yeah I I tell people I I told my kids I'm unaware in the context of Engineering in the context of finance of too many examples where uh resources have been have seen such a dramatic reduction a three over four magnitude reduction uh as as this you know if you look back I don't know the printing press the Industrial Revolution but these are massive step functions and they're often in fact exclusively triggered by technology and we think that Starks and zero knowledge proofs are exactly one such technology so um in the process of building out Stark X we implemented a particular set of business logic uh specifically this was for spa trading and invariably your customers and new customers come in and say well can you do this and can you do that can you change the business logic can you expand it can you mint nfts can you trade nfts can you do Perpetual contracts and as long as we were working sort of direct translating directly uh the business logic into bare bone raw metal math that was a very slow and dangerous exercise there were about five people on the planet fortunately all at Stark where who could do this stuff it's it's Error prone it's a security Hazard okay and so the conclusion quickly was that this is not sustainable that if we want the business to grow we need a programming language that allows us to generate with code proofs for any logic and this was this is how Cairo uh was born it was born as always out of a necessity and so Cairo is what was under the hood for starkecks for our SAS business but then very quickly we said hold on a second we're now using it for Stark X the thing to do here is to open up this language for any developer anywhere okay and this is of course driven by the the ethos and the the principles that that guide us in building out decentralized ecosystems but from a raw business perspective the motivation here is to allow the business to scale out if any developer can Implement code and deploy it on the blockchain without talking to us they win and we win and so Cairo uh one question we're often asked I was asked on stage an hour ago and apologies for anyone who's sort of uh hearing some of the same comments over again is why why didn't you actually use solidity and we would have loved to use solidity at it as is uh we came to the conclusion that this would be far far the far less than optimal okay and and and in fact by doing so we would be tying these weights to you know our ankles our proverbial ankles in a way that would limit the scale and efficiency of these networks for years to come and so we thought that the sensible thing is to take a bold step and introduce this new language and that is a bold step now I want to say in this context looking seven years back this is not seven years back and I often have to remind people on the blockchain ecosystem and they're often half my age but I have to remind them of this the solidity software stack is seven years old okay this isn't something that's been around for decades um and in that sense you could ask the question hold on a second why was solidity invented why didn't they just code smart contracts on ethereum using whatever you know C plus plus or JavaScript or you know one of the many languages that are used by web 2 companies worldwide and the answer is that they face the same deliberation they they thought we could do it with existing programming language but those are not smart contract languages they have different they were built for different needs and different specifications and what we need here is something else and so solidity was invented and so the concept that as you move to a new platform As you move to a new paradigm in certain cases new Tools in specifically a new language may be required is not something that should be you know foreign to the blockchain ecosystem so let's talk a bit about Stark net uh we announced the uh the concept uh beginning of 2021 so almost two years ago in June of 2021 uh we launched a public test and supporting smart contracts General computation and composability so everything that a smart contract developer would come to expect conceptually from a smart contract platform uh we went live on Maine in November of last year and what we've seen since is an astonishing uh Community form around starknet uh we've had multiple events over the past six months hackathon starting at CC events with uh four or 500 developers in attendance to each one of these our assessment is there about a thousand people right now writing Cairo code worldwide once again that number a year ago I would not have dared sort of uh imagined some numbers here on the performance of of starknet there's a lot going on um and we're very proud of that uh the test that is obviously busier than menet I want to stress that there's a ton of work ahead of us right there's uh you optimize the prover and the sequencer and the network becomes the bottleneck you do all that and you want to decentralize and that in itself is our gargantuan effort and so um you know we've been working our asses off a team of a 100 within Stark where many hundreds of developers across the ecosystem building the various tools that need to be uh used here and as of a couple of months ago there is a starknet foundation that's tasked with growing this ecosystem um you know some some uh some additional numbers on the excitement around this one particular number that I want to point out is there are about 20 star uh startups that are sort of starkman native and when they went out to raise money going a few months back it was far from obvious to me that they won't come back and say we talked to the VCS and they said too early we're not interested in doubling down on this whole software stack let's wait and see that it picks up some steam and in fact what we've seen is an abundance of interest from a very broad set of venture capital firms so what I told you about today are the three steps sort of in the product Evolution and starkware starting from the SAS offering Stark X Cairo that we had to build to support the sort of feature velocity that we wanted for Stark X and then the conclusion that Cairo should be put in the hands of developers in a permissionless fashion and that should be encapsulated in Stark net we've seen applications in a whole range of fields right a gaming defy social applications uh and we're very happy to see that from our perspective we're a technology platform we want people to explore far and wide uh I want to say a few words about the future we see and I talked about this on stage we started out of course with ethereum we built out starknet as a public infrastructure to scale the network and what we then basically encountered once again reality in the market speaking to us as applications came to us and said we cannot live we cannot scale massively on public infrastructure directly we want our own little environment and so these are the layer threes that we describe up here and there each participant can decide the pace and risk profile and risk parameters and Technology Stacks that they use to their hearts content and from our perspective that's a perfectly viable outcome we are not sort of the hardcore ideologues that say everything has to be done you know the world is going to look like many unit swaps we don't think the world is going to look like many of those swaps there are unit swaps and then there are other applications and other companies with particular needs and our job is to serve them and from our perspective that scale and that control that businesses need is going to exist at layer 3 and Thrive there the the tools to support layer 3 specifically the ability to recursively prove um that's been in production on ethereum mainnet for the past three months so we are today actually using recursive proofs for surer immutable uh uh diversify uh all the other systems we have in production including Stark net itself General computation recursive proofs is something that was a scientific paper a couple of years ago I just want to sort of impress upon you the pace of Technology Innovation here uh that is all uh happy to answer any questions if there are yes please thank you just one second that get you a mic because I know this is going to be a good question uh yeah I've got two questions um I understand that on maintenance ethereum you're you're verifying the the roll up and the you know the execution that's happened on the layer two are you able to bridge assets are we able are you able to transfer assets to mainnet ethereum from stocknet from you can withdraw acid of course you can withdraw assets from Stark X and Stark Nets ethereum and of course and it actually beautifully you I don't want to take up too much time you can do that even if the application says so rare and starkware okay vanish off the face of the Earth users can withdraw assets uh directly from the smart contract and by doing so we basically tie our hands in the sense of we are unable to censor them yeah okay good um second question zero knowledge proofs are often associated with privacy as well uh is that is that a feature is that in the road map is that something so yeah my co-founder Ellie is the founding scientist of zcash and of course that was the focus there um that is very much on the roadmap a because it's a beautiful uh sort of capability of zero knowledge proofs but I think the the business response is that businesses are going to need it and human individuals are going to need it right all of us but businesses are gonna need it no no even if we offer everything that we've described here to web two companies when they come in and say hold on a second we don't have proper privacy and we cannot engage in business that way so that's going to come it's a must it's further down the road now