Panel I Practical Zero Knowledge Proof Applications | Pragma Istanbul 2023

[Music] I want to get ready for inviting three of our amazing Founders we'll bring on Sandy from scroll Zach from Aztec and Alex from Alo they're going to be talking about how we can make zero knowledge proofs more practical how to think about it for their chains and overall the things that we can be excited about in the world of ZK so please give them a massive Round of Applause [Music] good all right we have a pretty large service area here to cover so uh we'll uh we'll kind of decide as a group how deep we want to go in some of these things or uh how how many fights we want to have on stage uh but uh that one's not going to be on me that's that's up to you two collectively or you three um developing Frontier technology is not easy um we've kind of seen n it's piece of cake trivia he check uh all right um what did you think you'll be able to get done immediately and the timings were off o now I say um wondering how honest I should be about this one because uh let's try the the 90% spicy question okay so honestly privacy privacy itself so just a bit of a like history um stuff thing basically in 2017 I wasn't the co-founder of Aztec I was a co-founder of a company called credit mint because we minted credit on chain it wasn't a good name and uh so it was a trfy product that used ethereum to cut out some middleman use decentralization to get some better trust guarantees uh but back then 2017 we were very naive and I was kind of assumed oh privacy yeah you can get that by just OB fisca transactions splitting them up into little you know packets sending them around as I said like I was completely in the space and we realized okay our clients we wanted our clients to be these like billion dollar funds they ain't going to be okay with their portfolios being splashed around on chain and actually owe analysis is a problem there is no privacy what do we do at the time I thought well you know like zcash is so this right and you know the nation make made L of noises about bringing zcash to ethereum as then they had these pre-compile was like oh yeah we can we can solve this this is a solvable problem and that was that was winter of 2017 and I'm still working on it U Sandy maybe I'll ask you for for scroll uh what as you were thinking about the zkv and what did you think was uh something that's going to be easy but turned out to be a much much longer test than expected um I would probably answer this question in a slightly opposite way is that when we started off um we assume this would be a very very very difficult task building the zbm I think most people estimated the the Endeavor would take 10 years and um and and we estimated building a team and assembling ZK researchers would probably take two to three years since it was very scarce and we thought we would have to train up a lot of people and to to to get trained as a zek engineer would would take a really long time and I think we we got very lucky along the way and and found a lot of talent around the world and um and that's made it kind of like go to market a lot a lot faster and finally Alex yeah um I guess my kind of immediate reaction is is like Zach is like wow where to start everything I think once you dig into it is a challenge but I'll maybe take the the um outside of the technical side I think it's just developer ad option and applications that use zero knowledge cryptography in a way that does something fundamentally unique and and seeing what those are I think we're still to this day I certainly speaking for Alo like applications for ZK are still theoretical largely right and so it's like helping develop a create a developer ecosystem where people grock this slightly different Paradigm enough to be able to leverage the power that this type of cryptography gives you you know that I think itself I think is hard um because you're talking about a different model I feel like we've kind of brought up these thems already um so there's there's in a way when when somebody looks at ZK you can partition that into a bunch of kind of high lever buckets there's privacy there's scalability there's compute uh they're obviously very overlapping with each other um maybe my kind of first question would be for Sandy how do you think about ZK for scaling and and sort of what was the goal behind starting scroll and and for for audience let's also understand a little bit about how scroll works and just overall how it's architect it yeah sure um I think how how we how how how I started was that I I've been kind of a in the ethereum space for a while and I and was a crypto investor before and I've been quite getting deeper and deeper into the ethereum core research rabbit hole and um and I think the scalability problem has always been a kind of a problem it's been there for a few years and everybody understands there's a massive on improving overload and um the the the aha moment for me when when was when I met yeah he was a first year PhD student and he had a paper on Hardware acceleration and for me I didn't realize at that point that Hardware was mature enough to process some of these proofs and and packet it to a way where it would be commercially competitive and also uh essentially make ethereum um very accessible to a lot more people and so we thought this is a very worthwhile Mission because ethereum has found product Market fit and um it has a lot of use you know it has the potential of bringing Financial inclusion to everybody that wants to participate in the network and um and we think more people around the world deserves to kind of be a part of this community so that was part of the mission and and part of the why um we we embarked on this journey and Alex uh you're kind of effectively with zi betting heavy on compute um tell the audience kind of what the the intuition behind Alo is and kind of how do you intend some of the new complex news cases to be yeah so ZXI is you know is a it's a is the name of a paper that kind of describes an extension of the model of zcash right so zcash kind of pioneered private payments Bitcoin and Zi kind of tries to extend that in the direction of ethereum right which enables kind of smart contracts right um is a high level idea so Alo implements zexi now I think there you know actually in the beginning we looked at doing Alo as a layer two on ethereum and one of the reasons why we decided not to was because the ethereum virtual machine the way it's kind of architected is is General it's not specific to this use case and to implement ZXI perfor in a way that's performant and efficient you kind of need to you know in a in a relatively bespoke manner kind of purpose build all of the diff the different pieces from the ground up if you don't want very expensive uh you know proving or verifying in terms of you know gas costs or just real world time so that's why we ended up embarking and going out on in a different direction and saying hey we really believe privacy specifically is important like you know for for different use cases like identity you know so that element of ZK the zero knowledge I guess you know snark being kind of the part where it's succinct and therefore scalable like the ZK part is what we we really focus on for for real world use cases and that was the motivation behind Alo now of course like Alo itself is a platform for privacy preserving computation so that's how we look at it could I could I can I be slightly spicy here in the most resp way everyone look up from your phon it's about to are coming off um so you mentioned that like so um one of the reasons you decided on the L1 pattern was because the zi protocol is it's not easy to implement ethereum BEC very expensive I think that was true when alio got started it's not true anymore um as then you you because um since zi was published you know we have these kind sort of half pairing friendly Cycles where you can use Halo 2 style recursion techniques now and folio schemes and God knows what these days um to basically crush the problem down to a like something which is trivially solvable to the point where for example Aztec is like the our architecture is hasn't really got a name yet but it's very much um like d like it we see it as as as as an iteration of zexi um certainly like the zexi paper was was a very seminal paper where we got a lot of the ideas for like how how how one could Poss how can one could architect a z system um but uh yeah we we um we uh Bic but yeah like but you can get that kind of stuff on ethereum these days I think so to me so I guess it's really important though to Define specifically what are we comparing to right I think you know the architecture that I think you guys are building it sounds really promising and interesting I think we'd have to see concretely for different applications how they compare in terms of all of these costs right and what other what assumptions are you making for example I think one thing one thing that I personally um you know I take issue with at the definitions of like crypto layer ones and layer twos is I think sometimes inside of this like stack based framing we introduce new assumptions around the permissionless of Technology I'm not I'm not saying this is the case for anyone necessarily but I do think it's not quite Apples to Apples anymore when you're talking about you know a layer two and a different layer one where it's per from the permissionless standpoint for a user perspective you know it's a bit of a different experience so I think at a high level I agree with you it's possible I think concretely I think we would just need to compare things Apples to Apples and have a like an in you know academically rigorous conversation about what is a fair Apples to Apples Compares what is an apple well so I mean Zach I think you've kind of been on the L2 side before you've explored TK rollups and I guess the the iteration that Aztec is now uh so maybe kind of tell us about sort of what that Evolution was and and let's contrast that with yeah sure so so why why is as L2 well I guess I guess we kind of came to it from a different from a different end where cuz we like an evolution of a protocol that was building on a to begin with um we had I guess maybe a more of a like an the way I see it um like my approach to like developing ZK Tech space has always been that from the persp of an engineer and not a cryptographer as you I I'm wearing my engineering hat and thinking okay imagine I when it comes to cryptography imagine I'm a smooth brain I know nothing I don't want to know anything it's complicated it's messy I just want to write private smart contracts I want to do like ethereum like things with the private State and it made very it made it was ethereum was a natural home to put Aon for several reasons mostly around greed and laziness uh um oh but alignment sorry alignment uh so what is azex called competence it's cryptography it's developing um Advanced cryptography taking it from research to production and then building out products and services and infrastructure Tech that uses it in in a way that's access to Engineers we are not domain experts in um distributed consensus we're not domain experts in um the technology the technology stack that's required to build an L1 so we were like we don't want to we we want to cut that all out of our protocol um inherit that from someone else who's already done all the hard work so that we can focus on our specialism that's the the laziness part the greed part is why is privacy useful if I gave everyone here Magic beans um and you started trading around magic beans do you do you care about um hiding how many magic beans you own no because like no one cares no one cares about your beans so so no one cares about your like privacy or or transparency for privacy to be useful you need value you need liquidity and that is on ethereum at the moment um overwhelmingly and so to to to make Aztec most access like to make Aztec's unique value proposition program privacy as as possible we want to be as cited as close to that liquidity as possible which means building an onium I guess though back to that point I think you talked about not being an expert in distributed consensus but to have a something an L2 that emulates ethereum's functionality don't you need to figure out how to manage State access concurrently right so you need to have some kind of you either need a SE as most l2s today do there's a sequencer that just manually or you know man yeah manually orders transactions right and then that and that's a new assumption around permissionless right now there's a new party that exists that you have to add an assumption about that so that now it's not no longer quite the ethereum model right and is that bad or good I don't know but it's different I think so so I guess like how do you think about that in that in that set yeah it's different but it's also I wouldn't consider part of like the consensus model I see that as the state model and that's my wheelhouse that's that's my that's my jam basically like how do you how do you create a useful privacy preserving application It's tricky because take a Unis swap or pretty much any def protocol you're going to have glob shared State variables for unisoft to work you need to know the total amount of value in every single pool so that you can do price um price matching uh in a privacy preserving world every state variable is encrypted because it's owned by somebody it's private uh or maybe a small group of people so how do you get Global Shar State well you can't so then the question is well what does that mean privacy useless well no but it means that like our strong conviction is that for useful privacy preserving applications you need a high hybrid State mod you basically need the ability of individuals to own encrypted States so that could be um like tokens that um uh or could be NFS or it could be like information in a SM contract that defines your identity where that's all encrypted using zal proofs to Pro statement about it but that can be connected in a very composable way with more transparent protocols like Unis swap um where you need that that public share state to um for the protocol to work to give a concrete example of how this would work take an amm Aztec well the am will be public so the the tokens going through it the prices the orders that would all be public what would be private is the identities of the individuals that had participating and any kind of logical checks and balances that would applies to validate whether that user can participate yeah and we have a public we have a public private model as well but I think still it doesn't all it doesn't solve the problem of like who I trade first you trade second how does that actually settle on the layer one yeah so that's that's Mev as in and I think it's l2s inherit the problem from L1 as in you still have the same issue with L1 so take ethereum right I'm a validat great I get to I get to make the next block well I get to build the block which means I can order transactions however I want I'm going to order them in such a way that I'm going to be able to extract the maximum amount of value does Astic solve this no uh at least when it comes to the public State part our sequences have exactly the same problem but that's not our problem to solve uh as in I I'm very confident that that that issue is better tackled by the community building on top of these networks than actually at the partical logic in the project directly I have a question for you too this is something I've generally not really thought about because I'm taking over your job introducing Sandy other moderator for panel so so so if I'm a user um who really cares about privacy on chain uh what should be I be thinking about when I'm thinking about choosing whether an L1 or or privacy L2 what like what are the differences in the trade-offs yeah I think you know I would love to get to a world I mean where and honestly ly you know just because I so I got into this maybe just let me just take a step back I got into this space working on the Syrian Turkish border with Syrian rebels you know seven years ago when I was still in the US Army and like a lot of those people were refugees right and so I have a very keen understanding of how you know a subset of the population actually maybe a large majority of the world's population is cut out from things that we all take for granted right and you know yeah I think that's just one example of those specific so to me like I want like that's that's a population for which privacy is important right I and and even beyond that I mean like how many people in this room take their paycheck in a stable coin okay yeah very few that's about right and this is like so this is like you know answering the question that's not necessar being implicitly asked which is like why is privacy important well that's why right because we don't even use for the most part our own products because they don't have privacy right and I think privacy is an essential part of the ux I would like hopefully whether it's Aztec whether it's Alo whether it's any zcash you know hopefully at some point like these things are just abstracted underneath a you know a bunch of other layers and like people don't even think about it right in the same way I don't think about the stack that enables Apple pay to work right I don't even really think about how my phone works at that point so I think for us I I don't I actually don't like the framing of users choosing the L1 I think it's developers who build applications that people that solve a problem for users and I want to enable developers to have the tools that they need to do that and that's to me the biggest challenge that's what I why I mentioned it first you asked the question yeah I think I'm mostly in agreement there uh if I'm talking to a user about l2s and l1s I've already lost uh like what matters are outcomes like what can I offer a user um you know uh can I make their lives easier can I provide them high quality Financial Services can I give them a bank account in a developing country can I you know can I help them coordinate with their peers um perhaps you know some kind of pseudonymous whistleblowing application if I got to talk about L1 l2's distributed consensus like that doesn't matter that's that's that's an implementation detail I think and it matters to the extent that it has an implication for how what types of apps you can build right so this is like why sometimes I harp on this concept of permissionless right because I think permissionless software I think they I think that's what makes blockchain special right is the fact that like anybody like anyone on this stage can deploy a contract to ethereum right now can use ethereum that's it I don't need anyone's permission right um you know Bitcoin same thing right so I think to me that's the key piece that's why I focus on a lot that's one of the other reasons why we went the one route because it's like the closest you can get to permissionless which I think then enables different types of applications which are new and novel right and I get that's where kind of everyone is going at the limit but I you know to me I think it's something that we sometimes can lose sight of and I just want to make sure that for me at least like we stay very focused on that outcome I feel like that this um this L1 versus L2 thing seems only relevant until we kind of get to a a point where liquidity is not as as fragmented which means I kind of need to know do I have an asset on this chain to be able to actually interact with it uh but that's a secondary piece that someone else also gets to kind of control or improve but uh but I largely agree with your points I I think that's fair uh so maybe let's flip that conversation on kind of the other group which is if actually this is important for is developers uh in the case of scroll um I just have to write my evm contract if I was a solidity developer that's excellent which means I just have to be willing to deploy un scroll if I already have a main net contract um are there any other benefits to to kind of picking scroll and and sort of or I guess future upgrades that they can get uh for the benefits of kind actual CK encryption there uh or just any just CK enabl Primitives um I'll start with that and then I have similar questions for why why Noir and Leo sure um I think the way developers can interact with scroll today is whatever you've been building on ethereum mainnet for the last couple of years whatever you've been running um deploying on scroll is just a copy and paste kind of job and um if you've been learning solidity for the last few years um your your learning will not go to waste you can still continue using exactly the same things without having to re audit or retool anything um our our our current product is um trying to maximize ethereum compatibility um to the largest kind of technically feasible extent so as to make the developer's job easier and um and to me that is the kind of the biggest opportunity for L2 because ethereum has already found product Market fit and not many products in the world have done that and um we're we're kind of an extension of ethereum in terms of its technical role map um but going forward there's actually suddenly a lot of um options available to us now that we have a working product um and and there's certainly a lot more options so there's there's a long way to go in terms of improving security there's also a long way to go in terms of kind of increasing um efficiency so like making it cheaper and faster and and there's also a long way to go in terms of decentralization um but I think the the the the first instance is really important because developers can now use it and we can start kind of getting feedback from developers about what they need and um we've we've been live for about a month now and um I think we've hit 3 million users faster than any L2 in history amazing thank you we're we we're very proud of that and um and I think um and and I think most importantly that the feedback for us have been have been very real and and also um some of it has been expected and some of it somewhat unexpected but but but all of it is super useful in terms of helping us shape the sh shape our priorities for the next the next kind of couple of months and years uh I I I'll kind of preface by saying I I know that's not an actual choice that you made I think but it was never like let's not focus on ZK evm or just making everything on Aztec or Alo evm compatible but kind of walk us through like why was there a need to create a brand new language and and kind of why that had to be done and what kind of benefits there were for for this path s i me to we go ahead yeah so I think you it's not I don't think it's helpful to talk about a language in isolation because every language is kind of intrinsically linked to one or more execution environments is in the place where you're going to deploy programs written in your language where users are going to be running this program so for example for JavaScript the language its execution environment one of them is node.js for solidity it's execution environment is ethereum virtual machine what is Aztec's execution environment well it's it's the Aztec layer 2 and the and the atic network and and and it's a weird one because we have private State that's encrypted owned by individuals it's a very unique novel execution environment that does not yet exist anywhere else in the world um present company excluded accepted um uh as in you just don't have you just don't have any any um existing like cisal paradigms to base your to to base your environment off of because um you know if you have private State you need your state model is it's um Atomic utxo based um updates have to be Atomic you can't modify State you can delete it or create new ones and then you need to compose this with this with with a more standard account-based database so the semantics around how to actually construct transactions that can mod that that can modify the data structures is completely Noble and does not yet to exist um outside of the the the the projects that are trying to crack this in um on web 3 so we're like okay well the execution environment is unique the transaction semantics are unique the language semantics need to be structured to make this as easy and convenient as for developers as possible so that we can try and create these abstraction layers that map these relatively unfamiliar Concepts to what so more familiar Concepts which meant that for the language we wanted to be able to very strictly control its semantics so that we could be very opinionated about how to do this uh which means well it means we need need language uh so that's why we we came up with Noir yeah I I think Leo is I think that your answer applies equally to Leo right I think it's large Parts because it's a new paradigm I think the one other thing I would say is like there you there are languages that actually predated both Noir and Leo for CK right Socrates was one of the early attempts right this idea of like you could synthesiz a circuit using a high level languages a language for ethereum smart contract it didn't have a state model it didn't have a state well it State model was kind of implicitly ethereum I think in the Zach's case right yeah which is not the stel that we private and the public right but but nonetheless and the the the as the reason I bring that up I want to focus on a slightly different aspect of what Zach said and part of that is just like you know you're talking about you we're talking about zero knowledge cryptography we're talking about doing computation and then and doing it in a privacy preserving way like that has overhead right overhead in a number of areas like one of the common ways we measure overhead is a constraint cost right which you can kind of think of like as a cost for executing a given program right and if you have a compiler so you know language like agree with Zach everything he said around it's like you have your execution environment you have a compiler that you take you know high level logic into like you know your execution environment you know doing that efficiently can make a huge difference at the end of the day for performance and cost right and I think that's the other aspect of it and then the last thing I'll say maybe just less on the technical side more on the developer side I I I used to be in VC as well I was I was a a partner at a6z in the first crypto fund and I remember at in 2019 when I started people used to say about ethereum they're like oh no one wants to learn solidity so to everyone already knows JavaScript in fact I don't know if anyone in this room if it's been around for a while remembers there used to be this project called ew was it ewm I think it was ewm right it was like the idea was like no one wants to learn solidity they're like so we need to figure out how to P like people you know other people writing from other languages and and that now is of course kind of defunct right and now the new narrative is like well solidity everybody already knows the end of the day solidity is a good language for what it is right and the you know we're trying to build Norah and Leo to be good languages for what they are at the end of the day developers my view like developers don't want to learn a new it's it's not like they don't want to learn a new language you just have to give them a reason to learn a new language if there's a reason to learn it whether you're a developer or not you're going to invest the time on it but if it's just for no reason well you know no one's going to want to spend their time to do something for no gain right but I think in this case in this model the only way you can get privacy preserving computation because of the different programming model that Zach mentioned I think you do actually you can't really just take uh an existing Paradigm at least if you want to do privacy preserving computation in a permissionless way so I mean maybe for for Zach and and Alex what do I need to how do I update my thinking around what composability would look like in this scenario for the case of something like scroll it's it's kind of obvious we know kind of how to interface because everything is on on the evm but if you ever want it to be cross-chain compatible or even kind of evm um adjacent does that break does that mean only certain things are possible or maybe what's not possible what does that kind of feure look like or did you make that trade off saying oh look we're not going to worry about composability just yet well we we care greatly about composability of contracts within the Asing network not necessarily cross like to cross chain as in could you transpile evm like solidity contracts to Aztec contracts Well yeah if you didn't want privacy but if you don't want privacy yeah I mean why are you using Aztec to be honest it's uh um we've made some very distinct tradeoffs which mean that uh yeah if you just P Focus if you just don't care what frzy at all then you're going to get you're going to get cheaper transactions elsewhere we'll be but we care about in like um intron Network composability and that was extremely important for us composability composability composability that is I want to write contract I want my contract to be able to call other contracts um and uh to do this in a completely famous way and that's very tricky in a ZK context uh again like I I I I looked at the problem starting from an engineer like I want to be able to literally like write a contract inart with functions there's functions that call other functions in another contract um and that a transaction can be uh like an arbitr long sequence of function course uh and that is I think as Alex mayre said exceptionally difficult problem uh but for us it was important to solve that so that we had our intuitive developer experience and so the way that Aztec does it was we have private and public functions in Aztec contract so private function modifies as private state public functions modify public account based State private functions can call public functions um but not vice versa like those are the heuristics and there sorts of reasons behind those ristics but the idea is we want to hide that away uh and then we use some rather um mind um uh bending uh Moon mat psychologically disturbing techniques to uh make that all um provable within a within a within a web browser um through like heavy use of Z russion but yeah that's um yeah like I think we've we we'll say when we go where we go live menia but I think we've cracked the problem of like composability of contracts within the assing network across different networks that's another that's another story yeah and and uh we have a similar Paradigm so I think you know like public private I'm not going to dwell on that as much I I think I'll maybe just mention that I think composability cross Network I think that's one nice I think one cool application for zero knowledge is you can have you know I think I think it's almost a more it's a better or more secure Paradigm for oracles generally right I think not only can you potentially interface with other networks like etherum and how would you do that well you know you can prove you know you can basically prove the latest State based on a state transition function of like another blockchain inside of the Alo Network this is kind of how bridges work generally today right the addition here is that you have a cryptographic proof that you know so in some sense it's more secure than the multisig way we do it right but in addition to like having cross state or state move between blockchains this way you can connect to like real world systems that use cryptography like one project that I think is criminally underrated that I recommend everybody check out it's actually I think an ether an e Foundation grante is TLS notary right there's a bunch of little examples where you can take like you know anything that has a you know it uses TLS like or like a signature from you know from the regular web and you can verify that on ethereum you know there like really cool ways where zero knowledge can be used to interact with the real world one project that we have built internally to kind of show this is uh something called zpass which is actually an implementation of a paper called ZK CR and part of that protocol requires you to verify a signature of you know from a issuing authority for your passport right so you basically like you're basically digitizing a passport and in that process you're verifying that passport is valid via verification of a signature and you're proving that verification in zero knowledge right so I think I I to me like that's an example of how you can have these composability with outside systems um you know using zero knowledge I think uh great mind s like we have an external project team working on a similar thing foras exek fport just to just to like to to just reemphasize what what Alex because is I think is going to be Paradigm shifting um so yeah you know passport modern God Jesus Christ passports yes um yeah the original nonf funable token don't wor don't worry um yeah these things have chips in them uh modern ones that you know if you bring it next to your phone you you can using um RFID your phone can trigger a digital signature to be signed you verify that there's knowledge proof Tada you've got some uh mostly permissionless way of verifying your identity this extends to things far beyond passports you can even use like bank cards as a very primitive form of ID again you know someone has a us or UK bank account or at least they have a card that does that yeah any anything cryptographically signed can be verified in this model which I think is really cool I think people are just starting to scratch the surface of what you can do there yeah um I can give another I want to ask something you Sandy um CU actually I want to focus on this is the good part which is let's talk about actually what interesting things we can build and now that we actually have the the infrastructure ready um one month in to the main net how are things going what are some of the coolest things you've seen so far or or even people that are working on them and uh that you think will kind of be impressive to everybody else in the audience yeah that's a great question um I think I think after the first week of launch we got to a stage where um I'm not aware of everything that's on scroll now which is a pretty good problem to add so that I I get told all the time that you know there's this this project and that project and uh so I I don't have full visibility um but that there are a few things that I think are are quite cool that that that are that are coming or or almost here or have been running for a while um there there are a few kind of AI bought enabled kind of um kind of cat Bots that are happening on chain that are helping people learn to you know work their way through their art documentation and to there's this like negotiation game where you can negotiate against a s bankman free kind of character um and um for your for your salary that's hilarious and um and and there's a lot of um there there's a few um projects from Turkey actually that I want to give a plug to so there's a nft Marketplace called cheese art and then um and then there there's a few wallets and uh looking for and we're looking you know we're looking to support more and more projects from Turkey um and I think um of the ones that I think are are ZK interesting um there are a few kind of um ZK applications that are working on the more established blockchain use cases um that that I'm aware of that I'm I'm expecting we we'll get quite a lot of traction from the existing players and ecosystem that already has found a business model so there are um super super fast um kind of l3s using uh zerge proofs um on scroll that that that's coming soon and then there's a order book Dex which is also a you know very well established kind of defi primitive um using zero knowledge proof it's also an L3 using this where we're sharing like you know the provs and um it's going to be probably faster than certain kind of orderbook dexes that recently went to Cosmos um and um and um and I think there's a and then there's a um there's there's an L3 that is working on like selective privacy disclosure of ins ions apparently um institutions have given them feedback that sometimes they want full privacy sometimes they want to be fully transparent on chain sometimes I want to be a little bit in between and uh I think there a project that's kind of working on some like tailor made solutions but these are kind of the early early shoots that were I'm I'm keeping an eye on awesome um for AO I know there's been a lot of experiments around just kind of data computations anything that you're able to share on what's kind of super interesting to you personally that you've seen on on the network yeah I mean I guess the first thing I'll say is like I'm really excited and proud that we're fin at the point where this the network is going to be live in January um I it's already actually mostly live um so the only part of it that's not permissionless is the consensus we've only got about I think half of the validators are external and half or internal um but yeah I think right now you know people are to kind of to s's Point people are deploying things all the time on this and I'm kind of only dimly aware of all the activity I typically find out after the fact um but yeah the use that I'm personally most excited about are the identity type use cases and then that that I think to me that is an essential precondition for building payment applications that are both compliant and private as in like actually usable in the in like the real Financial world right so um so you know like I said zpass is an example you know of something that we built someone else took that RS because zass is actually an SDK someone else has taken that uh and made basically an age verification um tool for the web which you know lets you filter content by you know if it's if it's age restricted or not which is cool and then we're working with a team called Arch named Arch block which I think used to be trui and uh working with them to do um a cbdc issuance for uh you know for for a country in the Gulf which I think is a really early experiment but I think it's really cool because to me I think that's the to me the biggest use case for blockchains is still are still financial and I think a lot of the reason why we don't see more adoption there as a it's transparent today but even if you solve that problem there's still concerns with compliance and so I think the combination of being able to you know achieve compliance through whatever logic you want by a smart contract in a privacy preserving way that's that's what gets me excited nice I mean for a whole country as an experiment for for issuance that's not a small thing so this sounds like a well very potential but you know to be clear it's early days for Alo and it's still early days for all of us I mean it's it's we've all been going to these conferences now for years and like we're still it's you know we're still not at the point where we've changed the world yet and you know but the the the developments that have happened in ZK certainly over the last 5 years I mean you know this better than me I mean it's been it's as much has happened in the last 5 years as did in the previous 30 and I think that's the exciting part to me about the potential of this technology and that's the Common Thread between all of us and you know it's going to take all of us including everyone in this room to actually push this and go beyond what's just a cool thought experiment and take it to something that really impacts real users uh I just want to just make sure if somebody has a question in the audience we we'll have enough time to just do two questions if you want to just line up we'll kind of adjust the rest of the thing accordingly but I have one question to go but I'll let Zach answer that question too thank you yeah thanks yes so I think um I think we're very aligned here uh why is PR useful privacy is Identity the sum of all your monetizable information is who you are as in that literally isn't how every every every other organization institution perhaps even people will view you through that lens and if if you can have onchain representations of your identity that are hidden where you can prove statements about them um that unlocks like that once propably properly executed will change the world because it will enable finally for like onchain assets to move offchain assets to move onchain and actually have digital representations of every single G down asset that exists from you know mortgages to shipping certificates to to dodgy ultimate Bill loans and then we can actually start to create a global settlements layer for the world's assets not digital assets just assets and that's going to change things where I think um where I think that something something is often missed is that I think when what we're doing with these private privacy preserving settlement lers is what they really are is we are planting a seed just a seed whose Roots will sink deep into the Earth where the actual onchain settlements component of this ZK composable privacy teex stack will be tiny uh because there's so much more value that could be added by ZK um that will where you have the much much smaller transactions that are much more vendor specific application specific that will eventually only Bubble Up and Bubble Up and eventually some of these will get settled on sh but most momently to give an example of this um so I think Bri from X Park said it best it enables ZK enables composability across data silos not just applications or web just so generic con of a data C let's say I'm I've got a web service and I want some information about your Facebook account and you're willing to give it to me well I could ask you to you know send an A A ping request to Facebook by the API and you get back at Jason Bob and you'll send it to me I can't trust you though well to be more accurate I can't trust the software you're running and also that data is going to be formatted according to e like Facebook's weird API that I don't know or or don't want to reason about um what do it enable enables data transformation um pipelines where you can take information from one data Silo and provably extract it by you know verifying DLS signatures TLS that kind of thing and then transform that data into a different format that is suitable for consumption by the requeste just that that basic primitive it it unlocks so much that we are currently missing and yeah I just I feel like we are on the cusp of some foundational changes in how we interact online uh and just how communication networks um uh can be CAU structured that that are going to going to change things that is a an amazing intuition around kind of you effectively zero knowledge will let you get proof of any data and you effectively abstract away the fact that I don't need to know where it is and and those cells are are kind of broken that's that's a good way to put it um well in this case let's end on something that you actually disagree on potentially um me that just launch for scroll Alex just announced that uh you're going to main net in January which is a big announcement Aztec is on its way to be up there sometime next year what what defines a successful mainnet launch uh how do you define things being decentralized how do you define things working is this is this the right approach uh we've kind of seen some uh some thoughts on this so maybe uh maybe I'll let Alex uh I guess this was directed to me tell us how are you thinking about your main net the context under this question is you know I uh I made a comment about uh Scrolls main net laun launch specifically on Twitter around I disagree with the framing of what mainnet means now that comment was not meant to criticize Sandy or scroll or and by the way I've made similar comments about other l2s before start whereare being most than most notable it's not it's not to criticize it's more of like to me what makes this space again back to going back to what it defines a blockchain defines blockchain technology like what what is the reason we're all here it's this permissionless technology started with Satoshi Nakamoto right Satoshi Nakamoto develop Bitcoin why was Bitcoin special because it's a peer-to-peer was like peer-to-peer system for electronic cash whatever the title of the paper was right and I think to me that permissionless was one of the Reason by the way like Bitcoin was preceded by about 40 Years of attempts to make digital cash right and it took 40 years to get to that point it was a hard Endeavor we've all talked about this endeavor for us is hard right ethereum to go from ethereum 1 to ethereum 2 took years right and I think to me I really want to make sure that we discuss these things and the and the terminology that we use reflects that that difficulty and reflects the features that we that we should that we should collectively agree on it shouldn't be one person's opinion you know and so this is why it should be a community discussion but to me my opinion is that you know when we say mainnet and we talk about ethereum as on mainnet to me that means this is a permissionless network right I think you can deploy software on top of main Nets but to me those are not themselves main Nets so that's my personal view um and again part of this is is not just as a member of like Alo and I'm trying to score points on people it's just as a member of the crypto Community Because by the way there are a lot of people outside of this room with a lot less Scruples who say things are main net and say things have a million TPS and say things are privacy preserving I mean we see these announcements all the time and sadly it reflects badly on all us cuz when the person walking on the street out there hears about crypto they might not hear it from you or me they might hear it from one of those people and then then they get Jaded by what this technology is and what it can and cannot do and then they decide to write the whole thing off and that to me is the outcome that I personally really want to avoid and that's to me why I really push hard on making sure that we all as a community agree on and understand what the standards are for using certain terms um so I want to applaud and also thank Alex for um Coming forth with standardizations about terminology because I think out of almost everyone I know of an industry I think somewhat agree that scroll has someone on our team who is very very particular about terminology and insists on calling out other teams on these things um just taking a step back for a second um I think Bitcoin is obviously you know a a fantastic project and that's probably why we're all here and it's great to kind of think about you know how can we be more like Bitcoin in terms of our permission standards but there isn't a single definition about what permissionless means there's no standardization about what decentralization means and there's also nothing about what security means either when it comes to you know l2s and l1s for that matter and for all of these um slightly inor first terms to people of different kind of levels of technical understanding um there there could be a huge range of what this actually means and we all all carry with it our own kind of um assumptions and or our our thinking into this um I think to to step back this that no one can really get to bitcoin's level of permissionless when you think about any kind of permissionless Standards like you can't if if you're a human person like we know Alex is a human being then we hope test with a extremely high standard of mayet definition um um like you know essentially um for for for for for for me mayet is a stage where any project feels comfortable enough to let their developers and and let people external to the team interact with the chain and when the when the block is confu uh is produced and and and when we feel like the security standards at a level where we're happy to kind of present it to the world and um and I think there could be you know multiple layers of definitions about what this means and you could you know go into it about how like you know how l2's decentralization um U and and and security standards are hedged on L1 and um and about various training wheels uh this conversation could be potentially endless and um and I think on on one hand I do think every team should Define their definition a little bit better I think we definitely make regular attempts at defining certain terminologies and um but I think um you know I but but I think I'm all for a a process where the team is being transparent and um and trying to get feedback from devs as early as possible and also iterating on that feedback so that's that's kind of where I stand philosophically that was amazing I think that deserves uh was an Applause yeah so so maybe I'll give the last word to Zach as Aztec V3 launches how do you think about essential ization for for how do we think about decentralization all right so watch him negate all the associate things when okay I'm going to go on a very slight tangent but I think it's relevant that thank you networks have intrinsic values associated with them um what do I mean by that okay so I'm going to lean into that this Roman Empire meme how many like you know how many guys think about the Roman Empire all the time so imagine um take the the road network of the Roman Empire if you showed an alien who had no contexual knowledge by Humanity that road Network I would wager that they will be able to conclude that the culture that built that network was a hedonistic imperialistic culture why is this well if you have a very like um Advanced well-developed set of Roads then that inies enables trade and commerce but it also eaches an enormous amount of potential and stability because it means that armies can now March down those roads very quickly enemies they can launch hidden run attacks and sack cities very easily before reinforcements can arrive which mean which would then completely undermine the value of the road uh so it wouldn't get built therefore if you have a sophisticated Road Network that spans a large geographical region there is only one Army marching down that road Network so you can also tell that it's an imperialistic culture because the road network has its Beating Heart at Rome and then you have these arterial roads that are radiating out but ultimately or Ros Le to Rome so you so you can see that that network is also enabling a singular singular entity to project force and power from a single point so now let's look at Aztec that networks values ASC is going to when ASC hits mnet we are going to be fully decentralized by that I mean there's going to be no actor that can control a network or Force upgrades there's going to be no actor that can control who can interact with a network it'll be completely permissionless anyone wielded contracts anyone will be able to send transactions um we will have strong guarantees against censorship attacks through having a fully decentralized sequen Network and a decentralized proving Network well these are both claims at least I'm talking uh I'm talking a big game we'll see if I walk it but I'm confident that we will not because of Any intrinsic not because of Any values or ideologies I hold but because the network requires it like azc as a network cannot succeed or survive or Thrive if it is not fully decentralized from day one um for a variety of reasons that are that are relatively unique to privacy preserving critical um and so even though I haven't arrived at that conclusion through particularly IG logical means if I could launch centralized I would it's a lot easier uh it shortens our it would shorten our road map uh but I can't uh because of the intrinsic values the network has forced upon me I think that's an excellent point to end this conversation this was a great remarks and and thank you so much for that thoughtful chat please give a big massive Round of Applause for our three panelists Sandy Al he thank you very much Carter for hosting