ZKPodcast: Exploring Aztec with Zac Williamson

Transcript

welcome to zero knowledge a podcast where we explore the latest in blockchain technology and the decentralized web the show is hosted by me Anna and me Fredrik [Music] in this episode we sit down with Zack Williamson to talk about zero knowledge tools range proves the Aztec protocol and the development of zero knowledge standards before we start we want to say thank you to this week's sponsor noise fund Nouri fund is building an IRC 20

compatible open-source technical environment which enables the tokenization of real-world assets assets like equity in a company or real estate northen does this by implementing different basic protocols that bind off chain assets with their own chain representations they are currently looking for a QA lead to help them test their daps back-end in smart contracts if you're interested check out the job offer at no fun org slash careers or review knife ins github at github calm slash boyfriend all links will be in the show notes so thank you again my friend and now here's our interview with Zach from Aztec so this week we're sitting down with Zach from Aztec protocol and

we're gonna be talking a little bit about what they're working on and a lot of ZK stuff so welcome to the show Zach hi there thanks Rosa me it's great to be here and as always be a Frederik hello we always start with a quick intro from our guests but I saw this comment on Twitter that sort of said why does everybody always start with just like please introduce yourself blah blah blah and I kind of I felt that and so actually this time I want to try something please do introduce yourself but also maybe just give us a little bit of insight into what inspired you to get into this okay sure thing so quite well I'm Zach and I am the CTO of asset check the asset critical and we are at

previously provided for the theorem where we allow our tools and technology can be used to create private representations of existing public digital tokens as for what really inspire meeting into this laziness to be honest pure pure indolent laziness to tell the honest truth originally I was working on building a syndicated loan platform for aetherium using smart contracts to purchase great financial instruments and we did a privacy from product to work and we were going to get it because there wasn't an existing provides a solution for hearing or other kind of trick complete public objects and when I realized this I started digging around into zk f ck

snarks and how to looking at the kind of work that would be required to bring a Zika snacks worth earrings at the time things like Socrates were there it's very young in the in the very early stages I have to be honest like when I when I first started the game Susie questions to the tech I started reading the papers I thought my god there must be a simpler way to doing this because effectively what we were trying to do was prove extremely simple statements and zero knowledge things like you know there's some of the input signals are some of the outputs like I know is he Christine the secret is plus an obvious equals equals at like 10 theory really simple statements and so the kind of the general purpose solution provider basically stocks felid kept cracking it up of the session

and I said yeah as I started digging into the tech behind this looking at the culligan arithmetic and range loosing all that jazz and realize that those probably at kind of a somewhat simpler solution specifically for this new song so you started at almost like a higher level of trying to solve general financial instruments applications stuff and then dug into the privacy parts have you gone back up to the financial instruments now like or are you still at the privacy like where are you and where do you want to be of course when it started putting together this your owns proving tech that we needed for our silicon lim platform really the

overriding thought was wow this stuff is really cool this time and that this is a lot more valuable than just what we wanted to use it for so we as a start-up we have primitives to basically providing the privacy solution that will enable other others other companies institutions to basically build what we were trying to build at the time was supposed to called credits what's what was it called credit men Christmas credit MIT okay yeah we don't we don't live a little bit of a pivot foot to propagate our privacy solution but we have kind of climbed back up the stack and so the we're putting together tooling zeros like construction and smoke mushroom verification engines they're focused very specifically around private transaction settlement make

trades both different private assets between bosses okay so let's find out what Aztec protocol really is so this is the project that you've been working on for how long I'd be working on privacy tech since the started twenty eighteen sixteen has been well over a year but you know and when today's tech start a sec really got off the ground in November last year that's when we publish our proof of concepts of the ESIC critical and created a private version of the dice table coin on if they remain lettuce and traded it around and she lost per second announced that hey this is now a thing that can happen when I was looking at our text the question that first came up for me was

how does this differ like how is Aztec using ZK for privacy different from something like Socrates so others fundamental level as take the essa critical is not based on an AZ case not it uses a very different kind of journals prove that does utilize bilinear pairings like it's not but fundamentally is constructed in a very different way and what this enables is effectively it's a lot closer to a bulletproof type system than it's not tech system in that you have a way of proving that a commitment is bounded within a range and then you can use

traditional home of arithmetic focusing around things like Sigma protocols to creates algebraic statements involving these proofs where they're kind of the algebra your your real Bell anything is like directly expressed in the exponent of an elliptic curve instead of having instead of reducing turning your elective Tonio program will suitor statements into polynomial equation then proving that you know a satisfying assignment to that polynomial it's a it's a kind of it is this a simpler form of tech so I do mind of it might take a step back to really answer that question because I British this may be from the opposite ends of things which was what all this thing to do and what's the most efficient way of getting there and what we wanted to do was create a kind of a private token standard private the way a

method of representing value providentially where users had agency over the specific semantics of these confidential transactions so that if you if you're building a private asset that has for example edge cases like say where you're building into central extremly of administrators policy community don't have frivolous lives or you're building a private security token and you need to prove that every trader has less than 10% of the total supply registry reasons you need that flexibility but it can't come at the cost of compatibility and interoperability because for example with the traditionals he gets up it is before sonic came along if you would coding up the kind of insurance the semantics of a private transaction

inside a Zeke if snark for a set of private assets another would really be able to communicate with one another because they would none of them could trust that each other setups were form correctly and this would be a nightmare buildings for example for a couple as journalized at which interfaces with these tokens and so we need the rule of our private asset all private assets using our protocol to use the same common reference string the same bytes it's basically that require inaudible from first suburbs so that we can create a standard that everyone can perform to you but that's kind of still need flexibility for these edge cases and so with aztec what we do is we use bilinear pairings and our kind of a stick commitment function which we can go and substitute to perform an extremely

efficient range proof and then what we can do on top of that is create modular single protocols basically it's the civic discrete zeros proofs that before discrete bits of business logic like for example like a dividend proof to prove that one value as a percentage or another or a join split proof so you have some impact notes and some happiness in prove that the encrypted values of the some like somebody it was equals or something and what a user or developer republican do is you they can when they're building a private asset can I take these off the shelf proofs and kind of fit them together in a modular fashion to create more and more expressive statements by the natives in zero knowledge are these are these like I mean are you building libraries are

you building what what are they a very good question so we we're building at the small contractor who were building single the cryptography engine which validates the family of a stick proof so we're the moment we have five but we'll have a lot more over time and we're building some libraries yes that people can install the client side of the moment we're building the building at services solutions as well that's constructed of quite proof so the end goal for all of this is that if you're putting together a privates and privates asset or some some service you can use a loud music oh hey I won so like I make me a joystick or make me a dividend pretty falling you know I there's a minimum trained special so I wants a

registry in the same way that you could very easily efficiently CDF say signatures using the library without having to know the natural like mechanics of how it needs a day is a signature work we want our stuff to be intuitive to the point where you don't need to be literate and zero knowledge cryptography teasers and yeah witness so we're building out us our standard library that I wanted to dig into more of the range proofs and what these proofs actually are so you mentioned these arranged because it's sort of similar to bullet proofs but as far as I know I think reputable proves our range proves as well right yes yeah you think you can use very some for arbitrary computation but the main location is

matrix so my question is like what our range proofs mathematically and then I'm also curious like do you consider what you're doing to be research are you inventing new methodologies or is it engineering like are you taking known mathematical constructs and creating code around them really as both its this nexus of cryptography research answer for engineering using cryptography as it's a sole problem sets operational when it comes to the range proof it is do you research it wasn't something better the shelf because there wasn't really anything off the shop I could take for the solution because we needed something that was much more succinct than the

existed unless you like that move of using a Zika snot at the time I really I really really wanted something that used the same common reference stream from Oslo proofs and snugs couldn't do that um the way I started was I started taking around with looking at digesting good luck marry papers kind of stalking the halls of Google Scholar and shaking down crypto fitness for the misfits of the taken IP the Beckett years and I you know March I came up with what I thought was a very efficient and succinct range group oh my god couldn't touch with Ian's growth here at the summers and he said you see oh I know it was like never like hey ends I've got this great range proofing succinct and efficient and what

do you think and he was like what doesn't work it's got a problem and yeah looking back if I had an amateurish security flaw but I so I went back beep it away at it and then came back again I'm like okay what right okay second I'm worried about this and he was like so that was fun but basically at that point read them so yes degrees she joins as an adviser for our company and I came up with him with for third time I'll finish reading the paper okay what about this fashion and he was like okay maybe but you need so you need to either you need a bit of samus food and then that was

kind of a something very issue at first as well over the ends to really like refine this the the asset paper and cleaning up to the kind of standards required we prove this like its completeness of Samus and honest verify as your knowledge properties and so yes basically this is new research I may take stuff that did exist in some form like it's a special case of polynomial thing the scheme uses the stuff as existed for a while we combine range proofs with a commitment scheme which hasn't been holds really Dean double for in this way we're effectively we praised what's like something which is quite similar to a pedestal

but the difference is within our circular lens if you want to prove that you know an opening to masculine so you want to create a general fixing here as here is an Aztek notice I know what it's worth but I'm not going to tell you what that proof it's basically the same as proving you know an opening to a better Singh commitments with a very very tiny modification which music by being able to present that kind of proof it is implicit that you know that the that of integer that you're less in the note is constrained within a range well I just I wonder have we explained exactly what an efficient range proof is like okay so what about Reggie is why our range proofs a thing why they needed what the hell is a range proof when it comes to private transactions and being able to

prove that you the validity of a private transfer range fees are extraordinarily important whether prevent double spending so what they do is they enables somebody to present an encrypted number to a verifier and prove that whatever that number is it's within a given defined range the the very basic result will be like for example imagine you need to prove your your age you're making is pretty over 18 be able to reveal so the verifier precisely how old you are that kind of stuff the reason why it's really important for proa transactions is because private transaction systems confidential transaction systems use something called holomorphic arithmetic which i measures copper Lots on your on your vodcast

being able to evaluate mathematical statements over encrypted variables as if they weren't encrypted you can use it if occur cryptography to do this but one of the problems with this is that the arithmetic that you're evaluating it's not over normal numbers it's over introduced at are modulo a large prime number so in this weird realm of commonwealth of arithmetic -1 isn't actually -1 -1 is basically it's is actually an extremely large positive number so if the if the home of the rhythmic was all modular 7 but minus 1 would be 6 and this is a huge problem because it means it means having several spending if you account for it very very

easy because for example imagine I want to prove that if I have a transaction where I have I'm sending you some some value so much I'm sending you some credit i I have ten private eye and I'll signify normally I would create an oath worth five and give you a net worth five and then my output sequence but now vision imagine my inputs are zero and I give you a dino worth one and I retain a dino worth minus one obviously my accomplice one is zero but under hamat well in the world of hormone arithmetic when I've done so I'm actually just created a note that has an extremely large positive number I basically create more value then exists in the observable universe which is a huge problem for a cryptosystem enraged we've stopped us it deals with the fun

with a problem that's you can't tell the difference between a positive and a negative number in a halal figurative assistant like the ones created over elliptic curves and so rain troops are used to completely remove the possibility of negative numbers creeping into your into your crypto system what's so there's like a family of range proofs I guess and you mentioned one bullet proofs is there are there other sort of projects or concepts that we might have heard of that are also considered range proofs so there are a lot out there and I wouldn't be able to give a full account of all of them because my anomaly doesn't complete so Hazuki stalks most of them contain various

other than the some episodes are some level purser examples e caches join sketch transactions they also rotisserie require view of what's an issue and they see it's not missing so this sucks every York icily providing a range brew there's bullet proof there's also I don't know I'm not sure why it's called the spirit disk in militias called the folklore bits commitment scheme sounds like it's because there would nobody can find a reference to the original bits commitment scheme because it's so old yes weaved into the oral tradition in history there's even logic ripped ography Wow I feel I feel like I should get like a like a like a like an illusion of Queen sauce drumming some some 32-man I'm talking about the

folklore bit commitment scheme but basically it's a way I would it wouldn't be out of character for the CK community wizard hats I know it I'm sure it's all good it's a reality interpreter in the folklore victory whisky basically you you just you provide a encrypted representation of every single bit in your number and you prove that your each encrypted bit aside 0 or 1 and then you can get compatible together to create your actual number that's traditionally how it's done that's not it's not very efficient and the Crucifier huge because if you want to prove a 32-bit number you need 32 individual discrete signals groups with all of our inputs and that you can get pieces of this the you know

dozens of kilobytes large and the verifier competition is extremely intensive bulletproof computer this massively massively massively enough is and it's a naturally amazing system where you can use it basically takes the traditional focal Bikram's you know compresses it down so the size of the proof is logarithmic with the number of bits in the number and does is through a kind of proof recursion so instead of proving that you know an opening to each and every single bit you true that you know basically an opening said the opening before with correct or something like yeah you have a vector of

bit memories so at the corporate groups this is kind of a vector product where you can take the product of two vectors and prove you know a given commitment as an opening since vector products and that proof is all worth making the actual length of the vector which is an enormous and through best got it so what is aztec doing differently then so what aspect does differently is it uses the moment what is really a unique range proof and a unique encryption method of encryption to create extremely short zeros proofs that focus very specifically around private transactions and settlements and private fundraising transactions it uses the range we have to make these priests very efficient and

it uses the fact that we don't be only want us to set up to create a large set of proofs that the fun of the goal of a stick is to present a solution to developers builders companies that enables them to create private digital assets without them having to know any of the lick of zoological ography so in that in that way it's kind of is the other scimitar was equities the tragedy but we approaching it from a different angle which is instead of giving people the tools to build their own zero knowledge proof so we're giving people the zero enormous proofs that they need to build their would like to solve their own problem solving problems well how how small are you know small proofs and when

you say that they're efficient do you mean size or like proof generation time or verification time or all of it right so so to answer that question it's a bit it's complicated because the range proof is extraordinarily succinct and shot and then the concern of logic that we may use this range principle that stuff is linear in terms of its proof size and the amount of verification time required then proving this because I'm required however the kind of the constants like multiplicative factors in that linear in that linear relationship is extremely extremely small so the range proof in terms of proof say is I believe at last the last count is about 700 bytes for tradition like a

typical joint split transaction to embed notes into album notes it does vary though quite substantially him because I mean if you leave your sending the pictures of our contractors auxiliary information not relating to the proof but you need useful near the information or backbone for aetherium so does that count as a proof size also when we send these transactions to etherium we don't compress the elliptic curve data with a traditional zero knowledge proof when you're constructing your transcripts you can represent a little curve point with half the data that you actually need because one of the curve that you can represent the white border has a single bit because the exponents any given x-coordinate Williams off to - Deker points and you can differentiate them with a single bit but right there him we don't profess elliptic curve points

because the amount of gas required to decompress the points is more than the amount of gas required to actually just send the uncompressed point in a sympathizer so that increases our brew size but it's not like we could make small groups if we had a launching platform where computation was cheaper but we could get up your some stuff like furniture bytes or less it's it's it's they're generally a little bit larger than a growth style as it gets not clear when considering construction that's one of the real key strengths of a snake is Griffin construction is trivial because you don't you're not performing this some pub like this reductive this you're not taking around one constraint system and then creating a satisfying polynomial

you're not solving the beam like 1 million duty polynomials you're just creating these resettle procedure protocols where the rate for if that's the really the hard part is basically taken care of because you just you use a lookup table really just select an element from our common reference spring and if you didn't have that lookup table like the max competition we we need to be quite large in a way I believe with the full with the full version our protocols it's accurate to say this both proof construction and pre verification is a level of sickness to both sides of that because because of the fact that we basically trade computation to storage by with these pre computed tables so you have this system that can efficiently

prove things you have the theory of you know private transactions on ethereum but how do you go from that theory it's actually having a working system there there's obviously engineering challenges along the way there what does that look like yes thank you that's a really good question there are immense challenges and this was the topic of a talk that was getting at the fantastic zero knowledge some of it that you're hosting lost month well in my word there are some challenges what we're doing is we're naming other people to put private assets or on top of the theory so that adds an enormous extra different level of difficulty with traditional journals

proving tech because well how do people normally interface with etherium DAPs applications through web browser or through something on the phone but definitely right now is the web browser which means that if you requires your knowledge technology it needs to be able to be performed inside a browser you using webassembly and that then means that the proof construction has to be extremely efficient because you take true detective traditional as it gets not prick with a large number of constraints but gets difficult for some of these crews to perform them on just like optimized code compiled down to form for computer processor running that webassembly that's an extra complication but what's the point

it presents enormous enormous UX challenges because interfacing with a blockchain is already quite difficult from a UX perspective you know you have private keys and accounts and weird hexadecimal strings flying around and then you need like maybe you need to build some countries showing you some solutions your customers don't need to know what the private keys are you take the theorem layers and velocity transactions and Craig digital signatures it's already really difficult and then on top of those our users are going to ask the required day users to build some construction or large proofs and we can take even expect our users to really understand the proofs that they're going to be asking they use it to construct this presents enormous your ex challenges and so what we are doing is abstracting away pretty much all of

the journalists that's the goal it's all working progress without ruining but the goal is to create a standard library a JavaScript library you can plug into a browser that's where you have very simple API calls to make the required zero inference ideally this is also interesting segment mask integrations and traditional integrations at a high level what's really needed here are standards when someone a developer builder was building a sec they need to know what our zone or scripts do didn't show how it happens they do what they do inputs that they need and the APIs that must be expected your you're building a system that's for aetherium and you're

building in these libraries the JavaScript side and everything else but you're still you're still only targeting theorem you're not like we had matter labs on recently and they're planning on building this whole separate network that deals with all the things that are doing like proof generation off chain and and etc etc you don't have like that off chain component rights like you're you're trying to do everything on chain precisely yeah we're not with other one we're not really targeting scanning solutions we're very much focusing on how do you get how do you represent value privately on chain that the load layer one maintenance level because we think that's very valuable array needed for lot of applications and watching

because our tech is designed to be used by people building right there in you want to make the absolute bare minimum requirements or expectations on them so we don't want them to have to be able to run some kind of parallel Cheerilee network on side chain and we don't want them to have to get their users to install clients just fast and we have a lot of things to do that because we're not doing scaling we're doing privacy and we do have the reserve the tag to do this all menu by small contracts because of our seconds of this is think that's just a verification the question of standards comes up a lot in all sorts of fields but in zero knowledge I know that there's like even a meeting I think that just happened all about standards right

what's your what's your take on that like are you participating in sort of these standards talks that's a really good question um yes I was at that journalist stones workshop that was in exactly I just just last week recovered from the jet lag coming and it was a fantastic workshop in and there was an enormous that effusion of ideas between a lot of different groups I think one of the greatest takeaways from it was yes we need standards for not photography but there's there's not a lot of consensus around what this tender should be and right now we're proposing etherium specific standards um when our own technology is more mature we'd like to

convert those into more formal stands that did justify to the clear mulching but right now there is there is no uniformity at all which is a business you back to that event so like what it when they when you talk about standards in that context are they academic standards like what is it only for blockchain like what conversation is that even about we've actually never really talked about this on the podcast and I know it's a topic I'd like actually like to dig into just when when you're talking about a sort of combo academic corporate projects open-source ideas where are the standards where do they live and what

are they for so it's it's again it's kind of complicated because there's no you guys know consisted on water scientist in this field so at least I mean you know some people you get ten different responses um it was more as an academic level than like an application-specific level because a lot of people that are older teens are all using sniffing or these but for very different purposes so you give an example of some standards that i've opposed for example a standard for composing and missing rank one constraint systems a little quiet physique szq stunt groups so that you can have multiple languages compilers physique it's knocks it open crate least there's the part of the same rank want

the strengthened system portly you can then combine two proofs across multiple different elliptic class other exams for example is right now the community is deriving a lot of new elliptic curves which is quite quite a new development because there's an enormous demand and desire to perform elliptic curve cryptography operations inside of Bezique it's not second and if you want to do that you need to very specifically choose the parameters of the elliptic curve that you're doing or if it's a con up to the point where the existing missed curse is none of them are suitable and so people driving their own cats and so there was a certain workshop on well hey can we can we formulate some standards around how to derive elliptic curves cause it is not because it's not

easy you can you can you can is crazy to get that completely wrong and basically just make a completely insecure because you've chosen your parameters poorly it sounds about that and that was kind of the those are the current narratives at that conference much more than a high-level so given that you are actually focused very much on bringing privacy to etherium you're working within the etherium sort of standards framework tell me a little bit about ER c17 wait ER c 172 for or is it ERC 173 had me both of these it's better than I'm amazed I'm amazed you knew the numbers because I typed my

stuff had forgotten them it's a shame where ones were taking so LC 1723 is a standard for our cryptography engine and the RC 1724 is a confidential token standard that uses the top of the engine standard and so yeah when I came to these standards well when I was putting together a second November and December I was realizing well hey I'm kind of happy we're going to having to answer a lot of questions that have not really been asked before which is like things like how do you actually do confidential trades of different assets again you also like to resolve neatly just every along a lot of sectors we focused on unilateral money transfer so bilateral track management stress how do i if I

want to buy the proof so we haven't fully implemented yet but adequate for example for partial border fitting so you have an automaker and you have a set of order takers that don't entirely film you want to but mostly do like and there's Iran's proof that guarantees that the either traits are fair and everything has been executed correctly privately but I can as I get into get into a standard how do I carry creates a systematic way of introducing these kind of modular composable journals proofs but in into a system that developers can use integrate with and can understand and so ELC 1723 is our solution to that is up cryptography engine effectively what it does is it abstracts away proof verification and takes it out

of the hands of the individual like smart contract developer builder so that they don't have to deal with it I mean there's a lot of other reasons why we do that I'll get to them in a bit but basically it's a suite of smart contracts that live on chain they act as a kind of ultimate arbiter of the correctness of a zero and specifically our kiss the pins asked a zero space but the standard is much more generic than that see this is really useful for a lot of reasons the first one obviously is that you can completely abstract away the preferred engagement then the digital assets more contract to the point where for example we categorize crucify identifies and different we have that meant five proofs they all perform there's a developer when it comes to

verifying this proofs digital asset level you just saying hey I have approved for the idea of you know like seven rather bigger sprees and so you can have a very simple interface that accommodates a large and ever-increasing number of those principles like different things without it being inconsistent with kind of what the additional asset is trying to achieve so are you guys are you doing that with other people is that like a is it just the UH naztech proposal or cut maybe there's some other teams working on ZK stuff on aetherium so yeah so there are the teams work as EQ stuff but it's it's very early days most of the other teams they aren't dealing with premises are dealing with scaling solutions so for example very white has roll-up and the method labs team when it comes to

private transactions other than in like hackathon projects and the other thing you know I can't really think of my head of at Tina's specifically focusing approach I guess me with is Easter guys but I think they're more academic at this at this stage I mean we'd love more cooperation this year on tech level it's just that we've kind of been kind of screaming into the screaming to either of them there's not a lot of people to get meaningful feedback on the cryptography front from a decent community know that all there is names but ease there it's just that everyone's focusing other and stuff and right now privacy is not the focus of a lot of people is going yeah there's also been a wave of privacy like private transactions like parody has built in private transactions but it's off chain private transactions and that was the

path that a lot of people who are going for a long while is it's not use their knowledge proofs at all but you know I know my endpoint like I know the pair that I'm talking to and I might communicate on chain of like how do I connect to you but then do your part of transactions off chain and you just commit an encrypted state on chain so you have private transactions but in a completely different way yeah though it doesn't really relate to what you're doing yeah exactly and I think it's probably different different differences to many different kinds of solutions as it was very much focusing on kind of the the kind of the have your cake in you said Mosel where you you want to enjoy the Consensus benefits of public okay I'm

curious like when you or Gianni's RC processes or like EEOC standards and generally in in working your products how much do you think about the kora level protocol like it indeed I mean at some point you must have gone if only this thing existed this other thing would have been so much easier and I wouldn't have to dealt and deal with all this stuff like what on the core level would do you wish was different so at the core protocol level of aetherium yeah um gee there's this yes I mean you ask any theory much developer part the devil theorems I have a mess a laundry list of things they would like to change and we're no

exception to that but I do want to kind of preface this with what I think if the room is being accomplished as a she is quite significance in that because it allows you the fundamental goal of the theorem really was this graters you know they realize they didn't understand not define the potential use cases that people wanted a blockchain form and so by creating the spoken language they allow other people to say how they can add value and that they were trying really trying to kind of follow in this underlying philosophy by giving people the tools that they need to build their own similar fascism doing for them in theory was not really designed to values in order cryptographic protocols it was designed to do relatively simple things like transfer around the RC twenty second business so the it's

virtual machine is quite inefficient when it comes to actually validating like before we the mathematics required to develop it is your own proof and because it's inefficient and you can only press us so many transactions a second it means that the the concrete cryptocurrency cost the gas cost of the urgency lithography is extremely high this has been approved recently with the addition of these countries pre-compile smarts contracts there are exceptions luthier brush machines even before elliptic curve cryptography on very specific curves but it's still extremely expensive because the implementations originally they weren't particularly efficient because they're focusing more on the security side of things in the wind speed but basically the the takeaway from this is the amount of computations that's Ana theorem is

performative other somatic zeros proof it's roughly ten times more than is actually required just because of the inefficiencies of the underlie the smoke washer platform and the prequel a spark of loss and so I mean I guess two things one is um there's there's a movement of the works to add some more great compiles to theory but which anyway before we lit occur arithmetic over a much wider range ability curves which would be extremely useful for our state as well as producing gas costs because reducing competition which is also used faster we're also very interested in the point of entering work with X proposals to to improve the decent virtual machine and make it a bit faster um run programs and perform mathematical operations because that would help us we

actually have a smart contract that performs majority of mirror if they're tickled quiet to village a house in all scripts so we didn't need the prequel pals in fact the small project is cheaper than the pre-compose even though it's written as a small contractor so actually improving the interpretation there but also enable us to be more cool and so why would you want the pre compiles them like more options for precompiled if you have a smart contract that's actually faster I mean we'd like more options that a faster as well um let go yes also things like the our smart contracts everyone's arithmetic over the BN once you age I don't that's cool beyond 1 to 8 B and T 5 16 2004 Danes but Z caches old curve it performs arithmetic on that curve

Lina very much like axes it was kind of ironic that that curve was added for Z cache interoperability and then Z cache changes oh yeah I heard about that yeah it's ironic isn't it because i zk shot using that curve but now we're using z caches care to do many different kind of check 0 instrument ii and all of our people in the same boat even there's either paper that does brother crease the whole want the whole thing what purposes they don't mean for us to settle and so they can use more secure and they don't require by than a pairing so they can use more secure curves but they have to use this beyond 1/2 a cup of a theory if they wanted because of the prequel class what was that paper you just mentioned is ether ether okay yes i have a question about like you

mentioned it for a second you mentioned as techn users and i still i'm curious like who are the aztec users are they deaf developers who uses Aztec so the asset users are not retail consumers they are debt developers and companies building debts and smart contract systems on ethereum so how do you like how do you actually see like we follow a lot of the zero knowledge stuff going on a lot of the zero knowledge stuff on aetherium as well and so what how do you see that developing if I do my job if I if I fasiq the phils it's what its mandate does what it sets out to do then by count answer that question because the people who will be answering my question

will be the people using a second using other Zillah technology because the fundamental I think the thermal challenge this is your industry right now is access instead of having this kind of the high priests of cryptography deciding what the problems are and they providing solutions for those problems right now that's kind of how it has to happen because it's so complicated so abstractions are difficult but we really wants to move towards a world where we're giving the tools nfinity users and developers they can solve their own problems using the knowledge Tech and we don't have to decide what they are and so basically I think that what between scene if he is is an enormous emotion large set of people using journalist technologies in those tools ultimately it looks a subset of those users will

even know that they're using general technologies because it'll be abstracted away but they'll be so have the kind of the ability to use this Texas to actually define their own like solve their own problems as and so how do I see it developing I wants to see like not entirely sure how they're going to use it to be honest and if people using is you're always checking is in exactly the way that I think that I prescribe and how I did how I see the world then I found because I've not provided the access that I seek to provide because I'm not be able to allow people to solve their own problems so you want to basically spur creativity and other

people and figure out how they do it cool yeah I have a last question and this is about the name Aztec yes where did you come up with that to be honest it wasn't really intended to be the name of my privacy technology wasn't they were my company but we needed a name for this thing basically add two zeros check it even the name so I wish I killed zap zero knowledge anonymous was it pre for critical but then I found this app was actually the company operating one theorem so I couldn't call is out and so I thought well even they need an acronym you know so I come from particle physics get experimental particle physics and if there's one thing this is a great for

it's it's coming up with obscure weird acronyms and so a sixth sense for anonymous zero knowledge transactions with a fishy communication some of this is a silent and so I can't because I we needed a name it it yeah it kind of throws a fork of the tongue in nurse the team liked it and so we we decided to young is a Socratic owns we roll with it it's pretty cool I think it's good so to wrap up where can people find out more about the protocol the research papers that are interesting like how do I learn about this stuff the majority of our activity is on mono repo a sec github repository so github lock all

four slash Aztec critical we also have a medium publication which describes in high-level terms what unless they're doing why we're doing it and this is really the two main go-to sources our papers on and I want to read them which describes the fundament like the fundamental assic tech we are going to be adding a lot more papers to that we can see we have we have a few papers not kind in the final final stages that we you know we're gonna throw out there for example about just a server but that's why we that's where we'll the activities you mentioned that you'd taken this paper to Yen scrath do you peer review papers in general or is that like do you just publish where you have so the only

paper was public at the moment is the original critical paper which was created with extensive feedback from Ian's growth and so yeah I mean we wouldn't wouldn't we were betting probably never had me the amount being interviewed by by and I could emigrate over from the show cool well thank you very much for being on the show thank you very much it's been an absolute pleasure thanks for inviting me for sure and to our listeners thanks for listening thanks for listening