ZKPs for Privacy and Scalability | ETHSofia

[Music] hello everyone um welcome to the panel on zero knowledge Bruce where we are about to explain everything you need to know without actually telling you anything all right um I'm going to present a list of awesome speakers starting with sir ingu Ru um from the found of C protocol an identity protocol that provides secure and practical infrastructure for Enterprises governments and consumers helloo all right next one Marco Marinelli from Horizon Labs Marco contributes to ZK verify a modular zero knowledge proof verification chain aimed at democratizing the adoption of ZK proofs he's also the one with the longest hair on the panel all right next one is Martin ianov from BL sense Martin is a software performance and metaprogramming Enthusiast his interest in crypto cryptography brought him in the space in the web tree space only a year ago and in that short time he gained experience in developing trustless oracles leveraging the highly scalable recursive succinct proofs I mean I've been in this space for seven years and I don't understand a word of this big brain all right the panel is going to be moderated by Rosen stanf Sant sorry Rosen is the co-founder and CEO of aedify a pioneering platform in digital asset back Landing welcome to the stage everyone and I leave the word to you okay you hear me well I think yes guys I want to tell you something first I'm so happy to be here as you can understand I'm happy that Bulgaria is hosting such a great uh conference ethereum Sofia pretty nice and yes I have to share one story with you actually a few days ago I was talking with my mother my parents actually and I was telling them that I'm going to be part of this conference we're going to be talking about web tree privacy identity uh zero knowledge smart contracts all this stuff she was looking at me like I don't know what you're talking about but I'm proud of you and the case is that these people here they are experts in zero knowledge and I can tell you that even an experienced blockchain developer sometimes feel like my mother he knows about zero knowledge but can understand anything and I really hope that these guys here are going to help us to understand this technology better and are going to give us advice on about what is going to happen in the next years with this technology okay guys so to warm up our audience a bit let's talk first about what has changed in zero knowledge in the last two or three years Marco maybe you can start here okay amazing so uh thank you for the question and uh I think that zero knowledge itself has changed a lot in the past years and the reason for that is like if you think about zero knowledge at its Inception and at the end of the 8S it was theorized as a concept but he didn't uh he didn't found an application immediately because of different constraints like all the computational power required to implement the system and also for verification and in the last year with the Advent of blockchain ZK and blockchain came together kind of as a complementary technology and uh I think that in the last years we have seen like a Cambrian explo explosion in the field of zero knowledge seeing more proof schemes developed the academic being interested in researching new technologies and improving on the existing ones and uh yeah I think that uh a lot of real world implementation are coming out straight of that research and it's very cool to see something being theorized and implemented just uh a few months later okay okay so as we can see it is now uh the technology is improved it is more secure more like finished let's say but is it easier for the people to use it like for people to understand what is it about how it works and is it easier for them to apply it in their products to resolve problems that this technology can resolve so I I will say in terms of is the technology ready to be used the answer is yes and no the point is that all the building blocks of the technology are in front of us the art part is to choose the best ones and put them together to have the most efficient system we can dream of basically Martin maybe you can help here as well like you've started working on these type of Technologies one year ago and how do you view it was it easy for you to find information to yes your knowledge about the topic was it easy for you to to start working on it and be good part of blockchain development team uh so I've been thrilled with the resources about blockchain uh they're very good uh but the zero knowledge uh information is not really out there is very hard to find uh it's also it's not hard to use per se um the underlying technology is getting harder and harder to grasp at Deep level but if you imagine it at a black box uh it's not really that hard to understand how it applies and how it can be applied interesting okay I hope in the next years we'll be able to see Improvement in in this specific topic because it is very important if there are no people that understand and can work with it it will be a problem to adapt it as a technology and use it because as I believe and maybe everybody here believes it is useful it can be useful uh about if the metric is easy to use it's definitely getting easier to use uh if you're uh metering how easy it is to understand uh it's it's harder to understand not easier uh as the time uh passes so to conclude it is easier to use it but harder to understand it yes and do you think this can be a problem like using something that you don't understand uh well there will always always be people who understand it at some deeper level and uh you can uh put your trust in them uh I'm not that kind of person I like to uh dig deeper and uh understand more and not uh trust anyone and take their word for it uh so it depends on the audience for me uh it's not a good thing that it's hard to understand I believe that's why you've made so much just for one year because you're interested in understanding the topic but uh yeah I believe most people in blockchain uh they all want to understand what they're using because the decentralization is what is important for us our privacy and that's about not being controlled for somebody else what do you think guys about this so um when I think of zero knowledge I think it is mostly uh a topic where probably a lot of PhD thesises of mathematicians have to be written still uh it is it is a highly Comm so I'm a computer scientist and I'm not a mathematician uh my wife is a mathematician and she somehow tends to understand what it actually so how it actually works we all know what it is it is actually pretty simple so I know something and I want to share with you that I know it and or that I have it without actually explaining you exactly that I have it just proving that I have it so that that is zero knowledge proof right this is what we're talking about and this you can solve with different mathematical uh theorems around uh and some are probably more complicated than others some some take more computation power than others some can work on a mobile phone some need a supercomputer basically um to actually prove things and this is useful in many many ways so um what so as you ask so who who wants to be in this field if you want to be in this field I think the best way to go in this field if you are a mathematician and you want to really change something in the world and you have your name written somewhere really big go on Research in zero knowledge proofs because zero knowledge proof is something where we have not seen the end right it is still under construction it has a lot of difficulties for me uh the most the biggest difficulty is that if I take such a product there are products out there there are algorithms that you can take and put into your product so we're building a blockchain so if we um if if we put this in and I don't fully understand how it works it does not feel me make me feel good so I think what what I would love to see at one point is a zero knowledge algorithm which is explainable so that people like me computer scientists uh can actually say yeah this is I understand why that works that that would be really cool yeah I think I can agree with you like building uh some level of obstruction that is totally enough to trust what you're using and also having many different type of implementations and many people included in such process in building these obstructions so we can all trust and use it maybe this is this will make it times easier and better for us okay uh then I want to put on the heing today so I want to make the difference in privacy identity I mean privacy and identity versus the scalability and where the zero knowledge can be applied better or it can be used in both worlds pretty fine Marco yeah you are our product guy so you can start with this maybe maybe I can take this one so uh zero knowledge initially were conceived as a privacy preserving technology we know that and uh in the last year especially in blockchain we saw that the along with the Privacy preserving feature of Z there is also the sness feature of Z proof and what I mean by Suess is that the proof is very small and easy to verify and in this way we can take a load of data compress it and then just verify the proof associated with the data and so this brings a lot of improvements in scalability because like at this point we are speaking about verifiable computations so you can prove the execution of a lot of data not re executing all the transactions all the computations but just verifying their execution and uh this is quite powerful and like in the blockchain space we are seeing that especially with ZK rollups you have a layer Sho that batches a lot of transaction puts it into uh generates a proof on the on the execution of such a transaction and then sends the proofs back to ethereum or any layer one blockchain for the verification on chain so yes I will say that in terms of scalability like we're seeing a rising use of of this te okay maybe for some of the technical guys here it will be interesting how actually you uh make it such way that your layer with the zero knowledge proof where the zero knowledge proof is submitted is secured enough and people can trust it how do you make it in ZK verifier so thank you for the question because it's um like what we do with our product Z verifi so actually the generation part of the proof is quite expensive computationally speaking and this can be done off chain for sure and also the verification can be done off chain but at this point you have to trust a server that cannot tamper with the verification result but can send back a wrong result a Mis leading result and so what we're doing with secy is basically having a simple modular blockchain designed to provide consensus and verify proofs in a decentralized way so we receive proofs from applications that use lzk technology we verify them and then we send the verification result back to ethereum or any other blockchain where the application leads so in this way they can go they can ask about the verification result in a smart contract and get back the result and uh you don't have any trust assumption basically apart from the one of having your proof verified by a decentralized set of validators which is like a battle tested approach in blockchain so the intermediate layer it uses its own consensus mechanism uh as a blockchain developer I can understand uh some concerns about that people here may have that uh uh it is not easy to trust uh specific blockchain and its own consensus mechanism without using uh the security of ethereum for example do you have some such component there like uh using the security and the consensus of ethereum one or your intermediate layer being something like Layer Two or it is totally separate blockchain well it's a totally separate blockchain and actually is based on substrate the same framework that padot uses and this is done for mely uh avoid all the technical complexities and the gas cost of verifying proofs on DVM so DVM doesn't have like any precompile and verifying a proof costs a lot of gas on DVM like for the more simple one it's uh considering 30 gas and ethereum at $3,000 around $6 $20 for each proof so like how can we hope for a wide adoption of proofs if uh their verification on chain cost that much okay uh we we actually have to continue with I want to ask about the identity part but one was question about this topic do you think that ethereum is doing the right actions to make it uh submitting such data like uh times less costly so we are going to be able to use it as the layer where we submit such data without being afraid of the high gas costs well I would say that uh we can make a comparison to what layer 2 did for etherium so initially there was the whole sharding idea and so on but then we realized that it was way easier to delegate the execution to other chains in a modular fashion so you have execution on the layer two and then the settlement on ethereum and I think that the comparison can pretty well fit with also the Z proof verification so using ethereum as only the settlement layer but delegating execution to another service can you please turn there as well uh because uh we we do quite a very similar thing actually because we also using substrate right so so we also build on substrate uh as as as a par chain to parka dot U and I think this this is what what you see a little bit as a trend if you are doing specialized things in blockchain you don't really only go on ethereum or something you have those side chains whatever you want to call them per chains which do uh specific work and have specific consant algorithms and are extremely cheap to operate for example um and you build them on substrate because in substrate you have far more flexibility because you're building your own blockchain so you build the runtime of the blockchain yourself instead of just accessing a blockchain through a smart contract interface so if you're building complicated things you see that also in in the cardano world they're building their partner Network also on substrate so this is what people do these days actually uh to not build anything everything on Smart contract but to actually broaden their uh their possibilities of building stuff uh with using technology which allows them to build their own blockchain yeah I totally agree with you that we need D specific chains and uh they will be part of our future but my biggest uh like important thing in this situation is that uh we need uh decentralization and security and it is just not so easy for the other chains accepting Bitcoin and ethereum if we can accept yeah I believe we can but if we ex exclude them uh the other blockchains are struggling sometimes to ensure decentralization and security as it must be so yeah that is my concern yeah but this is also so in subst you can do that too you don't let's continue with the identity in my opinion Ino is our specialist in identity so he can share us what do you think about that how your knowledge is related to Identity and how it can help yeah so I think that so we've seen quite a lot of identity projects using ZK in the last two or three years but they that may also be connected to the thing that ZK was like two or three years was very much a hype in the Venture Capital world and it was really good if you said ZK for for for getting some money um the question is do you need ZK in identity and if you try to explain it very simple then of course you understand of course you need it because uh digital identity is uh based on a decentralized identifier on one side and on so-called verifiable credentials on the other side verifiable credentials are things that you store and then you share with someone so for example a passport or an identity card is a typical verifiable credential in the physical world just we now it in the digital world but in the physical world we already had those things so uh ID card uh your University degree those things are verifiable credentials and when you look at your identity card you see that there's a picture of you that is nice because that connects to the identifier but then there's lots of information about you there's where you live there's how old are you you are uh what your name is and all that stuff and if you want to use this uh identity card for a special purpose for example getting into a bar where you need to be 21 um you show your identity card now the guy who looks at it knows your name knows your address uh that's not necessary the only thing that he has to see actually is the picture because he has to compare it to your face and the date of birth so that you can make a quick calculation if you're 21 or not right that's the two things um in the physical world you just put your thumb on this part where your name is that is what we call Selective disclosure so you only dis disclose what you really need to disclose like the date of birth um in the digital world one could go and say this is where we need zero knowledge proofs because we want to prove that I'm born on this day but I don't want to share the whole credential right so let's make ZK stuff on top of it um possible but to be honest much too complicated because you can also do that with merization in merization you can just have a um a Merkle tree around the uh the whole credential and you share uh an end result and then you can see and then you can just share parts of the credential and you can see still see if the end result is is correct so that that is much faster it can be calculated in micros seconds um and zero knowledge proof makes it more complicated and as I said before also it zero knowledge proofs tend to put a certain opaqueness on how things work while a Merkle tree everyone can understand and that makes it much easier for adoption because um when we talk about identity then talk about big companies which use that we talk about governments which use that and so on and they tend to uh want from us not only that it's open source but they also want to look in the open source and they want to understand the open source because it's really critical infrastructure for them and if they don't understand what's happening in the back uh then they will just not use it and this is why we at Kil protocol we use the simple efficient method of merization because that is easy to understand um and it's much faster so we opted against um using zero knowledge proof to solve this problem of course there's tons of other problems which you want to use with uh zero knowledge proof but in digital identity there's easier Solutions uh it was a pretty nice description of the situation but Martin as a young developer uh can you tell me what is your opinion is it really easier to understand merco trees than zero knowledge Concepts uh Marco 3 is are a concept that anyone can understand it's uh really very simple uh as opposed to zero knowledge proves that uh even if you have a m degree it will be tedious to understand so not this agreement on that side okay we have one agreement here okay and Ino then do we have something at least one that we can't achieve without using zero knowledge in the identity space can you give me such an example um well I directly in the identity space probably not because the identity space is well defined in the meantime so we have international standardizations we have dids and verifiable credentials signed off by the w3c um so they are now on the same standardization level like HTTP and for everything that we have in there we don't need zero knowledge proofs that doesn't mean that it's not allowed to implement Zer knowledge proof so you can what we do with merization you can also do with Zer knowledge proofs if you want it doesn't doesn't say in the standard US Zer knowledge proof is not allowed that that's not the case it's just that the customers don't want it um but identity is always connected to many other things so identity is probably something that's going to work cross chain and on Cross chain things uh especially when it comes to layer twos and so on we have seen really nice um implementations of zero knowledge proofs to actually uh put together data and uh and and have it in a very efficient and uh storage uh um low storage way on other chains and stuff like that so this is where zero knowledge proof really comes in useful I think it is a layer below the rather user Centric uh application type uh layer which we are working on I think when you really want to do things like trustless bridging and and and stuff like that then it really can become really interesting to uh to to use these things but I I don't see it really on the on the layer where I work thank you thank you very much Marco yeah I want to hear your opinion on that because as a product guy you have experience with uh zero knowledge yeah sure it's business point of view so please share with us no I wanted to ask something that first of all I didn't know you use Mar ation and actually is a very smart idea so I I agree on that but also at the same time like for the wider use cases like if you use internet you don't have to understand tcpip or HTTP so you can abstract away that and then and the user just goes on the internet and do what they they want to do and I think that the Final Destination the final goal will be to uh not having the user not understanding it but abstract away the system so that the user can understand the benefits but it's not important if it doesn't understand the underlying technology yeah Absolut absolutely correct so the user hello so the user is not interested if it's Z knowledge proofs under the hood or not uh the customer who wants to implement that in their country they are interested so I I can give you an example we were working with the uh DHS is the Department of Homeland Security from the US last year uh to Define how the you probably all know the esta thing right so if you want to travel to the US you have to fill out a form in a central centralized database two days before you go and then you get an okay back and then you can travel to the US uh that's called esta for whatever reason uh they wanted to put this on dids and verifiable credentials and they asked us uh for help and not only us they asked a lot of people for help and uh then uh we had this presentation where the DHS presented what they actually wanted uh to do and they said we need the system it has to be simple if your solution includes zero knowledge proofs please do not apply um because we are not going to understand what you do there and this is Department of Homeland Security so we're talking about the National Security of the country and if there is a piece of software which we don't understand and where nobody really takes the responsibility for how it works if it really works in every Edge case or if there's maybe a back door that we just didn't see then we cannot use that so zero knowledge proof out and so this is not the this is not the end user of course the end user when I apply for an esta I don't know if there's a centralized data base a blockchain or a zero knowledge proof behind that and I don't care I just want to travel to the US but the US is interested in the question well thank you for your opinion uh you were talking about about the countries and how they're going to apply the digital identity standard that you were talking about can you tell me more about it like uh what are the governments doing to standardize such thing because I have seen many projects that are for example doing uh some type of modern kyc which is some type of decentralized and it is super cool and nice and better but they are like their clients are not actually allowed to use this K provider because it is not standardized and this is a problem uh the the issue that that governments have is that they in in the in the analog world in the old world there were big players in identity because they were issuing credentials they were issuing passports they were issuing uh identity cards that were issuing driver's licenses big in the business and they were really a big trusted entity and then digitization comes and when you think a little bit where your digital identity actually lives the outcome is probably yeah on the Google server and on a Facebook server um that's unfortunate right so your digital identity lives there and if you want to lock into a new service you click connect with Facebook and then you're connected fantastic um so this is something that governments really don't like because they losing the touch in the digital world with their own citizens and this is what they don't want so all the governments in the world are currently very much interested in getting ahead of that and getting ahead of such a monopol or do duopoly or whatever you call it shelling point we heard before actually for identity Google and Facebook are shelling points I think um to get over a shelling point is not easy so you cannot just go and say I make it better I make it in a decentralized way this is not going to work what you need to do is that you have to agree on how it looks in a decentralized way and this is what happened in 2016 when we were founding the decentralized identity Foundation uh which has more than 350 member companies and governments in there the American government is there IBM is there Microsoft is there Google and Facebook are not part of this thing and they said we have to standardize in very simple terms like a protocol like TCP IP how decentralized identity actually looks and when we have this standardization and when this standard ation is signed off by the w3c then all the countries can basically invest in building those systems and they have decentralized identity as they had before in with the plastic cards and that stuff which is decentralized identity already so we just replicating the digital the the physical world into the digital world and we're making Google and Facebook useless by that but the uh the precondition for being successful there is that you have the stand and now since 2023 we actually have the standard it is signed off by the w3c it is on the same level as HTTP and now people start implementing it countries start Implement implementing it I already said USA their plan is to issue the uh identity cards on uh on dids uh next year that's probably not going to happen but wait two years uh we currently are in talks with the Republic of Indonesia um then there we're talking with some Arab Arab countries we see uh that uh basically all Asian countries are moving in that direction so it is actually pretty cool that everyone agrees on that standard because that makes compatibility right I can take then my American passport and actually travel to Indonesia and get new credentials there and add them to my credential list and have them in my digital wallet that is really really powerful but it needed the standardization which unfortunately took forever totally agree okay one last question very shortly please about the identity how how close we are from that from using this type of digital identity in our day today lives say what was the question how close we are to using this type of technology in our day-to-day lives like when we will receive it in one two 3 years when we will start using it yeah it depends a little bit on where you live and uh so I think if you live in Asia uh probably you're going to use this Tech technology or sometimes maybe you even use it already without knowing it it's another technology which is not really U transparent that you use it right so it's not saying hey did now it's just a digital passport that you have in a digital wallet um uh so I think in some Asian countries you already use it uh some other big countries are going to follow in the next one or two years uh us is going to follow they say in 2025 but I suspect 2026 uh Europe is a little bit slower so Europe maybe 2030 okay thank you very much uh I want to ask something Martin because he's our new uh like zero knowledge developer here uh would you like to tell me what is the process of being part of block SS starting with these Technologies what is the way for new developers uh that can follow your example and start working in this area could you give them some advices so the microphone please uh it's a different process uh you're still writing code uh if you're developing a zero knowledge circuit for example which is still the most prominent uh way of uh generating zero knowledge proofs uh IDE code it's not linear how you like how you're used to you're uh building a c arithmetic circuit uh connecting all the wires and uh you just need to understand that everything is being evaluated and uh all at once uh which uh can can be hard at the beginning uh there are not very many resources uh to learn about how to how to write those um the Whiteboard uh ZK sessions have been very useful you can try to Google those also the zero knowledge podcast is what I would recommend for starting out thank you very much Marco yes I'm listening to you yeah so I wanted to add something on top of that and uh basically what we're seeing right now so the old school approach was to write circuits for a specific proof scheme and that is uh very much optimized but very much time consuming like writing a circuits in the past years could require up to months of work and uh necessitates a very very deep knowledge of how the entire system works so the expertise out there is not very much but what we're seeing in the last uh in the last year is the rise of Z kvms that basically are virtual machine that allows you to write code in C++ rust and so on and then they take care of compiling that high level code to Circuit code which is not super optimized not as much as circuits but at the same time provides a entry point to every web to developer basically with the knowledge of this programming languages this is super nice and I believe this is the way to adapt this uh technology and actually use it if we need it and want it uh I also wanted to few tools like the best tools in your opinion that are right now used uh in the zero knowledge field just to help again people start using and looking for it uh I want to shout out a project there is a a language a programming language called Noir uh it's a domain specific language for uh writing zero knowledge circuits uh which is uh very much like rust uh in some sense in the syntax uh but it has some constraints uh this is uh some hybrid approach to uh writing uh circuits uh because on the other uh side you have something like ccom which is uh very hard to use uh and uh on the easy to use site you have the kvms which as Marco said are not very optimized uh in the middle you have n which uh still generates a circuit uh in the intermediate representation of the proving system that uh you're using uh but uh it's very much like uh the regular code you would write okay thank you very much as uh because we have to finish our discussion it is very interesting for me I hope it is interesting for you as well but one last question and please one or two sentence uh for it what are your predictions for the future of Zill knowledge like your biggest dreams maybe uh Marco can I go yeah okay uh so what I would like to say and what uh I hope I will see is uh zero knowledge finding real world use cases outside of blockchain because in blockchain yes we're exploring zik proofs for scalability for privacy and so on but what about bringing these uh features to web two projects web two applications that uh use ZK for these properties but then uh settle the verification result on a blockchain so like you have privacy you have scalability but also you have censorship resistant by publishing the verification result on a blockchain and I think this is a very uh overlooked uh space right now Ino and just adding that because it wasn't said um I think uh it will have a big future in preventing civil attacks here and there okay and Martin it's very early for me to say where the future is so I'll leave with leave you their opinion on that uh yeah okay so if we can conclude with something the zero knowledge is not our decision for every problem we have but it literally can help us achieve better and uh like yes better world so yeah this is it thank you [Applause] [Music] n [Music]