ZK8: Park Bench Discussion

Transcript

foreign of the ZK Summit 8 we are going to do the park bench panel this is always an experiment it's sometimes a little messy and as I said it's really important that you prep your questions without questions the thing doesn't work so I need you to have some like hot takes and good questions so I think we you know we wanted we often use this time to also talk about like somewhat controversial things and so I think our first question that we wanted to talk about was what is ZK evm and what isn't zkuvm

oh intro round oh good idea okay let's start right there um I'm Zaki uh I work on a variety of blockchain projects uh I have something called sommelier Finance I've got I've been working on a thing called Cosmos for a long time um some people have heard of it um and the uh the that's uh that's what I'm here for I don't know I've been like a ZK Fanboy for a very long time and I that's how I think I ended up on this panel um I'm Henry uh I work on penumbra which you may recall from the previous talk um so yeah we're building a private decks I guess and yeah I'm Alex I work on a Leo also a long time ZK and ZK podcast Fanboy which

is probably also how I ended up on this panel okay I'll ask the question again what is ZK evm and what isn't zkevm um so you know there's like I I will frame this thing like you know so there's like this uh like uh four kinds of ZK evm blog posts that vitalik wrote which is like an attempt to systematize this uh this uh uh actually like very difficult question right because um you know it it basically like anytime anything happens in zkebm and like the last big zke evm thing was uh uh Paris you know polygon comes out with her like we've made a Zeke ABM and everyone else's he gave him is not a ZK AVM and then everyone else with the ZK VM says

they're like all their other kind of parties are also Nazi abms so this is seem to be like a a little bit of like a kind of a a Mean Girls kind of thing going on where like you're either in or you're out and the in-group is being the one true ZK evm um I think like one of the biggest challenges of this is that um zkevms are like right on the cusp of being turned from being like research projects into something that like you can tangibly uh uh uh sort of understand and to end and build upon and it's going to be probably like the ultimate question I think is like you can tell us you will know a zkevm

um just like so this reminds me a lot about like when people were building like the first proof of stake EVMS um and you know it was like there was like are you by code equivalent are you RPC equivalent like all of these things was like this huge conceptual framework and you know what actually mattered what actually mattered was like if the developer Community could like adopted your thing and decided that like this was like good enough for them um and the so like ebm equivalence ended up being like very much this good enough question in the bft proof of stake setting I think the same thing will apply in the ZK evm setting I I have no skin in the game so like I I don't know to me Ezekiel ZK VM is like

something that somebody makes and they like write a blog post about like here's our ZK AVM and explain what it does and like you know then people can argue about it I don't know hot takes Also if anyone makes a meme of Mean Girls with all like the names of the projects above the characters bonus points um yeah so I guess you know I don't have much to add I think saki gave a great answer um I made my quick take I think is that in general is it kind of the lower down you try and Achieve evm equivalents it seems like there's a a performance trade-off or you know you have to kind of make compromises in certain things um that's I mean that was really the thrust of vitalik's post I think and I think that's kind of what we've observed with a different zkevm implementations

um and so I think it really comes down to the community of people around this that will you know end up adopting something and they'll really decide at what level does it matter if you're you know down onto the byte code compatible or if it just okay if it's fine to you know at a higher level you know if that compatibility is okay of course ethereum has huge Network effects right they have the entire ecosystem of dapps they have a whole ecosystem of solidity developers people who are familiar with that model and so I think that's that's why people try and really closely Hue to it but you know I think potentially there's a lot of really interesting constructions coming out that that are much more performance um you know different ZK VMS that you know kind of they kind of give you a big

performance boost which potentially will I think over time you know you'll see people building on those ecosystems and and you know and and we'll kind of get a I think probably a diversity of different uh VMS or ZK VMS that serve kind of specific use cases can I pivot the question a little bit sure okay so maybe to make this question more relevant to the people who are on the panel um Henry has this like sort of half-baked maybe VM idea and uh Elio has a more baked VM idea but neither of them are emulating anything with existing Network effects and you're like trying to restart your own network effects and we have had we have lots of people who are here whether it's the zika evm people or the risk

zero people who are trying to take advantages of the network effects of something that already has Network effects um or how do you think about that um yeah so I think I mean my opinion is uh for any new protocol or product you have to differentiate and offer something that is unique and new I think I hear a lot of time the you know you know because wait you know we're so Alia we're building our own language called Leo's a high-level language that you know compiles down to an IR Halo instructions it then gets you know rolled into r1cs so it's a new it's kind of a new ecosystem the developers have to learn to build in and I hear a lot of times like oh well who wants to learn a new language and I I think I actually

don't like that question because I think everyone in this room and most developers I've ever interacted with you know if they're if something adds value to what they're doing if they if they see something is legitimately interesting they will absolutely learn it I mean look at rust 10 you know I mean rust is a new language right so everybody in most of the projects today and this ecosystem certainly use it so people thought rust was interesting and useful enough and provided some value such that okay they're going to go ahead and use it and so I think it's really just a matter you know of course on Elio and every project that's starting from the cold start you know you have to figure out okay how do I want to be different I think that's really the starting point is how do I want to be different what's

the value that I can add maybe that you don't have to say it's better than ethereum or worse than Bitcoin or whatever right whatever the you don't really need to compare it in that way but I think it's it's what is fundamentally different what can you do in this world that you can't do anywhere else so that's how I think about it um so I guess maybe I'll say a bit about how we think about Network effects and then segue into sort of ZK VM generally um the the strategy that we have is basically you know try to build one useful application and um try to have sort of a very very narrow initial scope but build something that is just like genuinely better

because of the Privacy features that it has so we're not actually trying to build some kind of like programmability thing the the talk that I gave about like you know here's an idea of how you might do that is really just an idea that is a possible generalization of the concrete system that we're trying to build and so the overall strategy of penumbra is like okay how do you build like a DEX that's better as a DEX because it has these privacy features and then once you do that you're going to end up with a bunch of users who are using the product because they're getting some kind of value out of it and once you have that now you have some kind of audience that might be interested in interacting with some kind of future programmability system

um once it exists right um and so that I think is a different kind of um Network effect it's a much more direct like let's just build like one thing rather than let's try to build some kind of environment where other people can then build a thing I think that if you get that to work it's like obviously much more powerful but also it's a much harder lift as a project because now you know for all of these other people who are going to be writing applications you're just kind of like another project whereas the whether they do or don't do that you know for you as a project that becomes this kind of like life or death thing um the second point on kind of ZK VMS generally I think that

um the existence of of really any kind of ZK evm is just a really incredible like technical tour de force um as a matter of like snark engineering um however having to have compatible like having to have compatibility with evm byte code is not really just about the byte code it's about the state model and um my personal belief is that like the state model of ethereum was like designed to be a like transparent system and when you try to make that private it's not just a matter of like oh well we'll make all the state Trends we'll like snark

all the state transitions or something but you if you have a notion of privacy you have a fundamentally different data model because now you have different entities that are supposed to have access or not have access to different pieces of information and so it's not to me it feels like the the programming language or the byte code or whatever is actually this like you know second order like Downstream thing of the more fundamental thing of like what does it mean like what is the data model for like private computation on chain right and for example with with Alia there's uh you can effectively have sort of like every transaction be its own little like micro roll-up or something um for us we're trying to focus on like

how do you have interaction between like some like public computation and some like individual like private data and try to have some kind of differential privacy or something there but I don't think that it's like really clear what it actually means I mean in the sense of like you know I have my opinion but that's not the same as like a like a market validation right um and and I guess the concern that I have is that if if you start by saying like Okay we're going to take the path dependence of ethereum and then like make it CK you may end up like sort of paying an enormous amount of of cost to do all this like complicated proving but then it's like great well now we just

have this like centralized sequencer so you know what was that cool all right I want to see do we have another mic you have a mic right um are there any questions comments ideas oh there's one arm I went right up there all right I feel like someone who actually makes a ZK evm should like might want to weigh in on that you might want to weigh it can you make sure the mic is on sorry just give it a second what number is that there we go oh no no we're here yeah I um I regret I was not on the on the final one the question was asked but they can still be my answer up

here in a second so all right so like yeah yeah it's it's interesting that we're using in zkvm we're actually not using the ZK property properly right like it's not for zero knowledge it's standing for zipped with cryptography to use the succinctness to scale computation so you can think of it as you know just computational layer just like IEP protocol is for data and SSL protocol on top of that is for privacy similarly we can build protocols on top of the KVM that provide privacy my question is about privacy because you guys are working privacy mostly um uh and it's going to be a controversial question of course like a spicy one so um we see an attack on privacy happening recently with tornado cash being put on

our fake list um and it seems like a systematic approach and like it's it's been boiling for a while with uh elliptic channels it's like the soft sensory enough transactions by exchanges and now like straight on Attack on on the Privacy protocol uh how do you feel about the the path of development of this regulations what can you share about your ideas like how you see the space developing like is it going to affect all the Privacy protocols do you expect like compliance do you expect non-compliance like how how you you think about the situation in for the future okay we're gonna have you come up but first why don't you answer that question you're going to be taking Alex's spot Alex is going to switch for Alex keep it easy that's good we can

just roll okay there's enough hours um so I can I can take a first stab at that um so I think so first of all I think the the ofac regulation or that you know the action against tornado cash that you reference I think you know there's some news yesterday that I thought was encouraging personally where you know they laid out a path where okay if you had interacted with that contract you could you know potentially you know clear your name effectively by going through some process so you know I think the process that they proposed was a little bit burdensome and may not really be what I hope we end up at but I think at least now there's there's some understanding a more pragmatic understanding from the view of The Regulators now I think my personal

belief with uh existing um decentralized system today let's call them Bitcoin and ethereum since those are the two biggest is that that model is so much more radically transparent than the existing financial model of banks that I I really think that it's going to be difficult to if we imagine a world in which web 3 overtakes web 2 and finance 2 overtakes Finance one I think without more basically without stronger privacy guarantees than those existing systems offer it's just not going to have people adopt them I mean I think look I think we have to be honest with ourselves on a lot of the volume that happens on on ethereum you know people do defy and they do nfts and I think a

lot of people view that as fun and interesting and cool but very few people I think are getting their paychecks there very few people are doing what I would consider real economic activity and you know if I send Guillermo a us one usdc on ethereum not only do I know his entire bank account effectively but everyone who's ever interacted with him and their entire accounts and I think that model like to me that just seems to be a fundamental blocker for adoption now I think you raise an interesting point where is it do you need to bake everything in in one layer where you have privacy kind of guaranteed all the way up or do you can you separate it in layers I mean there's there's I think there's legitimate points on either side I mean obviously Elio's taking the approach from the bottoms up because we

think that gives you the strongest most unbreakable guarantees there but I think you know it's it's really interesting to see how this ecosystem evolves but overall I guess the last thing I'll say is I don't think any of us can escape the need to be compliant at some level um but I think what we can do is advocate for uh and the best or the best way we can advocate for what we're building is proving they're useful right that's why I really like what Henry said a minute ago was like he described privacy as a feature of a product that people actually use right I think this is this is something that I you know I don't think we as an industry focus on enough it's like hey let's really talk about like how does this matter to some consumer right what's the you what's the

consumer-facing product and then building back from there I mean one thing that I think is um just like interesting uh about this is like so like there there there's some regularity to transactions on ethereum that uh I control like sort of making you know paying invoices and stuff like that and they're like the attempts to hack me are directly correlated with like the regularity of those payments so basically like I feel like this is like kind of like the most practical thing that I've been seeing uh as like a exposure to to is like oh like the attacker sees that like typically we paid we may pay we like pay invoices

around this time of the month and then like spear fishing is just like pervasive like that that that time of the month um which has been an interesting which is like an interesting exposure like privacy exposure and I think a potentially a good articulation of like why that might matter um I think also Henry has like a really good rant about like how IBC is a good uh mitigation against these things which I like dear I'll set it up all right I mean that that wasn't what I was going to lead with but sure um set Spike so I think yeah like there there are some good comments on the kind of like regulatory side like explaining like the value of you know privacy or whatever to me personally as a builder the thing

that seems like the most useful thing to do is to build products where the Privacy is like clearly this kind of uh on like like core enabler of what that does right so for instance if you're trying to have um like Market making on chain and somebody can just like open up a block Explorer and see like who are the top 10 performing LPS and I'm going to build an index out of their strategies like effectively you cannot do like good Market making in that kind of environment and so if you build products that just like are better products because they have privacy in them I think it's a just a much better place to be both in terms of like having a a

larger user base and um a kind of like more um tangible near-term argument for why it's important um I guess yeah on saki is sort of like you know request for comment um I think the existence of IBC um makes the type of uh of censorship that was done um a lot harder because all of these different sort of uh chains just communicate with each other permissionlessly there's not some kind

of particular set of um you know this is like the address of whatever like you have to you'd have to like block entire chains or like entire Bridges right um and so I think there's like a lot more collateral damage um if if a sensor was trying to um you know put put a stop to like some particular thing um I also am not I'm not really sure that the the tornado cash ofac designation was dislike carefully calculated like next iterated step in this like big long game or whatever like to me it actually

seems like a kind of like ham-fisted like you know someone told us that we had to do something about this thing and this is something that we can do um you know like if you look at the actual law uh before that is supposed to to give that power it talks about like uh sanctioning specific persons so like I don't know did we accidentally like Grant legal personhood to a Smart contract whoops I don't know like that'd be a kind of funny implication but none of this really seems to be thought through like the fact that um there's like viewing keys where you can like see exactly what your account activity is and like demonstrate that to some third-party auditor did that matter

I don't know apparently not so like what's the point of trying to like build a system that has regulatory like capability built in if The Regulators don't really seem to care I thought it was interesting that like with this license thing that got announced they did not ask for the viewing key it seems like that would be like the logical step in like getting a license again to Henry's point I mean this might just be ignorance on the part of some of the bureaucrats who wrote the rule I mean I think there's often and have speaking as someone who's worked as part of a US government bureaucracy before I mean it's oftentimes you know people like to think of it as like the Illuminati but sometimes it's a little more like office space and so you know um you know so I think yeah these things

will evolve but I guess and I need to pass the mic to Alex the last thing I'll say is um you know look I'm I'm actually really optimistic about this I mean if you go back in the history of cryptography I mean I think we've made enormous progress I mean there was a time when the the fundamental cryptographic algorithms that we all rely upon were export control they were like and it was like illegal to take them around the world or like tell anyone about them right there was a and there was a time even more recently with crypto like there was legitimate talk of would a government ban Bitcoin or ban this stuff I mean that that's at this point just seems so remote as to be ridiculous you know and I think we're kind of on this next Frontier of this technology and you

know I think it's of course like early days it's too early to say which way it's going to go but I think there's a legitimate path for it to become accepted and legitimate and welcomed even by by institutions and governments so we'll see cool thanks Alex come on up Alex Alex tap in all right do a quick intro Alex G let us know who you are hello everyone um yes so uh I'm co-founder of matter Labs which is the creator of ZK sync the first CK evm the first hey all other ZK EVMS are not ckvms right not not just the first the only no no we we just we were the first to introduce

the term the first to implement the Prototype the first to open the test net and we will be likely hopefully the first to hit the main but the others are also gkvms of course all right I want to open the floor are there any other questions anyone who might want to weigh in on anything that's been talked about and I will wait we do need questions for this to work comments thoughts anything um considering that we now have a variety of trustless setups my personal preference recently has been

bulletproofs what's the continue justification for deploying new crypto systems using graph 16 and other trusted setups that's opinionated hot takes gets up although bulletproofs doesn't have a trusted setup he said trustless trustless setups oh got it yeah I I mean like the the setup is like one engineering access out of like you know a dozen or whatever right like personally um if I I don't this is just my opinion right but I

think that if you have a a situation like say like the sapling ceremony I actually really wouldn't use the term trusted setup for that I would say that's a decentralized setup um and I don't think that there's really a meaningful technical risk there compared to say like oh I wrote the circuit wrong I like didn't constrain something oops there was an inflation bug like whatever like those all seem like way bigger risks to me on the the security side than you know 299 out of like 300 participants were all collectively dishonest and like

there's like you know the one hold out that they like wrench attacked into colluding with the attack or something like it just and and then um on the other thing that's nice about growth 16 is that like it has like very good contents and if you look at like um I think in in Practical terms like who is the marginal user of some CIS like some privacy system that is like you know I would be using the Privacy system but it has this like setup like I think that much more important is like what is the actual performance of the

system how does it work like for penumbra for instance like it's actually pretty important that you'd be able to do the proving quickly because you want market makers to be able to uh respond to price information with reduced latency and you also don't want them to be on this like separate sort of transparent thing because the point of like the whole game plan for penumbra is like you use the market making to bootstrap the Privacy set right so then actually having performance is like kind of probably also like directly important for privacy so I guess like what I would say is like there's a lot of different um engineering axes and I think that like the specific kind of security

of having a setup or not is like not that high on the the list there's other reasons why you might want to like a universal system um but I think not for security in my opinion well the the the uh obvious answer like on the surface is that it's a little cheaper to verify curl 16 on ethereum uh like on on the system like ethereum where um it's quite expensive so like girl 16 verification costs around 300 000 gas a full start verification is currently at 5 million yes that gives you a lot of like 20x almost 20x that's 18x difference

that enables you for example to do more frequent checkpoints because you need to advertise you need to wait for enough transactions in this batch that you're running up if you're using it for a roll-up for example to justify making this checkpoint and verifying the proofs uh so it would be great if um uh we like if draw 16s are going to be used still in the future that they are made as universal as possible that they are just approaching the targeting certain proof system uh where the parameters the verification keys are passed externally as inputs so you can reuse it many times which gives an incentive so for many people to

participate in the ceremony and make it more secure but over time for for like in the in the long run I think that we'll absolutely have to uh to avoid doing trusted setups you cannot build the internet of value especially with privacy that no one can actually see inside and check that it's still um keeping Integrity uh on something which like depends on on 300 people like no matter how how legit and prominent those people are like to to have institutions move to crypto and move all their transactions there it must be completely transparent and that's where we're coming for to in a few years well it seems fine like I you just like

make make the make the setup bigger like I don't know just I I think that there's not like one of the things that's cool about God 16 is that because you actually have a hard-coded statement like the wiring is free so that's kind of like a fundamental thing that is cool uh about that from a from a performance standpoint I don't think that goes away like I think that the solution is just that we like get good at doing setups which is actually what's happened right like comparing the the the the trajectory of like the first like zcash setup to the sampling ceremony to like all of the the ones that are currently ongoing it's

like okay this is like a thing that it's like a collective coordination problem that people who build the coordination technology like get good at coordinating oh it's inside I I don't know like every ceremony is requires a lot of attention a lot of diligence in like how you actually prepare your machines for that you might have 300 honest participants but all of them for convenience just use some software that they download from this link that the participants uh the organizers gave them or they they participate from a browser uh app where they do everything in browser which is like you know like you you you want a lot of different eyes a lot of different

implementations um I I agree but I don't think that that's actually a problem with the setup I think that that's just a fundamental problem with making a circuit like I if like what you also have to get the circuit right you have to like have Assurance on that what happens if somebody like puts some kind of like subtle bug door in the constraints that allows that is transparent that is auditable you you know it like you you like there is a chance that people discover here you can do it like theoretically fully covertly and like not letting anyone anyone know and this is especially problematic in the context of private systems like with open transparent systems we will at least

notice that you know if something happened in the key sync with Blanc uh with ignition trusses set up it would be obvious that the setup is compromised but with the private system it would not be obvious there can be things that go wrong and no one knows right like some hidden um minting of assets Etc so I I think for for like why use of privacy we really want transparent systems in the future but like that that's not an urgent problem I think we're gonna we're gonna move on if that's okay Henry I think you're off and I think you're on thanks for being on here Henry and I have to be on a Twitter space is

it six so you got it you got to get me off this you're gonna get you have to get off soon you have uh 21 minutes I have 21 minutes okay do you want to introduce yourself sure hi I'm Luke uh also known as Kaiba online I volunteer around Monero and then I have my own project as well nice all right are there any questions comments thoughts anything on what was just set up Alex there's someone right there oh sorry um could someone describe what is difference between schemas based on inner product arguments versus like these existing Solutions I mean like the trade-offs

so that's a question so that's not a not an opinion okay yeah I think so the question is can someone describe my answer is I'm not sure I can't did Benedict already leave yeah what say so is can you say it again describe the difference in their product argument uh which is like the pasta curves yeah in zikash uh they have a way to do this without trusted setup but apparently there are some speaker performance um trade-offs which I'm not aware of maybe someone knows yeah what is there was an episode I did recently with Halo Dara and strad we talked about that and I don't remember

so you can listen to it yeah I think we talk about it that was what I was thinking about what we can't listen to this question I think like the the the like the trade-off is what like inner product stuff versus like kzg it seems to be like the uh the current trade-off space of people who are building these like general purpose protocols but I am not the person to try and explain this either all right any other questions comments ideas back there oh there um I guess my question's a little bit pointed uh at zksync uh scenes is uh you're up there on the stage I get to

ask you a question um it's really cool what's happening with the uh ZK EVMS but do you think there's more of a fundamental architectural kind of a different approach you could take rather than because right now with the evm contract the state has to live on a effectively on a single machine right there's there's a capacity to how big uh the state can be right and so is there a thought into the direction that that could go into expanding that into a different kind of architecture as opposed to something like ebm I know there's obviously shards and stuff but ultimately the data still

has to fit in a smart contract because with ZK sync right now I understand you put all the data back onto ethereum right um so the there is no fundamental limits uh on how large the state can become and how we can process it we could use some parallel transaction processing techniques uh um to to like you know like prepare the block before it's been proven and proving always happens in parallel we can parallelize all proving really well so uh it's not really a question of how you structure specifically zkavm it's just how you build your blockchain in general one

thing that the kvms can do differently from ethereum is actually diverge from Geth and the existing ethereum clients that are optimized for uh the broader usage like before for having a lot of full nodes run on consumer Hardware on ordinary laptops and uh build Special clients that are optimized for for for being run by validators where you have higher requirements where you can say let's keep all of our state in memory so it's going to be a lot more expensive but like in in the context of having only like a dozen of uh those validators actively running the uh the consensus it it can be a good trade-off so

uh and this is what we're doing like we're not using get we're building something that will be potentially more efficient and get as a full node but that's not that's not specific to zkevm in my opinion do you have any thoughts if not he's going to take your spot cool is that the the sacrifice is accepted okay all right come on up and thank you for being up here Saki and I hope you make your call all right here's the quote Oh wait any questions throwing it out there okay there's somebody there

uh hello guys so after D5 2020 and nfts last year I think now it's a comment for me of course but I think it's identity season and how can we leverage uh zero knowledge to to protect identity of one cell phone to blockchain do you have any thoughts on that sounds like a comment yeah comment slash question is your name Alex because you have to be named Alex do you sit here my name is okay okay it's 80 chance okay do you does anyone

want to answer the question about ID zkid or ID on the blockchain uh so this isn't really my field so I may sound like a total idiot here but I'm gonna try my best um to put this kind of in Monero terms just I'm not here to show it it's just the best way to understand it and if I'm gonna give my best answer I have to rephrase the question a bit um there's talks about his transaction protocol and one of the things it's doing is it's working towards I'm really not trying to show him an arrow I'm just sorry but my point was is they're taking a public key just as we all know them except this public key is actually the aggregate of multiple public Keys each for their own purpose and because yada yada just

create log problem you can't change any of the individual public Keys you have to have them all correct together and when we're discussing these decentralized identity I think the simplest way you could Implement such a system is simply by having an aggregate public key you can say oh yes I have the public key that says I'm a U.S citizen because it's not actually a public key it's just a point and you're just saying yes the point that says I'm a U.S citizen is in my public key which is what I'm using for decentralized identity but you can very quickly and trivially create a circuit whether it be graph 16 whether it be some bulletproof you're running on your computer whether it be a Halo 2 instantiation that just blinds it it doesn't actually change any of the value

that you've embedded in the point it just says Hey previously I was using this value but I know some offset that lets me transform it so an app doesn't need to know who I am at a base level it just needs to know I have appeared as this new static identity and I have these properties in my public key and obviously that doesn't solve the problem where you have multiple people creating IDs yet you can also use ZK snarks to say yes this offset is derived from a very specific process and because of that I'm not replacing it with anything else but even though I'm creating it with this specific process it's a snark you don't get to see how I did it so part of me wonders why decentralized

entity hasn't done better because I feel like when I try to describe it it's very simple and that's probably why I don't work in the field because I'm missing something but that's my attempt at a stab at it cool any thoughts ID yeah no I I agree I think it's uh it's gonna be the one of the next big things because when I am asked to like log in with ethereum with my address uh or like receive Bob nfts on my address and like also a touch um for convenience uh the ens name uh like with my handle to this address it's getting weird because like now you're logging in everyone knows who you are everyone knows where like all the

things you you've done where you've been with all the nfts with all like you know like all the like and tickets to some events that you purchased it's strange so you really want to be able to connect in a way like this Apple ID uh is connecting you like which is which is implemented better than Google ad because it uh creates a like unique email for every app that you're trying to connect with and it only shares this like random pseudo random or like a pseudomized uh uh ID to um to those things uh totally um and it's actually not that complex to to implement so it's it's a matter of

good ux and a standard that everyone will accept so this standard question is problematic you can think of this um XKCD uh standards yeah I think that's the problem in general is coming together on a standard that everyone agrees on and I think identity suffers from a usability problem too and so when It suffers from a usability problem you basically just get metamask uh installed in everyone's browsers and that's it um and you know it's very there's a there's a hundred different wallet providers all trying to break into this uh Market But ultimately everyone's installed metamask now so it's very

challenging so obviously as an industry we need to grow and we need to obviously keep working on on these new technologies what I would say is I think we're missing a layer of uh and it was really interesting I think one of the other speakers talked about creating something of actual value for users because that's what makes them install the new browser extension that's what gets them to drive change that's what gets Brave installed instead of chrome and I think we need to think a lot more about what technology and and yeah so I'm a the founder or co-founder of space-time we're building a decentralized database and one of the things that we're really passionate about is rethinking the whole model uh so yeah I won't go into it but I think I

forgot to do an intro to you okay oops switch with not Alex what's your name let's find out who you are and we're going to do another question in one second so thank you Alex for being on and welcome person who does ID what is your name uh Alex my name is Miguel okay different name what so just introduce yourself real quick my name is Miguel I'm working on a lot of Open Source software like secret infrastructure and the game yeah cool all right any more questions ideas comments I think I saw a hand right there um hey so

um well it depends on you ask a strong argument against switching to digital voting from ballot voting is that the ballot voting process is it's leaky it's faulty um it's not it's a yet has other thoughts but it is auditable by everyone and going to like a like a digital learning scheme would require very sophisticated one that then puts this into like an oligarchy that can then that is then the only uh the only set that can actually verify that this was done correctly um now this argument has um risk a slippery slope that it can just become anti-intellectual but I do think um some of this um applies to the question of

um let's skip out privacy for now but like using ZK for scaling um if you go for this kind of complicated you could also just be waiting for technology that is much simpler for scaling than using zero knowledge Technologies to achieve this um which is kind of the tension that Bitcoin has so my sorry my question here would be like what is your does this apply to zero knowledge Technologies as well like this this argument should we replace just for the sake of scalability you replace systems like for instance ethereum that are much easier toward it by a broader um set of developers or eyes than something um band technologies that have a much more restricted set of eyes that can

actually audit it should you replace the question again yeah are there cases where you would say nine uh ZK shouldn't be replacing this technology because the other technology while it is less efficient um is much broader clay can be audited by everyone okay say for instance Bitcoin you have any answers on that I can give it a go go so I think what you're asking is is there a downside to privacy and is there a downside to making things zero knowledge so no one can see what's happening right no not quite it's it's the technological

Implement implementation of that so how complicated the implementation is to audit for a broad set of eyes oh okay so like in the same way that like pilot voting like anyone can like check that very locally in their District pilot voting was done correctly yeah well I would actually argue that how many in here have audited a ethereum smart contract themselves or how many times have you done it 100 of the time for every smart contract that you've used probably not I think it's really unrealistic to think that everyone's going to audit every single contract that they use and so actually I think this is one of the things that we need to look at as an as a uh in web 3 is

okay how do we make it a easier for people to look at a smart contract and understand what is it doing uh and there is obviously an element of trust here but you still have to trust that you know the web dap that you're using is still using that technology and not actually going oh I actually want to use this other contract and obviously there's things around signing but again who how often do we check the contract key you know that we're that we're signing away from so I think one thing uh that we that we rely on today is the fact that people can audit and people do audit this technology and I think that won't change because it's CK technology I think in fact um really the the fundamental thing will be will still have the same problem that we

have today which is we have thousands of different circuits or programs that we need to find a way that people can audit them easily and some of those there are ways around this right rather than having everything the the developer to be able to change everything and write very explicit code can we create some sugar syntax to say okay if we're transferring from X to B do you really need custom code to do that because if you have custom code you can sneak into some some stuff like if you've looked at the openc Smart contracts they have a god Clause where it's like oh if it's open c key you can just do whatever you want right and so these things exist because we allow people to write custom code so

I think yeah that's probably my answer oh okay can I just add a sentence sure I was just going to say that I think no matter what protocol we discussed there's going to be a few different numbers to look at there's going to be amount of people who would like to audit it but don't try to amount of people who try to and that amount of people who do and I think if we throw CK into the stack the amount of people who will try to in any decent manner will immediately go down because they'll shut off they'll realize it's not for them and well yes we can dream of a future where it's automated we do have to consider if we practically want to get more developers involved we have to promote a level of accessibility and anytime we pick a more complicated protocol that's obviously

going to be a trade-off it's decreasing the amount of developers who can participate but it's gaining a sufficiency and no matter if we're discussing ZK or not CK if we're going to get more efficiency we're always going to appreciate that yeah just to conclude I think it's always a trade-off between education and design options so we have to to figure out what's the best option for each case and yeah that's it there any questions comments anyway anyone else throw it out there you had your hand up what happened no you can you can bring it back you can actually bring it back to any topic

we've talked about today it doesn't have to follow this is a very fluid format oh this is going to be even weirder actually okay this is really actually I think it makes sense for Ezekiel audiences I'm going to really put my tinfoil hat on okay uh and be and ask the question um what happens if one of the cryptographic assumptions that we have for a lot of these CK protocols is broken you know on the auditability process like how do we know how do we how do we know how do we know that this isn't like another uh nist debauchery where you put in specific numbers that happen to be surely random but probably end up having a back door how are we sure and do you trust yourself on that wow hard question

okay you're gonna come up just because you're gonna come up but I don't know if you can answer this question well a lot of the ZK technology is based on the encryption that has been proven out so I would say it's not just this magical new thing that we've decided oh we're just gonna uh try this but of course with as you mentioned um there is concern and I think that's why we need more people creating these different types of algorithms rather than taking them from the government which is definitely not a good idea uh we could maybe take them from from a broad range of researchers and have more auditing done and I think that's really the only way and ultimately you know Shah one was was broken right so we need to be aware of that as an industry and

be ready to move to uh something that's safer uh and something that's better is is this true actually because there's there's a bunch of weird algebraic assumptions that we make for a lot of the ZK systems that we have um but anyway sorry I think it really depends on what assumption you want to talk about if you want to talk about the discrete log like well if that's broken we have a lot of things going on tomorrow personally I'm waiting for uh P equals NP hard crypto to come out if someone got that let me know um but then of course we have a variety of other assumptions we have discrete logs sure we we trust that we rely on it every single day um and then in threshold signatures I believe One Security of model of Frost which I

cite just because I also have worked with that is premised on one more discrete log which says if you know a bunch of discrete logs you still can't find another and as a non-academic that kind of just sounds like the discrete log problem I don't get why it's a different assumption but obviously in Academia we constantly have these new assumptions what I think it's important to look for when we're discussing ZK proofs is one that they're actually formally proven have gone in review but two um my big red flag is when they introduce new assumptions because that either means they're revolutionary or you should be running away as fast as possible to Circle back to the narrow just because that's what I'm familiar with I have a frame of reference okay um we were excited for sure thank you I appreciate that

um we were looking at this one protocol called trip Tech um it was either triptych or Arcturus that had this property I think it was triptych um and it was based on a I think they called it a dual discrete logarithm assumption where it was impossible to find two discrete logos blah blah blah nope that was broken uh we the person someone decided to do a proof of concept implementation just try to move it forward nope it was worthless that assumption did not hold so obviously there's varying degree of assumptions but if we're not introducing new ones and we're going through formal review I'm more concerned about the implementation attack surface rather than the theoretical one I guess as a as a person who writes a lot of weird

peer-reviewed papers what is formal review to you and and specifically does that apply to AGM as well right you're getting up here you guys are actually switching places soon um so I actually stay out of the formal side of Academia so it's hard for me to comment hey I just work here okay I'll write your code I'm not going to build the next bulletproofs um Beyond obviously conference inclusion which of course doesn't magically mean a paper is valid but obviously it's a great show of confidence in it um it does kind of become an ad hominem thing where if there's developers I respect and trust you say yes I've reviewed this paper yes I didn't just skim the abstract I read over but when you start having other developers review it when other researchers shy man say

yes this is a valid proof yeah you believe it's a valid proof so obviously it's about web of thrust but the easiest check mark for me is always just going to be conference inclusion I did hear one thing today and this is a second hand actually uh but uh one of the one of the challenges I heard today was that people were because ZK proofs take uh quite a long time to prove people are a little bit more risky on the encryption level that they'll do with them so they'll just turn it down a little bit because it proves quicker so that's maybe something we need to watch out for as well okay I'll give you uh instant verification if you accept anything I say camera you're taking the middle

[Laughter] but yeah oh my I'm up here now dude can you answer your own question did I answer my can you can I answer my question oh I certainly have thoughts about uh peer review in Academia oh yeah but I think I think we've talked about that at an episode of our call correctly but that involved level of sake so I guess if you want to hear the full thoughts uh it probably involves at least one bottle of sake that's true that's an open open invite for anyone but anyways more generally yeah if peer review is kind of a weird a weird thing that is unclear to me and um you know unfortunately it's kind of the best thing we have as far as I know but if someone come up with something better

then and oh no oh no okay okay the hair color back on stage well I mean I just want to point out that we we have an industry where like the product is the bug Bounty this is true that's another alternative approach Henry do you want to come back up you could no okay any other thoughts questions people who want to weigh in this is the last session of the ZK Summit so if you had even if you had like thoughts from previous talks just sort of get that conversation going all right um so let's say you don't do math for fun and you're at your first CK Summit and you've had that bottle of sake and

you're like okay um I want to build stuff for users and we know that Zeke amazing scale we know that ZK can make transactions be like worth microsense um what's the most important stuff in like the last six months or the future six months when it comes to new capabilities that allows you to build product for users um that is awesome or is in your opinion that's a great question I don't think there's an opinion in there I think that's a good question right it's pretty balanced I wasn't sure I understood the what was the last part of the question sorry could you repeat the very last part not including scaling and like microsend transactions um what new capabilities of like current ZK attack um that's showed up in the last six

months or you expect to show up in the next six months has the most new useful and interesting capabilities to build user applications right so what is and what's not use like proposed use cases for ZK basically not scaling yeah or like I think zika can be very useful for compliance issues uh then Bonnie actually talked about about this Elite uh like uh compliance is very important like you can just submit SDK proof and and like do your taxes and stuff like that and I I'm very bullish on compliance scalability was very talked about today privacy as well so compliancy I think is

my autotrike so I mean uh yeah I I'm a bit biased but I think uh our blue haired fellow right here actually had probably one of the like craziest weird ideas where ZK is kind of used for something quite new uh but it required creating new mechanisms so specifically for for things like shared like very Global State uh things you can do with ZK are not obvious at all that you could do before uh so I think that's a that's a very cool application I think I mean and this is uh I'm assuming by kind of like cheap transactions you mean like you don't just use the succinctness property of ZK you actually use the ZK part of ZK um I think is a very uh like I don't know anyways that's what I'm assuming

that you mean by like cheapness or like easiness like you don't count that as part of the Cool Tech that's coming up I think that is a cool one uh yeah probably a couple of things like I I think this space is moving so fast that it's hard to keep up with all the new improvements like every couple of months a new paper will come out with a new exciting idea that's kind of like making things faster and I think if we can get ZK Tech they'll become an inflection point where ZK Tech becomes so fast and so good that even from a scaling perspective right it just makes sense for uh it to be used and I think the big thing that web3 is missing in general is that usability layer like I'm not going to use an app if it like takes

20 seconds to like submit a transaction right and so there is coming and it is whether it's six months whether it's a year it's coming where you will use adapt and you will barely notice and that's where we've made it and I think then a lot of useful things can be built on top of that I want to weigh in on this so you said six months before six months in the future feel like there's ZK for compression was an idea for a long time like like clients it's sort of in the scaling department but it's like making something smaller um I think we've seen more implementations of that what you talk about at least the way I always understand is like using ZK or other privacy Tech in defy that actually

allows for like either strategies to be really really well masked which is like it's privacy but it's like a very unique form of privacy or where it actually kind of recreates the way that some of these things that you study work those are the things that seem very kind of new in in that regard I'm trying to think of it like because that's what you're looking for you're like you're thinking like ZK use cases because I think you're talking more like ZK optimizing like it getting faster but like you're saying what is possible what's becoming possible past scalability but but I think if users don't use it it's not possible to have us like like how many users don't use doubts because they're so slow like I feel like usability and speed is a

feature at this point because it's like I'm giving you something so much faster it's such a shift and I understand what you're saying that would just be my Counterpoint is that like I think it gets to the point where it makes a big difference oh what's the what's the most stuff in ZK oh that's so harsh wait do you want to ask a question or do you have yeah yeah actually I I would have another question if it's okay like like regarding security uh how rather relevant do you consider like upcoming Quantum Computing solutions for like CK especially snacks so uh oh just so I think this is more of a question about

like Quantum Computing in general I guess maybe like I can weigh in on this one a little bit because I did work in a Quantum competing lab for my PhD unfortunately but anyways the answer is at the moment exactly not relevant uh maybe over the next 20 years probably slightly relevant but it's shocking the progress that we have made it is also equally shocking how extremely unsophisticated we are at doing essentially anything that is currently like truly quantum um in fact I if you know there's been a whole set of things about like Quantum Supremacy and whether that has even existed and some people have claimed it has and then someone goes and writes like a algorithm runs it on a super computer for like two days and it's like congratulations we emulated your results

um and this happens so often that the the the taste in my tongue that it leaves and and you know I've seen it in in lab settings as well is very much uh you know we're still 20 years away uh what it was said you know 20 years ago and so we'll continue um but you know then again we might say that about whatever Fusion or something but maybe at some point it'll actually happen so um I think that's that's the highest level version we can get more into the details but that would be again that might require social sake so one question are you more confident in like latest cryptography or uh snarks oh I I actually don't know a lot about real cryptography yeah like actual cryptography unfortunately I know very

little all right so any other so that I think that was a question so thank you for the question but so uh any other questions comments the way this always works right it's like the more opinionated the more likely we're gonna do a switch oh I do have a question is that can I ask the questions like everyone is that a cool uh oh my god oh wow you are hacking my format sorry for it yeah yeah I mean should we separate Justice this is piggybacking on the previous question should we separate succinctness from like what we call zero knowledge like as a whole like is that a good thing to do like right now you know we kind of when we talk about zero knowledge we mix in all of these things

um is that bad is that good is that you know maybe yeah maybe Alex's question one answer too I don't want to answer that yeah I was um thinking about the answer um oh and I guess one side comment is can we do better by being just a synced and not zero knowledge do you know if there's a gap there do we know that there's a gap there or maybe not maybe well I think Starks are by default just a succinct not really zero knowledge you have to do extra steps to make Stark's your knowledge you have to go right like uh beyond the uh the the domain field so uh yeah it totally makes sense to separate and I guess why don't we maybe as a an

additional so I'll I mean I'll give my take quickly I think like Alex increasingly You're Gonna Get You Know systems that are you know tailored to the use case specifically so I think where you know the way I like to think about zero knowledge systems is they conceal and they can compress right so compression compression is the succinctness piece I think in like cases of Roll-Ups right you kind of the zero knowledge is kind of an extra nice property this may be like okay you can verify that this is correct but you know the Privacy The Hiding piece isn't as important and I think you're going to increase because everyone is aiming for maximum performance it's inevitably going to get to the point where you rip out everything that is not absolutely essential to the use case that you're serving that's going to be the way that

it goes so I would predict that you see these systems start to Branch off and just serve one or the other things so I think in some cases it could be good it depends on the use case there was a question back there who had their hand up I think important too you can actually just specify you don't want zero knowledge in the configuration does it actually help you though uh I think it's I think it's faster if you don't use it yeah oh it's actually like not just marginal yeah no they actually say in their dogs like don't use it unless you need it okay that's a actually I did not know this all right sorry note to add on to the point I think it's actually in a way dangerous not separated because I've I've heard people Founders actually um telling me they were doing this

wonderful ZK app and when I asked about this UK it's just because oh because we are on this one ZK roll up like that was it that was the Halls UK so I think it is actually dangerous and there's a third thing it's not just um the zero knowledge the sickness there's also just the validity of what's going on and you could that's a third like axis that we have to sort of separate out I like that I think you can come up that sounds like opinionated you've been volunteers I think you're off thanks for being on the panel do you want to introduce yourself yeah um hi I'm Nico and I'm a cryptographer at geometry we have a cryptographer here now so we can ask all the cryptographers

yeah now you can ask hard questions Alex is like let me get out of here thanks for participating thank you for your awesome answers questions okay yeah any any wait you had a point so I don't think that wizard would you want to ask me no probably just the points okay does anyone else have any questions comments thoughts about what we've talked about already here thoughts about talks that have happened things that you think this group might be able to answer we do need some so questions or this doesn't work so yes sorry to push but actually what I need

okay so this question has to do with nomenclature if you will so as terms matter do Roll-Ups should you consider something to roll up or an L2 if it has the ability on L1 for ethereum to be able to um immediately change that smart contract for upgradability um so should you if you can't if they have the ability to stop those transactions from happening on and from you being able to escape and force through transactions on L2 should they actually get that ability I don't think they should I guess is my opinion on they should have a separate classification based upon the ability to change their protocol to kind of freeze funds within their L2 so

yeah from the way I see it roll ups will just use like execution of chain and then use the main chain layer one for security and that's how I see it I don't know if you want to add anything I mean yeah I I feel this got sent to the is a like as a is a well a fish or a mammal kind of question like you know it goes both ways like you know once one feels like a much more useful classification but the other one feels much more technically Adept so I don't know I mean upgradability of smart contracts from from the L2 itself seem yeah like it it isn't you know maybe it's a useful distinction but I would

add it as like a you know trademark after L2 as opposed to like make a whole new nomenclature just for that if I can Riff on the question maybe for the for the panel that might be interesting is so you know there's a lot of discussion around where economic value will accrue in a world of L ones and l2s right well the mid so in your opinion well the majority of the economic value accrue to the L1 upon which the L2 is built or will it accrue to the L2 which could potentially serve multiple L ones can you actually hold that because I want you to come up here and then can you ask that again yeah please come up please please come up no you're right [Laughter] thank you for being on the panel and welcome

okay you're gonna have to so so I'll actually just first quickly introduce yourself and then Alex if you can just hello I'm Brian um I I'm immigranty with the ethereum foundation so hold on so uh the question restated uh in your opinion between the L1 and the L2 in the L1 and L2 Paradigm does most economic value accrue to the L1 upon which l2s are built does most of the economic value occur to l2s which could potentially process transactions from multiple L ones or yeah so that's the question where does the economic value occur I see Guillermo scoring which is why I asked this question do you want me to answer sure um so I guess my opinion

I think as we move forward into the future in onboard like a ton of users I think that we will see the um the value accrue to Applications more than we will to the actual Network as much as I think the networks will actually continue to accrue value I think that the if we see true economic value of the real world come on to networks I do think that they will hopefully suppress what's actually happening with the l1s so I think l2s are going to take those applications and they're going to probably do like whole l2s that are like wrapped up in that one application and I think they will see a ton of value accrue to them I actually think if if it doesn't happen

if the applications don't get more valuable then the project has failed like the whole thing like it sort of has to it seems strange that we constantly just focus on Rails as the primary but for the application I mean applications to be secure about there needs to be economic value at some level right where you're settling certainly like in a financial application where it's settled this is the idea of ethereum right is why it's secure is because the miners have this huge economic incentive to play by the rules that's the whole premise so I guess that's the I agree so great first of all great Dodge of the question that the application but yeah like the the problem with this question is that and I feel this way about a lot of weird economic questions

the kind of silly thing to say feels right uh but it's probably totally wrong and we'll have nobody until it actually happens um so so you know like we can talk about like weird supply and demand curves and being like Oh like of course like the thing that should you know that everything is running on should have at least as much value as the things that are running on it but and that feels like a totally sensible argument but my suspicion is that it's completely wrong and we'll have no idea why um and then someone will make some horrible model you know five years from now and be like oh of course it was like obvious and everyone should have known that was not the case that the l1s wouldn't need a crew as much or not a crew as much value as the l2s that are

sitting on it or something so uh I will Dodge the question as well um but I will do it with a with a something that sounds intelligent which is uh we have no idea and in fact I don't think we'll have an idea uh but it's gonna be really fun five years from now to be like to write some paper that says obviously we should have known ahead of time that indeed and it's going to be peer-reviewed everyone's will be like oh yeah yeah I was obvious you know reviewer two is going to tell you sorry I rejected because this should have been stupid you know we all did know that back in whatever 2022 or something I have one final hope to have my question answered but like what do we mean by value right is it just the the gas fees that we're

spending on different layers or is it something we charge users for a service other than just computation right it's a good question um and I don't know this panel is an expert question Dodge us to ask questions so I do politics okay so because there is another question so I wasn't sure if we should wrap up now or after so Alex I think you're up oh me you're taking Guillermo Guillermo if you don't mind being the mic person and there's a question right there thank you Guillermo no I think he he had it yeah he had a question oh I didn't actually have a question I had a comment and the comment

was not actually going to answer the questions also going to dodge the question serious I think the the value if we get it right will be actually for the users of these platforms right like ideally if you can have like a system where you you own the data and it's all privacy so you have it and you can prove stuff about your information then all of your information and that value goes back to you right and so actually a lot of these economic systems will eventually just kind of be very competitive in terms of like uh competition in in terms of providing the service whereas the users hold the key to all the data anyway sorry not answering the question so the tldr is like like anyone who saw the Ariel's

talk earlier today is I don't know wait Alex why don't you answer the question what's my what for your my own answer question yeah I think any so my own answer is I think anything that supports applications that people will use and or use it in the way that we think of web 3 wherein we have this economic incentive at the base layer that will be what a cruise value I I think the L1 L2 framing is wrong because I think there's a potential things we think of as l2s today could serve multiple networks and then kind of over time transform and become some root layer but I think there will always be this root of trust and I think again in this in this Paradigm which you've all accepted where this game theory and the economic security model I mean if we

don't have that Economic Security model you don't really have a blockchain in my opinion that means I don't know I mean maybe that's a more fundamental question is a blockchain a blockchain without an economic security model [Laughter] twist if you want to say anything more because I'm going to bring him no I'm not going to say like I'm hearing dodging your own questions you're up here and you're thank you for being on this you've already introduce yourself again and I know there's a new question here or comment or Dodge ball is what we're playing now I guess that was just a question so I'm wondering do you think like the the succinctness of validity Roll-Ups or in

Shrine Roll-Ups is like the only end game for you know blockchain Center distributed Ledger technology or like if you want to build you know one billion user application um or can you actually build these applications on like Solana or Cosmos or you know other other uh sort of scaling Solutions no I think the architecture for existing blockchains of course shouting will help but you're just taking an architecture and making more of it I think is not the right way to go personally I think Roll-Ups are able and if particularly I can't remember who it was here maybe it was you who were doing these like micro roll-ups oh yeah yeah so with the Elio every

transaction you could think of in fact right and I think like taking taking that data and doing these smaller Roll-Ups uh is actually the way to go and that's something that we're looking to to implement as well in our decentralized database is to have the ability to basically split the data Out and because you've got the ZK proof you can actually take data from one place and prove it in another place very easily so you don't need to have this conglomerate Global hash that just lives right you can actually you know prove any piece of information uh anywhere else and I think that's the future of scaling and a lot more computation because snarks will enable us to do all the computation or a lot of the computation on the client side actually what these blockchains will become is verification networks

um and all of the data will be owned by you there may be some level of data that exists on the network because you know you're not online all the time you know your phone's not online all the time but fundamentally yes we do need to re-architect the way that we do yeah I agree with that um I think yeah I think that was very well said and you know I think it answering you know without dodging I think the answer is yes this is the model and in fact I think you've seen this like this whole concept of Roll-Ups has kind of expanded to this concept of a modular blockchain and really even a more fundamental level it's separating execution from State storage and separating all the different parts from each other right and this is one of the reasons why you know I think if you look

at a lot of these snark based systems the the key bottlenecks are often how big is the proof right for a decentralized because a decentralized system every node has to store the whole chain right so this is like how big is the proof and is there an ability to recursively prove it down to be very small but even if there is that's fine for clients but somewhere someone has to store that state right so the state like state is one part of it but then the execution I mean because you're proving off chain it actually opens up this amazing new dimension for applications right like you know if you can if you're not limited by the runtime of like an evm you could run a program for 50 years years spit a proof out at the end and just and have the network verified and

that's something you can't do on ethereum right today that's like a fundamentally new model which is cool but the bottleneck becomes verification time and this is one of the reasons why we talked a little bit earlier about snarks versus ipas or Starks I mean this is one of the reasons why despite all these really new interesting proof systems snarks have remained the tool of choice because the verification time is cheap and the size of the proof is constant despite no matter how large the computation grows do you think was that a question or a comment me oh his kind of in the middle I wasn't sure are there actually the question here is are there more questions comments because if there aren't then we'll wrap but if there are then

we'll continue can I just add give someone time to think of another question maybe just add on to the end of that sure uh I think there's gonna be an absolutely huge opportunity as well like if you think about how big an industry cloud computing is like I don't think people realize that right like cloud computing is massive uh and if you took all of that computation that happens on the cloud and you put it onto the client like that's a lot of value now going to the client and that's just going to be super interesting okay in that case you are coming up let's let's get some new faces thank you do you want to introduce yourself

hi my name is Philip I work as a research associate at Venture Capital firm called Rockaway based in Prague cool welcome uh there are there any other okay there's a question over here um hey can you hear me yeah um a little bit of an opinionated question but going back to the one that everyone dodged about where it would value a crew um curious to hear do you think where value will accrue will change depending on your time scale so uh one year out versus three years versus five years versus ten years do you think 10 years from now it'll be application Level but Five Years it'll be a one three years

it'll be L2 do we think that may be a better mental model um do you think it will change over time like that I can hear you trying to put comment in there that's a good question I'll take a stab since I I articulated the original dodged question which I myself dodged I I think it will absolutely change I think it'll change based on you know what parts of the system become less or more expensive over time and that has there's a lot of factors there right so I mean technological improvements I mean if you think about Moore's Law like just Moore's Law and what that's enabled over time for computing I mean that's that's made certain things that used to be extremely expensive Commodities that everyone just assumes are free now right and I think you could you know the same

thing with storage right we've all gotten used to having you know terabytes you know on our laptops around which is like before like terabytes where you know that concept was extremely expensive so I think as like as the hardware kind of matures I think different things are going to be viewed as cheaper and and then but of course there's a market I think the market will also determine for different applications uh like weather storage or like state state growth or State uh maintaining state with some reasonable degree of you know retrieval latency is more important than executing something so I think it depends on the time scale but also depends on the application can can I add something on that um before you do that I want to switch out okay so do you want to come up I feel like there's more questions I saw

some other hands so that's good okay oh actually wait he's gonna introduce himself you have to wait for him to get here uh hey guys I'm Matt I'm building a private multi-cigs on Aztec and Alia cool um yeah so I actually think the economic model is going to change uh completely like today we think about companies in a very different way and I think the future of web3 companies um is is going to be completely different so today as a company you uh you build a product for example and then you get monthly or yearly annual profits from that and that goes basically back

to the shareholders who put in some money at the beginning and I think actually what we're seeing uh with with dials that have been created and around protocols is actually you take the value at the beginning so you tie the value proposition of your uh protocol to um to your coin and so the more usage your protocol gets the more the coin goes up so all your initial investors get returned once in a big amount a big chunk of money they get returned and so that's great they're done and then we can use uh the rest of the kind of economic system can actually balance itself health and not be as much of a prophet anymore so I mean this is the dream yeah I mean I think there's a lot

of other stuff happening that pushes those prices around the utility of if you look at the time yeah and actually most if you look at most protocols there's no reason why the price of the coin should go up based on what's happening on the protocol and you can build that in there are economic systems to build that in uh the graph is one of the people who are doing it right and they actually have a system whereby uh it's it's called the utility token model and you can basically tie using a ton of Taxi Medallion type methodology you can basically say okay if you want to use make money on our platform you have to stake some amount of money and then that that makes it work any other thoughts questions comments we

have oh we're kind of at time maybe one last one I didn't have one uh one comment on them okay yeah because or we take it from here go for it um I I was my opinion is actually more in dies Camp about you never really know and it's gonna be a complete up in the air thing that we'll look back on and be like oh that made sense but I do think it'll be like there's so many different ways that it can go that that wave function will collapse and everyone will be like oh of course it went down that path but like Cisco was during the initial internet build out was the most valuable company worth 500 billion dollars at one point that everyone just projected that oh that one's going to be the one and they're going to charge the most and

that wave function never collapsed until it did and then it went down 80 and Facebook so I think that's a good point to end this on we we don't know we don't know anything zero knowledge all right thanks everyone for the park bench panel it's a lot of fun uh I believe we have another slide we can maybe put up which is our last slide I don't know if you have this uh maybe you do maybe you don't uh I don't know if you do yes okay so we are now going downstairs there is drinks there are drinks I can say that correctly and uh least Authority has a special kind of gift drink that you

should definitely check out they have written there's like a textbook called the moon math manual I don't know if you've heard of it it was in the git coin rounds I think I've mentioned it on the show in the sponsorship spots but yeah you should go check it out because it's connected to that this is like a ZK Tech textbook that we've been waiting for that's at least like a first you know we're they're starting to come out they're it's not the only one but it's really really cool that they're doing that so please enjoy the drinks and thank you all again for coming to the ZK Summit eight see you at the next one I guess right