zkSharding Crash Course with Crypto OG Misha from =nil; Foundation

Transcript

welcome back to the rollup High Roller Community as always nothing on the rollup is to be ConEd as Financial advice we are just an Educators in the defi space toss us a quick like give us a share a repost send it to a friend who may want to watch this and finally give us a sub or a follow it means the world to help support credibly neutral educational content in def5 we're here to educate Empower and enrich you and let's jump in [Music] [Applause] [Music] [Music] [Music]

[Applause] [Music] [Applause] Misha what's going on Welcome to the rollup how uh how's your Dev connect been going so far about about 10 dos I don't know literally everywhere so I'm I'm done with it but it was it was it was cool it was cool it was worth it it was totally worth it what's your favorite zkm uh you know the one which has no security issues which is your least favorite

zkm you want me to call particular names for something just asking questions like it's like those ones which get those ones which get audited and then they and then and then those ones which uh you know get reported as as being hacked it's like yeah those ones suck I don't like those ones um what's your background well initially Al topology I mean then it's uh I don't know complexity Theory simplical topology y y and here we go like cryptography here we go it's like the background to you want

like you know projects before that or something okay I mean well bit message in 2013 it was like a no heck know infation that it was like bit chairs little bit of a stimate uh then like we had with Quest we had the Russian dedicated for of steemit as like we've been running it for like about 2 years Still Still Still Alive still there still works just like steemit is so you know Justin shant yeah yeah he he acquired stim it uh back in 2019 or something and then they forked out uh creating a hive block thing it's where the original Community

went and stim it became filled with uh I I know whatever Justin son has filled them with has filled it with that's what it is it's a long story long and complicated story a lot of forks in there like a big like a drama that's the thing and yeah then April of 2018 nil and since then it's nil five years yeah it's like a lot of things happened within nil if you wanted but like in terms of for all these five years but it depends what was the uh first first uh uh what was the price of of Bitcoin in your earliest purchase in 2013 era damn sub

$100 I don't even remember it's like I was more on Litecoin back then of people that were buying Bitcoin at sub $100 are sketchy but we're just going to keep that conversation out of this m may want to hold the microphone a little lower just when you when you talk um so I I think my question is because we were talking a little bit before this about ZK sharding um and then maybe two years ago maybe a little bit earlier uh vitalic uh transitioned the ethereum road map from sharding to rollup Centric um so could you explain why like maybe why he did that okay yeah that's

actually I mean that's actually quite I would say understandable and uh that's you know it was very logical step the Nuance is you know the transition to full prodal level sharding like on L1 directly first of all it's like very risky second of all let's say this way there were still back then quite a lot of design questions about how this should be properly implemented so in fact the transition happened from you know like protocol level sharting to application Level sharting it's still sharting it's just application Level sharting it's like the difference is that you know uh like rollups are are application Level sh like in fact okay let's be honest and uh the difference between that is basically you first of all it's like okay well

application Level Shing works I mean it's fine it's good the Nuance is that uh it as we have learned over I don't know like a past year or something uh the the problem is that each of those application Level shards um turn out turns out to be like you know completely separate database with completely separate liquidity with completely like you know isolated everything in in there it's like a modular approach things like this right while protocol level charting was supposed and it is like always designed to maintain compatibility between different chards and to you know some something like like it's like like the whole whole and to maintain um to maintain compatibility and to keep the liquidity unified it not you know balcon

some something like this so that's like you know the primary difference what do you mean by vulcanized uh well it's like let's say this way right uh currently currently existing modular approach supposes for you know each rollup to uh you know get some liquidity I don't know out of east or like some somewhere else and to keep it in there without being able to uh share it to you know like more efficient execution layers or something if those appear like if those arise for example so it's like each roll up stays isolated World Garden right so that that's like the thing and one way say that uh I mean this this has been so by don't know like Bridges and something but Bridges okay I mean that's an application Level composability okay

application Level composability each bridging I don't know how much would it cost you to I don't know Bridge a single thing 20 bucks 30 bucks I don't know something like this between different RS there are no Bridges and they resist uh to building to to just like to folks try to build those bridges so that's like the new on there's also like another consideration like uh some some some other like rollup projects try to push forward the idea that you can actually try to maintain compossibility between l3s by sending the transaction to a sequence to sequence directly and uh um basically you just putting the transactions between L3 coordinate and that all L2 right but nobody does that between l2s that's that's the Nuance L2

is still separated still L2 is still see still you know isolated and L3 is no matter I mean no matter um if you put the transaction through the sequencer between different threes each of them still is a w Garden so that's like you know the it's like the primary problem it's like no matter how hard you try with a model architecture with the current approach you still hit the wall of you know like a world Garden of a world Garden issue so that's like the thing and uh we were like hey how can we how can we I don't know put together like the modularity idea like keep keep the modularity idea to avoid the TR trategy of Commons and in the same time how can we you know keep the keep the

compossibility in there and keep the liquidity unified and you know to not to be polonized basically that's what it is that's that's the first thing second thing is we as uh we used some kind of like you know as as as we try for the interaction between uh as as as we try for the interaction between like ZK sharding as you know like more modern sharded execution layer for e be as transparent to the interaction the interaction of applications deployed in there to be as transparent with original e as possible that it means um we try to prevent the liquidity outflow from eth in this term because uh we basically allow people to you know

like read the data seamlessly from eth and uh next year most probably to write the data seamlessly to e as well I mean of course the fees would be required to be paid but like you know L1 L2 composability is something which is you know ZK sharding also you know basically features to avoid vulcanization of e to avoid know people tearing apart that think yeah I think um the the the big thing with the zkm is that um you can utilize shared provs and shared security mechanisms to kind of uh allow these different uh chains within a certain stat to to kind of interoperate with each other I know that um that's something like the polygon team is

working big on on their ZK evma Jord often refers to it as constellations um is is the is the interoperability with zkm something that that you guys see as like a big benefit for zkm um is is is your llvm kind of let's say zkm as a service is that uh is that kind of taking uh into account Shar Prov shared security are you guys kind of helping to build that uh that that stack for for people who want to use nil to launch a z cavium okay uh let's get into that it's like I see a little bit of misunderstanding here in terms of what ZK VMS are in fact okay so there is

like a very common misunderstanding in this industry that ZK VMS are you know like people confuse ZK rollups and ZK VMS as them they they think it's the same all right but it's not it's like zkm is just a circuit I mean it's just it's just I don't know like the the description of an algorithm you know mapped on the finite field that that's it like it's just a second there is no protocol in there is no protocol in there or something it's like rollups are still yet to get their protocols uh to be established as a separate database that's what they call you know that's what uh the decentralization road map of theirs is what they call this so there are no protocols in there yet there are only circuits and just one node which processes data from from itself to like

to to to to eat via Sam suck so that's like the thing uh no rollup is decentralized yet as far as I know no Z is decentralized yet as far as I know that's like the thing so regarding Ezekiel lvm on our side yes we use it to produce this circuit not a prodical it cannot process any other thing of a prodical it's like the prodal itself is like it's like a z sharting stuff right it's like ZM is is being used only to produce a circuit which proves the state transition of a protocol like L2 protocol to back to that's like the thing so can people use this compiler to produce their own ZK VMS uh to you know just prove their state Traditions to e yes of course and we have like you know some interest from

different rollups to use that I mean Tao were talk Tao guys were talking about this like you know publicly electrified know a couple of days ago or something at at one of the panels so like just an example or like we had several inquiries I don't know whatever so yeah it's like people can use it so does it do will Tao have to have to uh pay to pay or do some sort of uh you know agreement with nil or XYZ Z KVM to use llvm or is it permissionless open like how does that how does that kind of structure work if Tao wants to utilize Neil's lovm well it's like it's like Z lovm is just a software okay it's just it's just

a circuit compiler it's like you're talking about the man monetization strategy monetization on this side I mean we cannot prevent them from you know just taking the software and using it okay so in the monetization strategy on the on the circuit compiler level can only be done by by the thing which we call a proof Market which we were running around like for the with for the last like year before we went out that hey guys there's something like which Powers Market underneath so that's the thing and uh that's you know just and in fact this is just uh like a service for pro generation as we call it but that's what it is there's quite a lot of proof markets currently and everybody's building proof markets but like that's

you know was like maybe the first one and we've been with it like for the year and if they want it if they want to use it to you know to I don't know like increase the pro generation speed for their roll up uh you know by relying on market dynamics instead of you know instead of Contracting somebody particular who would you know get them some improvements but how can you be sure that it's the best B can offer yeah so if they want to use that for the sake of proof generation sure that's going to be the monetization strategy it's open permissionless um the end point for it is going to be deployed on East I know till the end of the year proof Market itself is running on top of ziki sharding on C Shard and like is that in q1 so like that's the thing so that's

like what it is so so the these approved markets um are these EK VMS EVMS now like what what is and they're all they're all kind of like uh an organization of nodes and they're uh organizing some cost and some incentives to generate these proofs what is the missing piece like what is what it means for those proof markets and those in those ZK EVMS uh to make those into ZK full ZK r UPS what is the the missing piece uh to take those ZK VMS into ZK rollups because you said

there's no ZK rollups that really exist yet there's only ZK EVMS so like do we need ZK rollups or why like why are these only ZK VMS and not ZK rollups yeah I see what you're saying so like see it's like do we need do we need the do we need uh for you know rups to establish their protocols like in general um yes it's like because first of all it's just data safety it's just data like it's just data safety okay but like in but like you know one can argue that data safety is been guaranteed by e da anyway but there's also such a thing as liveness so it's like State liveness I mean data on this like it's like data

safety on L1 is being basically guaranteed by Saving pieces of like each transaction happening on the RO database site in each commit log it's like in e chain whatever Bon chain okay that's the thing uh like in cold data and that's the only way you save it in that and uh but end users what whatever you do on a roll up you need its state you need you need its unpack State and it's unpacked state is something which can be only guaranteed by a set of nodes which keep this state unpacked uh and Nod which keep the state unpacked are basically nodes of prot of a role of protocol so so if anyone wants to you know like if anyone wants for

their rollup to be used uh by some by some by someone who requires liveness who requires some SLA or something then I mean you got to have the protocol in that it has to be decentralized at least it has to be stable enough it has to have at least I don't know like five or something nodes it's like but again it all depends on the performance requirements or something it's like if folks have no performance requirements and they only serve I don't know just e applications or something it would be enough to have for such a roll up I don't know like five noes or something if folks Target like high load use cases um I mean like on our side we do understand that would be required to happen in know like hundreds of them but like luckily we have hundreds of e

validators so it's like we'll just go ask them guys you want to host that as well yeah so so like we're seeing like kind of these like sets of eth like like Ian da they're they got a bunch of deposits of eth and now they're choosing to use their stake for da and maybe reaking yeah so very similarly it sounds like nil has a set of eth and a bunch of validators and you're choosing to apply this stake to a different use case is that is that it's like a bunch it's like it's it's like it's like um a bunch of eth and like you know to apply to different use cases in terms of you know just

collecting the stake which secures like the roll up and things like this I mean yeah but that's but that's but that's traditional thing that's like you know traditional thing which everybody does so it's it's it's it's it's of a new on there is only a difference the difference between different cases is only about the amount the the amount of of this e that's the first thing and the second thing the fit of a protocol if it can be fit within the within the shared within the shared security system or not so it's like some of the protocols and like the majority of rollups cannot be fit into a shared security model yet because of the way their protocol function so that's like the thing so

those ones which fit I mean they can use igen layer or something those ones which you know doesn't fit uh I mean okay well we'll just get security our own so that's like the thing so the last couple years the uh the ZK space has gone from uh basically a an idea and white papers and a lot of cryptograph cryptographic research to now you know actual practical functionalities use cases or zkm that have hundreds of millions of TV combined probably a couple hundred million across the big ones now where do you see this uh progressing over the next five years and uh how are you guys planning to fit in well it's like let's say this way um the bet on the circuit compiler is uh

the future Prof is the future future proof bet of ours I think if we're talking about like you know ZK space in general cuz besides you know the obvious ZK rollup application uh that are you know quite of quite quite a lot of applications of this thing you know out there like you know just one example is I don't know back in 20 I know 21 we we have put out such an otion as ZK bridging so now it's became now it has become a thing and uh I mean there are even projects which call themselves this way and like actually building a zq region or something so that's like you know the first thing for example another thing is that there's quite a lot of there is quite a lot uh maybe not really obvious but uh zkl applications you know

for like Risk parameters Landing protocols or something that's pretty obvious so that's future proof I would say uh I don't know Z oracles quite a lot of zi oracles have you know they they they they make sense because uh like as our you know like as our recent Li the uh use case with ZK Oracle report and TV the report tot total tvl have shown there is you know different data aggregation needs D5 protocols have and the fact that these D5 Protocols are guarding like you know bills billions of dollars and uh you know security assumptions for them has to be redused and the only way uh they can be redused

this by these kind of like Z zq oracles so like for example the current Z the Oracle like of ours guards I don't know whatever light this T 17 pill or something that's like the thing or yeah so that's there's quite a lot of applications of those I got to admit and uh we even had we even had a presentation like literally a couple of days ago with folks presenting CK Doom stuff I mean it's actually provable Doom you speedrun doom you prove the fact that she speed vaned like in some particular time and you present the proof of it without disclosing which tricks have you used to speed ranit in this particular time so is it fun yes it's fun I mean it's like is it like you

know some economically viable application I don't know but it's fun so even that thing exists I would say so I would say it's quite future proof bat yeah I think um like any space with Innovation there's going to be a ton of uh a ton of new entrance in space I think there's a big barrier to entry with ZK as far as the cryptographic kind of understanding that you need right I mean there's tons of cryptographers uh who are working in this space but um I I if I were to take an educated guess I would say that there's probably not that many in the world who understand cryptography to the level that you need to have a technical team uh to build a ZK application that said there's certainly some competition in the

industry now although we have an ethos of collaboration and pushing the industry forward competition is healthy so kind of regarding your competition what's the uh Unique Edge that you guys have uh in the market or the vertical that you're looking to uh to kind of you know uh be the uh the main player okay so if we're talking about if we're talking about like you know there are basically two major markets which we are like uh currently at and and then we can just trate over both of them okay so it's like the first one is uh like a tool chain for Z applications market so it's again circuit compiler and proof marketing of ours so this thing this thing is is I would

say this thing is this thing targets those cases when there's a need for you know proven performance and the minimization of a circuit size because I know quite a lot of quite a lot of folks are doing you know proven computations by manually defined circuits of a library which as you said requires you know technical team to do this it's not fun it's expensive it's just you know takes time it creates you know all like you know friction between people in terms of when you coordinate and things like this it takes time problem um it's like another approach is to I don't know like use just some general purpose zkv I going to prove it to the and that's you know that's cool but um the performance of

that you can prove a lot with it okay so it's not like fun so from a tool chain perspective we picked some you know Middle Ground between these two things it's like we still can't prove high level languages I don't know rust CPP or something uh basically to theum but in the same time there is no performance overhead induced by Z KVM because it's a circuit compiler it produces plain circus that's like the new on that's from a to chain perspective from a perspective of uh you know like the KE sharding stuff it's uh again it's like I'd say there is quite a large Green Field uh in um trying to you know fire High load use cases uh on which you know

we're willing to work on to of E again most sh ons programmable data avity it's like what that's that's that's what it is uh Builder sequencers this kind of thing uh like intent solvers block future auctions uh that's the thing and uh you know just pick at least one roll up which can process this kind of float within the unified model not like model but unified execution environment none of them it's like uh and it requires a lot of like DPS at least so that's the Green Field we're covering with the ZK shard in part of ours um that's that's the thing I mean obviously obviously we can cover like you know more traditional evm appications migrating from eth with

the ZK sharding stuff that's you know very traditional Market I don't know if that makes sense to even like talk about it everybody's it's obvious for everybody now um you say there was a a use case with Lia in in production could you explain that a little bit more and the findings that okay so so what's this use cases about so let's say this way uh initially Lia was designed um was designed to be that way so it requires the total tvl data from a beacon chain to be provided to execution layer by a bunch of oracles it's like let's not get into details why and who designed that because milil was also involved into that but

anyway yeah it was it was it was a temporary U we just you know like it was it was just put together in there but in general yeah so in general that was designed that way so let's just leave it that and um it worked and it worked and it was good enough for you know for those years which Li was you know just developing getting getting getting its users making itself more more more popular more interesting things like this right and uh it was fine because these oracles were were run by the same same validators same operators which were hand in ethereum nodes so it was kind of you know Fair trust assumption that you know if they want to screw up something they would have screwed up

with everything like long time ago uh so it was a fair trust assumption now when Li is basically like a backbone of the industry uh the need for you know reducing these trust assumptions um came up because I mean the more that's like the less trust assumptions you have the better it is so about a year ago the first version of uh the first version the first implementation of such a z Oracle providing total tvl information from Beacon chain to execution layer uh was implemented um it ran in test mode for some time wasn't performant enough so we decided to respin this project uh it's like to reiterate this project with a more W tool chain without Z kvms so we

used the circuit compiler for that and uh in fact what resulted right now what this whole thing resulted into right now it resulted into an oracle which reports total Li tvl information to like from V chain to execution layer so the issuance of stake ethereum and not only but so the issues of stake ethereum could be done correctly and properly based on the amount of you know based on the amount of TL which is which stakers have locked so that's the that's that's the thing that's what it is about and uh we try to guarantee the liveness of a pro gener for this thing with with the proof with the proof marcket thing because I mean

it's like the proof has to be submitted as fast as possible okay so we were like okay we'll use this we'll use the proper tool chain we'll apply market dynamics to accelerating that you know as to to to to to those to those theoretical limits which are possible so the combination of this you know like like a perfect use case for a perfect use case for the tool chain basically that's what this Z orle is about that's very cool I think we're going to uh to put you on the spot here where you don't think and just answer and you have two choices um and uh it'll be fun are are you ready all right optim siger or ZK rups ZK rups uh zkm or

plasma uh ZK VM Tao or zking plasma or state proofs State proofs um Litecoin or Bitcoin cash Litecoin everyone says Litecoin Satoshi or Craig right whole that's a trick question there's they're the same person oh yeah oh yeah very funny as I said whole that's it ZK oracles or chain link lto or any other uh staking node

operator H right what are you talking about uh which which risk is is is more important centralization or um or bad security um B security is uh is worse that I don't know how to put that but I was trying to get at a trade-off uh speed or or decentralization no I wouldn't speed is better because speed guarantees use cases decentralization yes it's censorship resistance but it's censorship resistant but if nobody needs that what's the point yeah so uh ux or decentralization ux ux or

speed kind of go hand in hand speed he he likes to go fast this SC did do you have a fast car I don't have a car there we go typical I appreciate you sitting down with us and having some fun explaining some technical Concepts and a pretty easy to understand way well at least for us and hopefully our our community that has been watching and and following along um we also have have a ZK Academy that we've been working on uh nine videos kind of taking you from Zero to Hero if you will in the ZK space so I definitely check that out and check out nil and yeah thank thank you yeah thanks thanks for Rising cheers to the Moon