DeFi by Design EP129: Everything You Need to Know About Zircuit zkEVM

Transcript

welcome back to the rollup High Roller Community as always nothing on the rollup is to be construed 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 welcome to D5 by Design we talk all things blockchain and cryptocurrency while striving to educate Empower and enrich what's going on guys welcome back to the D5 by Design podcast here today with Rob of course and a good friend Jan

whom I introduced at Epic eth infray in istan buul this past year talking about zerit some really interesting H rollup designs so gonna learn more about that today going to talk about the rollup space as a whole some interesting things um Happ happening there um so yeah why don't we just jump right into it and Jen GM welcome welcome would love to hear uh a bit about your yourself for our lovely Community um and perhaps how you got here yeah sounds good thanks for having me thanks for having me back actually I should say I've been on um a rollup podcast here before and um yeah about me uh I I'll give you a quick intro uh I

was a former academic to some extent I sort of did did some graduate degrees I was finishing my PhD and I saw that web 3 was sort of an interesting space to work in and so I I was fortunate to to get a job in Security in this space and I've been you know doing stuff ever since so for about five years I've been uh in the web 3 ecosystem largely as an auditor at at quantstamp but now uh I am working on a project called zerit zerit is uh a new rollup which you know as you said we'll talk about today and um you know I'm really excited to talk to you about that and I can you know definitely tell you more about the experience and and what it takes to build one and what we've

seen and just rollups in general because I got in the space when uh when plasma was hot and I guess now itala kind of wants to bring it back a little bit he's had some talks on it recently but uh things have transition to rollups mostly and and so that's what I'm focusing these days very cool yeah Andy and I both had the pleasure of meeting uh meeting you in person uh Annie and uh epic eth in for day I met you at uh misari Main in New York um and at the time I saw like this Insight report that you had put together about like all like every rollup design you could think of like ones that I had never even heard of um so kind of like you mentioned like

you kind of got in plasma was hot and now we're kind of like coming full circle so like Trace that Circle for us a little bit as far as like the preliminary research that you've done and kind of your foundational knowledge and then we'll we'll get to like the building blocks and and why and how you're building circuit on top of that yeah absolutely so I mean I think you know starting from the the core principles right plasma was sort of this nice idea that evolved alongside the Bitcoin lightning Network and other early scaling solutions for for blockchains and the idea was you know let's somehow inherit some security from an underlying blockchain but take some transactions off of the chain make sure that people don't have to sort of always

look at every transaction uh there was this sort of nice notion in plasma I think it was plasma cash where you could sort of only watch your money which is sort of more akin to like real world situations right you're sort of you're concerned what's going on in your bank account but if someone else commits fraud you're not really watching that um I mean if you're the police you are if you're some really some Authority or you have a really big stake in someone else's funds you might be interested but generally speaking you know if you trust that your money is is is um operating in the correct way under the right assumptions that's sort of good enough for you because then other people can trust you and you can also watch the specific accounts you're going to

interact with and through this plasma design um which you know I I haven't followed for for quite some time now but there were there were dozens of them I remember there was this giant sheet of like 50 or 60 plasmas different variants everyone wanted to do their own little trick it was you know watch your own money or watch only some of your money or uh you know a different way to to to challenge if something went wrong or something that could I don't know change how the data is started there were tons of variations but throughout all of that one of the big issues was sort of data availability I think and so people said well why don't we just require things to be published and that basically how the

design for a rollup came along the general design not any specific rollup which is you know let's publish a bunch of data do stuff offchain and put it on chain and then just check when it's on chain that some conditions are met to make it final either a fraud proof or a zero knowledge proof and that was really cool because as that was happening you know the cryptography was catching up as well the the strict uh mathematical stuff that sort of needed to happen to remove the trust from all of those actors sort of rather than you know trusting people to to watch their own money we could just say well actually they're going to do it right because of the math or you know some very incredibly small probability that it go wrong and um that obviously

allowed rollups to you know become really really powerful and I think now the idea is you could put that back into plasma if you really cared about it to to sort of get some some different changes but the design Simplicity I say Simplicity but it is quite complex still rups is probably not going to go away in in the short term or long term really because um some of these networks are already out there right there's a ton of rollups already developed with main Nets uh with good test Nets with with new ideas uh with a whole bunch of things and um they'll become entrenched of course and people will keep using them regardless of if there's a slightly better solution and when you're talking about like pennies for transaction fees

like it's got to be really good for you to to to Really Care About the extra couple of cents uh or you have to have a lot going on and the average user probably and maybe I'll I'll get Flack on this on Twitter or something I know I don't imagine you know six cents versus seven cents in a transaction fee is really you know going to be the game breaker going forward certainly going from what whatever ethereum is now to to those nickels and and pennies that's that will be huge and so of course the rollups will win out um yeah so uh that's basically you know how of fundamentals happened and then personally I was looking at all those things right like I said I was trying to figure out you know how are these plasmas working how are these different

rups going to work and at some point um sort of started getting my my hands dirty and started building parts of the system or looking at specific research questions um things like security things like what data is actually you know being posted or or needs to be posted can we can we change that a bit because that could affect the cost it could affect the trust assumption where do the data go Etc um other things like Bridges U you know big big interest in in bridges fortunately there's been a lot of money lost in in bridges and uh you know rollups depending on who you ask again uh or just Bridges plus another blockchain some people argue that too um but certainly a bridg is involved and so

having these systems be secure is um is very important especially because of the history that Bridges have had which is which is not amazing it's pretty good but you know last year was like2 billion dollar lost and I gu I said this last time too there's a lot of money lost in bridg so I'm really worried about you know what rollups will will have happen to them hopefully nothing I really hope nothing happens but let's be realistic you know someone's probably gonna screw up some yeah so this perspective comes from I guess years of being in the trenches with regards to being in the the development Community as well as just in the broader crypto Community what what parts of this kind of and you know theal put the uh rollup

Centric road map out I'm curious what parts of this of this current uh existence that we are in for rollups did you you know kind of predict or see coming and what parts have kind of uh surprised you um you know you thought maybe things would have done a little bit differently yeah um I don't know if I predicted much of it uh I'll be honest when I join the space with all the plasmas I was playing catch for a while and I think when the shift happened it was pretty clear that actually this is a simpler design so I think the prediction that it would take off very quickly um was a reasonable one and I probably had it uh I didn't predict perhaps I don't know famously I don't

think I ever said it on record that that ZK rollups would be so quick I remember thinking like what would it take to build R like it's going to take years of effort and you know now that I'm working at circuit it's I'm doing it so like you know it's it's changed definitely definitely changed um but I I definitely thought it would take a lot longer it's still going to take some time to get them right to have everyone's sort of Niche um applications or or Innovations put in but um it makes sense to me that there are a number of of of role UPS out there I think I did say that somewhere on a podcast uh there's going to be more than one you know that that will be out there because if we actually want to get regular internet scale activity on a

blockchain we need more block space and one rollup will help with that but 20 will really help with that and there's probably a limit to how many people can sort of swallow I don't think you want a thousand rollups um but a small number of tens seems pretty plausible because you could have a chain that's you know very good at at gaming or something or you could have very Niche new OP codes introduced in some places that make um some apps work really well but scare some developers because you know it allows different languages and they're not familiar or you know they they don't like that for whatever reason so um I think the prediction that there'd be many of them is one that was totally fair uh the surprises were definitely

how quick some of the technical innovations came um even the the ZK stuff you know we we've ramped up on quite a bit over the last couple of years it sort of went from oh yeah this could happen like when plasma was still stalking I was like yeah if we ever get zero zero knowledge I think Stark net was around they were one of the the longest running players in the field and I think it was like oh if they can get it to work it' be so cool but not only did they seem to do that so did like a bunch of other people and that was a bit surprising because I I didn't think people would sit down and do the work for the first little while uh but I'm sure glad they did because now it simplified you know air quotes simplified uh a bunch of other

development um what I think the next predictions will be which isn't really your question uh but I'll I'll take a chance to De yeah I was gonna ask what what are the challenges and like what you know what's the next steps here yeah I think the next steps will be really trying to differentiate um features and functionality and rollups going forward U because where I think people are starting to see is um evm compatibility is definitely a really good first step and I would suggest most rollups follow that path um because you can get smart contracts very quickly at all of this but then when you don't have to rely on everyone in the world to agree with your virtual machine you just get your core

user base which is you know hopefully substantial and and well defined you can do really cool things that you can add other functionality that other people either don't want or don't care about um and that could be a new OP code so that you can do something you know like a different cryptographic primitive could be done on chain built in and then you support a whole bunch of new you know signing algorithms or whatever you want or or it can be you know ways to share data so you have different interoperability or you can do a whole bunch of other things so I think the next prediction is we'll start to see everyone really Master the evm compatibility very quickly over the next year maybe year and a half maybe two

years and then we'll see cool unique Innovation that really leaves just being compatible behind right something else that that will help differentiate each roll up um but also add core functionality so like not differentiating just for a marketing point of view but literally to do something different and unheard of right like and that could be I don't know baking in Bitcoin or something I I don't know and I'm actually hearing things about Bitcoin layer 2os these days so maybe that's you know a very easy pred to me oh yeah they are coming and they're coming in hot right now Rob have you been seeing much Ian from espresso has been telling me so much like yo man

you gotta pay attention you guys gotta do a series about Bitcoin l2s yeah I I think I was on a call um I don't know much about them to be honest I have to have to research can you tell me a little bit more about like are you excited about that do do you see I'm excited because and I don't want to get too deep into it and kind of uh uh sty away from ethereum the promised land but what's exciting to me is that the bitcoin's got this proof of Works security problem wherein if there's not enough transaction fees generated if bitcoin's price isn't high enough in the future then bitcoin's kind of screwed but no with this we bring act activity on chain fees are generated Bitcoin security is much more uh sustainable uh as a model without necessarily needing

the price to go berserk that to me is like like sure I will let the DG dgen on all the chains and all the nfts and Ponzi it up but if we can just contribute to bitcoin security for long-term Health the broader ecosystem that's kind of is like really prominent for me currently I think it's very early stages but you know we're working with a project cordal they're very cool um there's a couple Bitcoin l2s like uh botonics uh orifi um and of course different NXT so yeah very interesting things here cool yeah I mean what's good for Bitcoin is probably still good for the ecosystem right like for better or worse a lot of people still only think of Bitcoin as web 3 and I think we should we're trying to change that but

uh this will help I hope so yeah I'll have to look into it too and take a look at when when zerit compatibility first when Bitcoin l2s in zerit so what is zerit so yeah let's let's open this can of worms here sounds good yeah so circuit's a new L2 that uh I'm helping to build and uh I'm technical lead there and it's a it's a evm equivalent zero knowledge rollup with a focus on security so to start just as I said you know we're going to be evm equivalent and and sort of meet that that bar but then we we want to do more for the ecosystem in one way we're doing that other than a whole bunch of cool research that honestly probably isn't

super sexy to like 90% of people like we have unique MSM optimizations in our GPU stuff which is cool but like most people won't notice the difference there right it'll shave the pro time a little bit but uh you know it's not going to be mindblowing there what instead we want to do is focus on another area of the ecosystem that we really care about which is um security so everything we do at rese at um at zerit in terms of research is to try to help produce a more secure system because as I said you know hacks have happened on Bridges and other systems of course in this space but every time that happens you know the media picks it up and everyone goes crazy like oh web 3 is over it's another

hack we can't survive this one possibly um we have and I can think we will continue to survive a bunch of them but if we can do something to prevent them from happening in the in the first place you know I think it's sort of our duty to at least try so one way that we could do this for example is you know had add some cool functionality into circuit where we check some properties of transactions and if they're not particularly helpful or they violate some invariance um we can slow them down and make maybe notify people so you know if you deploy a defi app and there's a transaction in the me pool that's going to wipe out your funds and it's not coming from you you know maybe you you think that's a hack and so there should

be a delay um maybe that's not the exact approach we're going to take with this system but the idea of of drawing in these type of circuit breakers in defi into um the sequencer of of a rollup for example sounds really interesting and it's something we want to explore and first and foremost um circuit as a research based project like it will go live there will be a main net uh it will be a functioning network but we want to try to do stuff that is different we our end goal is not just ebm compatibility we want to see what we can really get out there so it's it's a bunch of things but the short answer is is it's going to be security so good talk about the base contestable

rollup structure and with regards to ZK but you go ahead shoot your shot buddy we did we did just have lost my thought we did just have a conversation with the uh the CEO of Tao um working on the ZK rollup and uh he was he was educating us on um based contestable rollups the based part not super relevant to this conversation but the contestable part um does have some uh some relevance here because as you say like the circuit breaker idea where if there is something that seems like it's malicious or maybe an attack on the

network then anyone can dispute or contest this transaction very similar in this case it sounds like you're you're you know suggesting that uh there's like a there's like a pause feature like a circuit breaker feature while that contest dispute resolution process is happening is that yeah I that sounds pretty familiar I'm actually not familiar with the the contested notion I've I've seen other stuff from tyo the boosted rollups for example I think is a really cool idea uh this is I think one of those things where I said you know people are deviating from just evm compatibility right they're adding out cutes to have possible really cool layer threes or layer twos or wherever they're putting that output um but this

contestable idea you know could be very much related I'll have to take a look to to be sure that I I know exactly what's going on there but yeah it'll be similar I guess in the sense that you know it will slow things down if if it's particularly problematic and if we can prevent just one hack that way I think that's a really good draw right yeah and obviously we have to be careful about how that's defined and and what it means to contest or you know quarantine a transaction which is the terminology we're going to use um because you don't want to censor people and you don't want the the good transactions to go through and then probably there will be people who argue code is Law and there are no bad transactions but if your transaction you know wipes out millions of dollars from

the average user uh right I'm not sure Cod is law is the right defense for that so uh I'm willing you know to at least explore the opportunity and see see what we can do and maybe maybe it's very very effective and maybe we can use it in other ways right once we have the system to detect invariance and do things with the men poool we can do other things too so and that kind of seems like the general sentiment of ZK rollups at least the responsible ones going live is like the validity proofs need time to mature in terms of being able to be let loose into the wild and so there's got to be some sort of mechanism incentive design

on the most decentralized and and uh well-designed manner to allow for these kind of uh quarantine periods fallbacks contestations um and that's just a matter of building responsible Tech and letting the Technologies uh kind of mature now on the flip side of this do we really need ZK proofs for every single transaction on zerit on insert ZK roll up here like is there a world where there's a kind of a hybrid mechanism that that zerit adopts or how do you guys go go about applications which are perhaps not as financialized or not as much value at stake for smaller quicker transactions you know do you need a ZK proof for all

of those how are you thinking about the rollup landscape when it comes to the necessity of having a ZK proof for every single transaction um and kind of how you're looking at that okay yeah that's that's an interesting question uh I I don't have a a solid you know yes or no answer it wasn't yes or no question but like I don't know where I I really land on on which side of the fence my gut says you know ZK should be everywhere um because we want to always be able to trust the um computation whatever that computation is that's sort of the whole point but some people like metallic have always said there are some things you're willing to lose right like I think he he

had a blog post recently where you know you could say ah if if you're just losing sort of your Social Network status or a like right if it goes down you can't recover that data or someone really was fraudulent about it it's it's maybe not the end of the world um but if you lost millions of dollars in a transaction which you know I'm not doing but if other people did that right uh that would be really really bad um right I you know I think that is a good notion of like where where do we need full security versus sort of partial security and and how much we invest into the tech that security systems but I I think the gut says let's do everything on ZK at

least for now is reasonable because these systems are actually you know becoming very feasible and very easy the the sort of black magic the hardness is being you know made understandable and and implemented in ways that it's almost more effort to change um some computational models on a system so some sort of hybrid where maybe you only ZK prove some things um is is a worthwhile exploration and I think um one that people should undertake I don't think we've really thought too much about it from a practical point of view at circuit but uh it would be very interesting uh on

the other hand you know if everything works in ZK let's just do everything ZK ZK also has some nice benefits that you can do you know like co-processor sort of natively embedded if you wanted right so you could you could have offchain apps running ZK systems that are different ZK proofs and then just verifying them on chain and you could possibly tie that closely together if they're using the sort of the same ZK system um and I think that's really cool because then you stop even executing onchain you start just verifying onchain and it sort of Skips ahead on the on the road mapap for ethereum um to be clear you know that Tech would have to be developed in a meaningful way and and the specific applications obviously need to be built but when ZK is so

attainable for for now on like let's put everything in a CK chain because we can't and you know that probably means there's you know going to be a long-term victory for things like zero knowledge rollups over optimistic ones even though optimistic rollups right now have a lot lot of the user um space uh I could be very much wrong about that and you know there's a world where where the optimistic rollups actually play that lesser secure hybrid model that you were just talking about right you could imagine a world where a blockchain that tracks your gaming state for I don't

know World of Warcraft or something isn't an optimistic rollup because maybe you're not so concerned if you really lose your player stats right and it's very cheap maybe it's even cheaper I think z um zirc and all the ZK rollups will actually close the gap on the price quite a bit so that it's it's not quite noticeable as I mentioned or there'll be extra value to make up for the difference if there is a difference um but this world where where optimistic rollups can play that type of part be very very interesting and it actually be very easy because most of them are are are pretty close to mature if not mature right maybe they're just missing a couple fraud proof testings or or extra clients for diversity but they're actually you know they're they're

probably more mature than most of the the ZK proof systems and and you're right we do need the maturity to happen for these these EK systems so um for now it definitely Mak makes sense and they have a good place so Zer is all in ZK Z no optimistic none of that no no I don't think we have any plans full on full on yeah on the test net you know there's some some blocks that aren't proved right now just because like we don't want to spend tens of millions of dollars running prover on test n blocks but in in main Net Zero knowledge all the way it's gonna be you know every has a CK transition that's you guys are optimizing for security um got it okay and this is something

that like I've heard about like ZK and it's like still somewhat of a mystery to me maybe it is to other people too that's this concept of a circuit which like exists outside of ZK proofs but like when we're talk I bring this up mostly because the name here is zerit there's obviously circuits involved and I've heard them related to ZK Tech so like could you kind of like Dem ify this concept of a circuit and how it relates to zero knowledge proofs oh I can try I definitely you know IED to give some talks on this in the past and I feel like I never nailed the concise argument but I can I can definitely give you a high level view that people can pick aart and say actually that's wrong but I can I can try so it's basically the

notion that you're you're taking an arbitrary computation which in this case updating a blockchain state and you're you're making it into a a virtual circuit which you then sort of virtually execute right in the same way that you can take a computation and put it on a hardware circuit a real circuit like the laptop you're running right now um and it'll do something all the time uh you can you can make those circuits virtual and say you know it's it's not a a new stream of electricity per se going through it but a signal and the signal is virtual just some bits and as you do that you know you add some extra stuff to make sure that the only way you could have computed the circuit simulated it run it effectively um is if you had the right inputs and so

you get this computation which is the same as the computation you wanted but in a system where the circuits only were simulated correctly if you knew what you were doing you didn't lie and the circuits were built correctly of course I'm gonna I'm gonna I'm gonna try to refine what you've said up until this point just to make sure I like clear for my own understanding which is that it's it sounds like uh circuit is like a series of State machines that commun with one another and execute State changes is that and and like the most simple explanation of that being like like this like a configuration of State machines and like I'll simplify it even further and just say a state machine is

a ledger right like a a the state is a ledger it's a certain certain balance in all these Accounts at any given time so a circuit is basically a configuration of ledgers yeah I mean I think that that's reasonably fair I think there's some technical stuff the the academic and me is sort of saying we should go back to the precise definitions but I you know it's not a lecture so I'll avoid that but yeah the state machine analogy is probably pretty reasonable just think of the the circuit as like it's even less than than that in some way it's it's it's it's a gate if you know what a gate is right like it's if two signals are on and it's an and gate then the output's also on but if either one of them is off or both are off it's

it's off so it's a it's a collection of these very very small state machines if you will um that eventually go up to something big and and that's also why you hear like these big numbers where you say oh I have many many gates in my circuit and that's why my approver is so slow for example uh it's because if your gate is so small and you want to do such a complex computation you need a lot of gates right in the same way that your current CPU on your your your laptop you know evolved from just that and gate I was talking about you know in the 40s too A whole bunch of of of gates connected to make you have you know the ability to to to do whatever you're doing so uh yeah I think that's fair

um and of course you know most of the time you can abstract that away and and you can sort of use a high level language to actually write the circuit and you don't have to worry about the and Gates specifically it's generated automatically thanks to a lot of pioneering research by a lot of people um but yeah that's that's very cool okay so um like ones and zeros here would be the signals right like if I one is a signal zero is a signal if it's an and gate and it's one in one then then the output is one okay I think I'm I'm catching I'm catching your drift okay so then like uh bring this a little bit more to like current times uh

there was a prover Network that went live a couple days ago called succinct yeah so like like that just kind of like inspired some ideas in my mind it's like if I was to compete in this prover Network frankly I have no clue how to run a prover but I'm starting to learn so if I was to try to build the quickest prover and kind of like win these proving competitions and and earn money from doing so I would essentially try to build the prover with the least amount of gates because that would be the quickest proof is that I I think um I don't know if that's entirely accurate so generally you're probably wanting to prove the same thing right you're trying to prove a specific circuit was executed correctly um so I

don't know that you have too much flexibility in the gates you can change on the circuit because you know you're trying to prove the same thing uh which is a state transition for example but you could change how that is simulated and that's sort of the key I think bit here so when you simulate it you know you have to compute a whole bunch of outputs of of these you know Gates per se uh or something else like polinomial and things like this and through that you have different operations that you could possibly do faster and better right so if you had for example um different ways to compute uh multiscalar exponentiation in over some finite field in whatever that means it's a mathematical operation which is like an

ad but on crack somewhere right like it's it's much just uh if you could do that faster then you're still accomplishing exactly the circuit does so you're not changing um what the gates are doing but you're changing how the gates are being simulated then you're starting to win on on on that and and maybe there are other optimizations like seeing some circuits are not always used so you can discard them or things like this but I would bet you know most of it will be these other types of optimizations where you can either parallelize it or optimize operations or throw better Hardware at it those are the probably the three most naive approaches and maybe you can do much more maybe maybe you can do very different things um that's where I would start

so you know maybe but I would say you know optimizing this is also why you see program uh people saying oh we have really cool um MSM Tech or fft Tech or various operations of ref finite fields that become faster because you're like oh that actually translates into better Pro time and better Pro time you know translates to better costs because running so so much and so zerit is like can you know we started out this conversation by saying like you know all these rollups are starting to pop up like some of them are good some differentiation and and variation is very healthy at a certain point do we need like an infinite amount of rollups

probably not so there's some happy medium there where there's enough variation to to provide something new and useful and different um what is zerit doing like what's special about the configuration of of zerit circuits what's special there that makes you guys new and different and useful yeah I mean somebody say zerit circuits is Circus yeah it's it's a great uh it's a great phrase I've definitely used that before it's a I think I said circuit circuit circus sometimes when I'm looking at at the code um there's there's some stuff that is definitely different um you know we started by looking at what was out there of course because people had

started to to do some of this in open source it but we quickly realized we wanted to change sort of like how the proof pipeline Works um so you can change which parts are are are proved where you can change how the proofs are then aggregated you can change you know the layers of proofs so that you sort of compression is not perhaps the the correct word but you can sort of aggregate a bunch of proofs so you sort of prove several transactions for example in parallel and then take the the aggregation of that of those proofs so now you have a single proof that represents all all of those proofs and you can just keep doing that recursively to a certain point that is useful um and you know then you just get ultimately one proof at the end of the

day uh so what we did partly was change how that aggregation works so we have our custom layers of of aggregation for some nice reasons I can't share too many of the details in part because I'm not working at that level of the code and in part because we haven't open sourced the code just yet uh so I don't want to give anything away but we've also focused on that other thing which is sort of improving these optimizations for specific op operations so uh we have custom MSM implementations and ideas for implementations which are being implemented that will change how the MSM is computed uh not what it does right fundamentally it'll be you know the same

inputs and outputs as anyone else might be doing but we'll have a certain set of unique optimizations which is pretty cool in the past these optimizations you know are things you could patent uh I don't think we're going to I don't know if we're going to patent them I depends how effective they are and probably going to open source them so I don't know what it'll mean to have a patent um but we we are looking at at that type of optimization where you can get some pretty substantial wins on the lowest level Primitives that make sense to sort of optimize and deal with and this is part of the reason we started you know actually saying oh we should have our own it's because we had an idea and we didn't want to just you know go give it to someone else we wanted to see

if we could do it on our on our own and um so yeah the the circuit pipeline is definitely different the fundamentals of some of the Primitives are also a little bit different but otherwise the circuit know Faithfully implements or we're being testing at least uh all the off codes on the evm maybe not could you unpack those acronyms just in case someone wants to like take a deeper look like in this in this context what what does MSM mean and the other fft I think yeah uh MSM is multiscalar exponentiation so it's a it's a primitive where you I guess multiscalar exponentiate it's a finite field

operation uh I don't know that I can Define it off top of my head particularly well um the other one is uh fft which is a fast 4A transform so this is an operation that deals with polinomial and does something which we could also talk for hours about but oh yeah yeah two years of calculus to really understand what what we're gonna leave we're gonna leave any and all 4A analysis out of this conversation I mean Rob loves math but I think this might even be over his head I mean I took I mean 4A analysis is a doozy uh that's for sure so I think I think uh dazy yeah so leave it we're gonna we're

gonna drop on this DK circuit coding train 30-day boot camp build circuit's best app get ready yeah what's the uh what does that process look like Yan what's the two things on zerus ecosystem what is the developer tooling infra um and how is the ecosystem looking like how are you guys planning to build that out and then also how are you guys going to be able to talk to other rollups and interoperate yeah so uh I can answer almost all of those um I think the only thing we haven't really considered is is explicit external interoperability with other rollups it's not that we don't want to do it we just haven't gotten there yet on the sort of building out

the core infrastructure um but to answer the questions about developer experience it's pretty straightforward you change an endpoint on your you know I would say hard hat but I guess foundy is the thing right now that everyone uses or metamask or whatever your favorite wallet is and it it it just works um there might be an the occasional exception if they're not integrated correctly but uh more or less you change the endpoint and it's just like you're talking to ethereum and everything can be deployed exactly as it can on the Chain so very minimal overhead um there uh the test net is is currently down because of the update from spolia but it'll be up in a in a day or two I would expect um but you can try it out for yourself everything there

will work just like it does in ethereum and if it doesn't please let us know because that's what test Nets are for um but everything has gone smooth so far for us um so the development experience pretty straightforward pretty easy we're building out the rest of the tooling that that sort of necessary to complement the sort of actual development so we have a block Explorer obviously we have you know all the stuff you sort of want um we do have a native bridge to get stuff on and off the chain and whatever we don't have we're building it and this brings me to the next part of of Partnerships and and uh these types of Integrations um and we're talking to a bunch of people so we're talking to a whole whole bunch of projects so that when we go live on a main net where it really makes sense um

we're going to have a bunch of apps on there you know with all the the stuff that you sort of want on a chain and um admittedly that's not really my departments I don't know who they are but you know they they're the big names I've heard of them I just I don't know where where every project is on on the the sort of scale of of ready to go or you know still preparing for it um so the the short answer is you know we're talking to people for sure we have lots of interest and if we haven't talked to you and you are interested please reach out it's not necessarily my department but I can point you to the person who can help you out with that so um yeah it's going to be a seamless ethereum like experience that will encapture

everything and and what we you know what we're really interested in is some of those Partners who want to take advantage of the security focused features right so if if you in the past you know we we're reluctant about some some smart contract security and you think that detecting a variant at a transaction before the transaction hits is something that's very interesting to you like come talk to us so that as we're building out that functionality we can you know have that in mind but if you're not interested in that either it'll just work like a the and it'll be smooth and it won't take you very long to to change your endpoint you have 30 seconds you're done according to Rob doozy so I'm sure you're gonna have tons of uh

here since we can use code ZK apps now because know the rollup is evm compatible you're not building anything directly ZK right when we open the code if you want to look at our code and suggest improvements cool that's a whole another can of worms that's that will be complicated uh scary and you know welcome would love to have that feedback but if you're just building a Dap developer the ZK Almost Doesn't Matter right it's just a chain with fast finality we have we have no privacy so you don't have to worry about like shielding transactions and hiding it um that's currently not on on the road map right now we're getting just faster finality than than other rollups and cheaper than ethereum that's the goal so

are you guys GNA post the the data to ethereum right yeah yeah for now that's the plan um we're certainly thinking at other da Solutions um but with 4844 4844 IP 4844 uh actually really close closer than than sort of we when we started this project uh we're going to at least try that see if that works as well as we want it to be the first time uh I'm not sure that you know monolithic is the end game or the desired thing of everything but certainly it's going to be somewhat simple to develop on the test net you know obviously we're doing that without um 4844 um blobs right now but we'll we'll explore that option we'll build it out and then we'll see

and you know maybe there will be other approaches we can take either you know you can switch where your your data goes or maybe we'll look at other da solutions that end up being cheaper or trustworthy but we definitely want to make sure that security is a focus and so to start the fewest trust assumptions comes with everything on a theum and when we start to relax those we want to do those do that for for the right reason and with the right consideration we're we're going to move faster than ethereum but we're not just gonna overnight be like ah let's switch to different da because that makes Zer it a true rollup right not a volition or a validium or or optimium yeah it'll be a strict ZK rollup um you know I think the idea of a

volition is very attractive uh I I don't want to promise it because I don't know what the engineering overhead will look like and and I want to get to Main net before we we I promise other things um but yes it'll be a full ZK roll up so from the Builder perspective like a Founders perspective technical you have your sip of water if you need there sir um the key consideration for for you from switching from away from ethereum's data uh storage data publishing data availability however you want to call it um is strictly with regards to security and kind of like decentralization that's kind of like the key thought as like okay maybe we won't

do it on Celestia or Avil or igen or you know insert your da here because of these reasons or yeah I mean what go through your mind we want to make sure we understand those reasons and and what we are trading off from that um you know I I don't think any of those projects are bad let me be very clear I think they all have their place and I honestly think we'll we'll consider each and every one of them at some point right to to some extent and um we may very well go go down that R but for now the simplest most straightforward approach is is is everything on ethereum and until we understand exactly what the tradeoffs are or how they fit into the road map or you know what our priorities

are right like for example it would be really nice to also have a decentralized sequer we don't have one right now um like most rollups we don't have one right now uh but which is more important decentralized sequencer or a decentralized data availability layer I don't know that most people have a strong answer to that and it's actually not something I've given too much left to just yet either because we don't we don't have a main net once I have a main net I can start talking about what I really want you know the dream rollup to look like um I have and I my dream rollup might consist of both of those options um but there are more practical aspects and that's what we're focusing

on at the moment so it's not to say we won't ever be with a different da provider or solution um you know this anything could happen um maybe this is too much of a cop up but we could we could switch Layer ones if we really really thought that was a really good use case I don't think we gonna do that but you know it's it's a research based project we'll go where we think the the best results for the end users using the chain you know will will go and want results from because that's y do when M that sir hopefully this year that's definitely the plan um I don't want to commit to any tighter deadline

than that because when you commit to a deadline people start getting upset when you miss the deadline so um test Net's been up for a couple of months now um other than the small change dis aolia that happened um recently it's it's gone pretty smooth there's been no issues so um we're chugging along we just got to finish cleaning up some of the internals and we open source it and get some security you know reviews or take a look at it from other people and um and then hopefully we can get it out this year as you know as soon as possible without risking security considerations where we don't want is to release a product that um anyone is scared to use or that we are scared to

have users on because we haven't tested something or we haven't considered something and this is also partly why you know to go back to the DA like maybe and and if we haven't had the time to do that I'm not committing to to use it is there like a like like also on this on this uh security train is there like an idea to kind of like like roll out main net in stages almost like aan layers doing with their caps like only allow a certain amount of tvl at any given time oh yeah that's actually something we've been talking about um we don't know exactly what those stages would look like but uh you know certainly what we're not going to be doing is launching the

network networ where we're sort of a stage three or two from vitalic you know stages roll up uh where we have a fully functioning thing where we just hands off and like ah we're good we're done um it's it's going to be some some level of stages where we can possibly intervene in in case something goes wrong because you know what would keep me up at night is is someone putting on a bunch of money something going down and we can't recover those funds for the person um ideally that never happens I I want to stress you know that it's not the goal at all but uh there definitely will be stages so that we can prevent that just like most the other rollups there'll be some level of centralization to begin with and you know as long as all the tech works out we'll become more

decentralized as um as we can um but even the centralization you know it's going to be some level of of baby steps because as mentioned already on the call the ZK Tech needs you know time to mature and while you know we think we're pretty confident about what's going on it doesn't make sense to just say here it is good luck right that's that's that's reckless so yeah we'll probably do stages I don't know exactly how long each stage will be or when the stages will be um completed or whatever the reality is the role it will be under development for a significant period of time still there's there's

stuff to be done so we have to keep doing it but maybe we can get to a point very soon where we're very confident that actually you can use this for all of these features but maybe not this extra feature or something like this go take a look go test it and test it even better than just a test net just like actually use it and and if something really does go wrong which we don't think it'll happen we'll be there to help and support got it okay I think we've come far enough and I think it's time to to un unlock this this this uh onchain elephant in the room which is that if you look at certain block explorers for certain rollups that won't

be named you'll notice that if you if you look at the eth contract address on on such rollups you have way too much eth you have something on on on the order of like trillions of eth like don't go check in the the eth market cap it's not trillions of dollars some of the rollups have trillions of eth just sitting in a contract like like just garnered bu a multisig or like some upgradeable contract so again without like calling maybe all right if you're curious as to like which networks are doing this please just DM us or DM DM Jan Yan and we'll kind of like dig

through it a little bit deeper uh because we're not trying to publicly shame anyone right now but maybe later so tell us like why certain rollups have more eth on their rollup than the maximum supply of eth on mainnet and yeah why they think that's that's a good thing or or why they're doing it at all yeah so I mean I not having worked at at you know these rollups I can't comment always 100% but uh I'll try to try to you know hypothesize as to why and I also don't want to shame any of these rollups which I'm not going to name um because I think it's just a design

decision right part of it is uh this particular rollup did it to simplify um how deposits happen on the L2 right instead of creating it out of sort of thin air all the time they did it once the beginning of the their system and every time you deposit just sort of takes from that fund and that that is a very interesting idea and it it's stems partly because of how that these circuits work I think uh where you know the circuit always says oh well the money must have come from somewhere so where did it come from and the only you know the easiest way maybe not the only way but the easiest way to sort of to handle that is say well pre fund the Genesis block somewhere so that the circuit always says oh yeah just comes

from there it's fine um but it's exactly these types of decisions that helped generate uh like zerit as a project because we saw things like that like is that the only solution can we do something different uh and I'm not sure if we'll we'll take exactly the same approach um our current internal plan is to do something different um it might be a little similar but it you know doesn't sound like you know I don't want to put two trillion eth on a on anywhere or I mean I would love to but if I only if I earned it correctly and I I certainly didn't and it won't um but this is this is part of why zir sort of existed was to see you know can we do something

differently because when I look at that that block Explorer and I see that my first instinct even as an experienced feature like why why is this the way it is and I I actually have no doubt that there there was a design decision made that said yeah this is the easiest way you know let's do it but maybe the easiest way is not the most convenient the most confidence inspiring because um certainly if that is behind you know credible contract it's it's kind of scary that that something could change there um so these design decisions changing and and those those rollups making that specific one is part of the fun and research of

this area um and you know I'll say it's it's it's interesting that it was done this way um like I said maybe circuit will have to go that route for mayet I I don't think that's the current plan um but we're going to try to figure out some other way if we absolutely can and it's these types of research questions that will always make rollups even rollups which are all claiming to be evm compatible and and sort of air quotes the same which they're not uh a little bit different right there there's subtle differences across all of these things and and unfortunately those differences might incur different risks I don't think that that particular issue is too

risky um certainly if there's upgradeable contract and the keys are not kept in some way that's meaningful that maybe is a risk but I don't think it's particularly like shameful in the end of the day if it works and it's secure enough whatever enough is yeah the the issue is it works until it doesn't right like yeah yeah you have a pile of money sitting in there and it's like yeah like we only go in there when we need some money it's like all right but like what happens if someone gets in there and they just yeah so like I really hope these chains are taking this desci decision you know do everything correctly in every other area

and and I have no reason to suggest that this isn't the case but a design that sort of asks this question in the first place to me doesn't sound like the best design and and then and maybe it is but then I I feel like I should have heard more about it places like don't sweep it under the rug be very open about ah this is this is a point of risk that we unfortunately had to introduce and here's why we're doing right that would also be very nice to have yeah which might be out there to be totally fair maybe I've just missed it on some of these rups um but it without going like super without opening up like another can of worms here like could you try to make it

make sense like why are rollups in this position where they have to kind of like pre-fund the Genesis block with a certain amount of eth we have Bridges right why can't we just take the eth from the main chain yeah so I think I think I can comment on on on on why this is and and that's it comes down to how the circuits are implemented so at the end of the day to prove you know that you took money from somewhere to to somewhere else legitimately you need a zero knowledge proof that says Ah Yeah funds exist Ed um wherever in a and you took it to B um and and you can if if you if you're taking from this

pre-funded account for example that's very easy for the circuit to see oh yeah there were there was money there there's nothing invalid about this transaction as long as the transaction is signed and it only takes the amount that's specified in the transaction etc etc etc everything works um but if you tell the circuit hey I'm taking money from nowhere and I'm creating money from out of thin air like there's nothing fundamentally problematic in a computer science point of view from saying you know add something to nothing in a variable um but it sort of changes what people want to see in the circuit because now the circuit looks less like ethereum because you've added this sort

of exceptional functionality that means on a deposit you need to Mint stuff out of thin air which is typically you know only oberved whenever ethereum is normally minted which is you know during consensus on on the layer one and but you you've taken that functionality you've added it somewhere else so you could you could change the circuits to say oh yeah I'll just do that but now your circuits look different than say an evm or they could look different um and maybe they're more complex even if they don't look different because maybe uh you know you can make it look close enough but you still have to branch in some area and when you have to do that the pro time might go up and so maybe there's a trade-off that says oh I really want to have a fast prover so I

can't have this type of branching in this exceptional case in short it's a very subtle technical issue that's sort of at the bottom of of of sort of the stack um that is not trivial to overcome um it's not incredibly difficult they did it I think the other way of doing it you know by disabling say circuit signature checks you could also do I don't know that you want to do that because if you screw up a signature check fail uh like disabling and you do it too often or someone forges a signature or something someone screws up in another way you might have different risks um but I can also see other worlds right you can see you can do things

where you pref funded with a couple of these five 10 20 and then you top up that account regularly as part of sequencer fees and maybe that that is sufficient um maybe there there's another fault that happens if you if you go down that approach which is why I can't say for certain exactly why this has been done um but I'll tell you at circuit we're definitely exploring different options um something like that one I just said where maybe we'll prund just a little bit so that if you really didn't trust us and you did look at our Genesis account it might look like 20 e which is a lot but it's not two trillion years right it's not the end of the world it's just right shitty um It's A Hard problem yeah and it's these small subtle things that actually like

really add up to making these these systems complex and and it's therefore why some systems have design decisions that are I don't want to say questionable but like not well explained or not the way that your intuition would initially say let's do it this way um it's because there's there's a bunch of these these small little issues that arise from all the technical stuff of trying to get the evm in his Z knowledge proof system or into a different type of bullet like even optimism changes how um deposit transactions sort of look inside their gu right you can look at their L2 G and say they've changed stuff they have they have very good documentation on how they've done they're optimistic

so it's slightly different for them um but there's a lot of subtleties that make this type of project not nearly as simple as let's go Fork something else and call it our own chain like you you need to do a bunch of different Innovations and Technical bug fixes and and revisit some designs to get some you know progress as opposed to just trying to get something else to work yeah and with that comes tradeoffs exactly oh well this has been an absolute journey and uh an absolute pleasure of one

um yeah I mean if we keep if we keep going any longer I'm going to start bashing forier again so like all I got to say is it's a doozy and I'm building the next 100x on zerit follow me at a andd we're going to the Moon baby we're GNA put horse racing on zerit zero knowledge proof you'll know for sure that your horse won without knowing how much when where you just gotta log in Click a couple buttons forray analysis Rob's gonna code it all up going to the moon so see you guys on there with with your astronaut suits and your horses all right thanks in the Z cats yeah thanks yeah thank you John