Debating the Multichain End State: Service Providers vs Immutable Protocols | Bryan Pellegrino

Transcript

my friend Citadel who I have this custom relationship with and I'm just going to fill all of these priority orders with them they're going to sit at a higher priority of packet than any normal person coming from the source chain it doesn't matter what order they submitted to doesn't matter that they were later in the queue I'm going to fill them on the destination side because I'm more economically incentivized to do I think you're going to end up with a world that potentially is a worse Financial system than the world that we have today because of a lot of this and so I think if you're going to go down that route you should care deeply deeply deeply about the properties of the underlying messaging layer and I don't think enough people do right now hey everyone this episode is brought to you by say the Blazing fast

parallelized blockchain which is unlocking salonik performance for the vast ocean of e devs out there now you're going to be hearing all about say in their new V2 upgrade but if you take away one thing evm is here to stay there are some problems with it which we're going to get into later in the episode but say and especially their V2 upgrade is helping solve that so thank you very much say for making this episode possible hey everyone this episode is is brought to you by Unis swap delivering the best onchain trading experience in crypto period Bar None one thing I want to call out is the Unis swap extension so say goodbye these days of these annoying as sort popup wallet extensions you lose your place trading you have to open it back up unisoft now has a nice Sleek sidebar that persists no matter

where you are in the web it's much easier to use if you click the link at the bottom of this episode you can join the waiting list and I'll see what I can do to get you moved up that waiting list but definitely go click the link check it out all right everyone welcome back to another episode of bell curve before we jump in quick disclaimer the views expressed by my co-host today are their personal views and they do not represent the views of any organization with which the co-hosts are associated with uh nothing in the episode is construed or relied upon as Financial technical tax legal or other advice you know the deal now let's jump into the episode all right everyone welcome back to another episode of bell curve this is a very very special episode which has

been in the works for a long time uh very excited to introduce my co-host for this season heart and today we are joined by Brian pelo of layer zero Brian welcome to the show thanks so much happy to be here this gonna be great thanks for yeah thanks for setting this up mate it's gonna be great yeah um hey this is this is my pleasure and frankly I'm just going to do my best to almost stay out of your guys way I think in terms of uh this is the interoperability episode guys so if you've been following along with us for this entire season this is of course the multi-chain endgame season and what we've been asking ourselves right is trying to update our mental model in terms of how this uh these many different blockchains are going to speak to one another and obviously the core

concept there is interop so basically for listeners in terms of an agenda here what I want to do is kind of tease out from a high level uh Brian your perspective on what this multi-chain endgame is going to look like how are all these different um hundreds or thousands of or millions of chains going to ultimately end up speaking with another heart going to ask the same of you and we're going to try to tease apart uh the parts that I think where you guys agree on which despite you know some of the spirit interactions on Twitter I think there's a lot of agreement actually in both of your worldviews and then I also want to get into the parts where you disagree and maybe dive in a little bit deeper there so Brian why we give the floor to you first here just just to start off can

you sort of starting at a high level I mean where do you think we ultimately end up in terms of this world of hundreds or thousands of millions of chains and you know how are those chains ultimately going to end up speaking to one another yeah I mean I I again so so we try we try to stay pretty unopinionated in terms of what the end outcome will look like for us it doesn't matter if it's literally a thousand completely orthogonal layer ones uh or like ethereum only as a settlement layer in a ton of layer twos I think the goal for layer zero is to build uh the lowest level possible primitive of the highest possible utility right like that's the function that's the framing we use when we think about the world and I think

when you think about that you think about like how how should that thing actually be built so if these things you have a bunch of random State machines or like execution environments that exist uh how should they talk to each other what should that look like what properties does that transport layer need to have uh I think like those are the things we care and focus on and like very uh strongly opinionated in terms of like what properties that layer needs to have in order to like achieve kind of end state that I think all of us want um so I you know I does I don't think it matters what that is whether it's a thousand or a million or or literally just 10 I think you need the co-c commmunication no matter what yeah you

have this you have this really great analogy Bren I've heard you describe sort of the birth of the internet and comparing it to uh the development of crypto as it is today right there was kind of like in the beginning there's kind of you know little computer clusters whether that's at UC Berkeley and then there was arpanet and those things didn't really talk to one another and even in the early days of the internet right you had these sort of competing corporate intranets where these corporations were trying to put everyone on their same uh structure and I've heard you compare layer zero the protocol almost to sort of the the most open sort of permissionless uh thing that ended up winning which was TCP IP or TCP IP so can we get into that a little bit like talk a little bit about those that kind of like early days of

the internet how is that similar to where we are today and then like where would layer zero sort of fit within the sort of analog internet context yeah sure so yeah early days we had a bunch of intranets right which were just distributed execution environments you had DARPA you had Stanford you had all of these different clusters right and if you wanted to take some data and processed that you want to do some computation you literally take your floppy disc and you would like fly to Stanford and you'd pop it in and run it and then you'd have have your result back right and then eventually we invented this very nice stack the internet stack and allowed these things to talk to each other and that you know basically became the internet and the lowest like the primitive of the internet is effectively a packet it's

just like do some compute on chain a take those btes that are generated from that compute move the btes and then like do something with those btes on the other side layer zero is arbitrary contract invocation with a bites AR rates invoke a contract here and generate some btes move the btes and invoke a contract on the destination Chan with the bytes and so uh I think you know that is effectively a transort layer and so when you when you walk down like what properties do you want that to have that's that's where I think it actually becomes very interesting because I think a lot of the space was built on like you you like immutable permissionless censorship resem right like those are the three primary properties that I think most things in the space were built around and I think when you talk about like how you actually build that

layer and what it means and then we can walk through this uh throughout the episode because I think it's it's going to come up quite a bit there's so many problematic structures that you get into if you lose any any of those three properties and I think for us the goal is again like like that primitive is all that we want like it's layer zero it's not trying to be uh you know tcip is is a dumb protocol it's incredibly powerful layer zero is meant to be like reasonably dumb in the sense that it doesn't add a lot into an individual packet right anything that you can do on any blockchain any amount of bites that you can generate that can be processed by another blockchain so like the two blockchains are compatible in terms of like size of uh uh size of payload um

you can do over layer zero there's there's nothing added there's no restrictions there's like all of that is there and then I think it's really about that how you build those systems that matter so much um and I think for us it's it's really and again I think I think we should get into this but we really try to focus on pure infrastructure and I think a lot of other people are focusing on like becoming a service provider and like you talk about early internet like Microsoft I you know everybody had their own version uh the information super highway there are all these different privatized versions of of trying to basically make this uh and and the thing that won was the thing that was uh was was both uh most open and uh least least like a service provider let's say at least like

a privatized internet u i I want to hard turn it over to you in a second here but Brian could you say just a couple more words about that for folks in the audience when you say the difference between layer zero trying to provide infrastructure versus others that are trying to be service provider you dive into what that means a little bit more 100% like when we built layer zero there were three those three maintenance a mutable sensorship resist and permissionless right and so what that means is when layer Zer deploys an endpoint that code is entirely mutable we can never change it and most importantly we can never remove it once it lives on a system it will be there in perpetuity or until like the chain itself makes breaking changes right um so when eth moves like dial trees and we're using the original proofing you

know you might have to there might be a new endpoint there everything else will be forever um any body can come in so permissionless side anybody can come in and run the infrastructure so when you think about like a a and this gets into and again we can go down this road and the modular thesis when you talk about a world of of thousands or tens of thousands of chains or millions of chains uh Every Chain is beholden if you take a system so other systems Wormhole Axel all these other systems that exist they really what they're what they're selling is a validator set they're selling uh trust right and so here's a validator set r19 validators our 75 validators stake whatever it is here's a of trust that is willing to be priced and then here's a transport layer on top of that the problem with that Hub and

spoke model is that all of the work falls on them basically to do any the deployments and to continue to support and at any point in time whether it's them whether it's chain link whoever it is they have the ability to say we will no longer support this chain it's not economically worth it for us to support this anymore and there's nothing that chain can do about it uh it always the power always relies on the economic viability for the central system with layer zero when there is an endpoint anybody can run that infrastructure forever there's one app who cares that application can go and run it itself if the chain itself wants to run it it will always be available uh and anybody can run any piece of the stack and that is like a hugely different so like we say this really nicely in the white paper

layers are is the only system that allows the applications uh in by uh proxy the chains to reason longterm about both the security and liveness properties of the protocol everything else is basically up Tob so all of these other uh systems are effectively they're they're providing something very similar if you take it at face value they're moving bites everybody's trying to move bites the difference is are you a service provider are you reliant on that third party to do things and is there any case where you don't need to be reliant on that third party and layer zero like layer zero lab still does a bunch we run an Executor we like we're we're active within the network uh but you're not reliant on us anybody can run that infrastructure if we disappeared

from the face of the Earth today if the government came and and effectively issued an order that says you cannot run this anymore we got censored in one way or another we cannot affect an application who chooses to basically be separate from that the protocol will live forever and it is truly meant to be IM mutable rails a mutable infrastructure in perpetuity yeah I I really like I think that's a really uh critical thing to to point out and I think one of the things that people misunderstand about layer zero is maybe see layer zero competing directly with like cctp or Axel or something like that whereas you just my understanding right brand correct me if I'm wrong is you belong in a very

different place in the stack and in this sense yeah so again very easy to confuse because people like to they want to create a similar mental model they are selling the model of trust that is a problem they're trying to solve is is creating a let's say trust minimized the most trust minimized validation set right that everyone is trying to solve the verification layer and we're agnostic to the verification layer we actually think over time it becomes extremely commoditized down uh we think uh one we were did not think it's practical to be opinionated early on the model has already changed drastically you see people shifting from what was uh considered you know there's a very uh optimistic messaging was like a very big

thing for a while there's all these different Paths of uh I mean Nomad Sal optimistic messaging there's a very different path of like um going down this route um and so we've always been like extremely unopinionated we're not solving like we don't have a validator set that we're telling you is the right one you can fit any validator set in any of those people I name axillar warmal Etc can be used within layer zero layer zero is focused on the transport layer only that is the thing that we care about cool yeah and I think actually for folks following along at home maybe one analogy Brian you can push back if this isn't necessarily correct but there's almost like there was a discussion a little while ago about almost like Unis swap Labs the front end and the backend protocol and the back end is just kind

of like immutable sensorship resistant protocol and then there were different service providing entities in this case Unis swap Labs which built a front end on top so in this case un uh layer zero is aspiring to build the sort of backend um immutable part and I I love the fact that you really honed in on those sorts of characteristics about what makes blockchain special so if I could maybe try to sum up and then Hart I want to turn it over to you to get your sort of vision of the future here but basically Brian what I heard you say was you know going back to this analogy of the internet um what what we ultimately had at one point was this kind of discrete uh sets of uh different Computing networks and we needed a way to unify all of them because where we're

ultimately going to drive to here is the user doesn't really care they don't care if they're sending work to be done by a Computing set at Berkeley or this other Computing set it just needs to feel like this integrated holistic experience right but we want to find a way to do that while preserving like in blockchain land these really desirable properties of immutability censorship resistance Etc um is that is that about the right way to describe yes only that I don't think in the early days they they like the focus wasn't on this needs to feel amazing this needs to feel Etc right all of those things came down like economic incentive right the goal was that it was actually useful there was some net utility for being able to do that right it made the world the systems Etc more

useful brought higher utility so like I I think that kind of is the early focus and everything else all everything else like even these days most web developers don't care about the internet stack they don't care about the lowest level stuff everything just gets abstracted away up once you have a sufficient amount of utility for like worthwhile systems to be built absolutely that's a great Point yeah I think well I was just going to joke for a second but the idea of uh yeah the era of the information Super Highway with I think that was Al Gore's phrase and uh floppy discs are those all happened before most of our listeners were born I think so ux was not the major consideration my go ahead no I was just going to say um I completely agree with that there's a great uh you can actually

go back in time and find some really awesome uh Time Magazine cover articles about the information Super Highway and some of the visuals that just turned out to not actually make all that much sense but maybe felt intuitive at the time Hart I want to ask you the same question we asked Brian so we got um sort of Brian's Bishop about being relatively agnostic about um how some of the stuff plays out by providing this sort of low-level Transportation layer har walk us through I've asked you this question on the season before but maybe for the audience that hasn't heard this at this point like where are we headed in terms of tens hundreds thousands of chains and ultimately how are those chains going to speak to each other yeah so one thing I think Brian and I definitely do agree on is like okay there's going to be many chains but

whether there's like 10 or a million or like 10 million um seems very hard to predict and how exactly they play out although I do think Brian we should spend some time trying to like trying to predict it um I think it's an interesting conversation so we're going to have these chains we don't exactly know the market structure the one thing that I think is interesting to push on um and I'll keep it high level and then maybe Brian we can get more into like the um how this all works um but this tcip analogy uh I I've been thinking about this a little bit more recently and I like it and I like the idea and there's definitely this need to connect blockchains at a low level um but we aren't just moving bites or at least not

most of the time we're also kind of moving money or moving value and so one of the things I sometimes wonder is if the better analogy is more like connecting Banks uh like correspondent banking systems that need to be connected uh versus uh con connecting internets um and you know my my reasoning or my idea here is if you think of um remittance payments like Western Unions trying to you're trying to send money from the US to the Philippines in a remittance uh sense um it's not like every time someone in the US sends money to the Philippines there's an international wire that gets sent in the banking system like Western Union batches together these payments

they make the payments in real time and they send like the international wire equivalent once a day or once a week um and think that uh probably but it does they don't send a flop disc although we're talking a very old school way here too but I'm just using the example of generally speaking you you have something fungible which is money that like can be fronted on the destination chain and maybe because of this property you actually want to uh front things kind of kind of make loans on the destination chain this is going to lead into like kind of my views on intense maybe we want to

actually front things on the destination chain and and batch them together and verify that settlement more slowly yeah it's it's just it's functionally just a question of how many messages get sent in all of these models like there needs to be State basically being moved and so you say there's a difference of moving money and moving data to to me it's the same thing when we we got into the space originally they like there were just Bridges there wasn't like Universal mess like it didn't exist and our our basic thesis early on was it's the exact same thing like moving uh assets is basically just a subset of that that was basically the foundation of the oft was was like why are we why are there third party bridges that are doing this when the

token itself can literally be the bridge internally you can move a 100 million rather than paying six 8 10 12 bips to an external bridging provider the token itself effectively can provide you know hundred million dollars for the cost of gas so you're moving hundreds of millions of dollars for sub Cent and then like that layer of efficiency is such a world of difference of needing to rely on on any of this other stuff so I think fundamentally they are the same thing it's just like whether whether you're batching or creating this external social contract or like all of this stuff like you all need a way to move state to be able to do that like that's the fundamental problem that's being solved um well I think that's exactly the conversation I want to have with you um which is predicting like how many messages are getting sent uh

because I think that really does uh it does really uh color like what the endgame is for this this multi-chain world here and so like I fully agree you are sending messages in some way shape or form but this question of how many is is very different and maybe that's where we can like lean into this intense thing which is not something I've actually heard you talk about so I'm actually very curious to get your views on this um and so to back up for a second my mental model of how you connect to blockchains we have blockchain a and b um uh and I want to send send value between a and b um and one way way to do that is you send a message and this is sort of the the layer zero approach um

although you can push me on this but uh send a message from A to B and you want to do that quickly cheaply and securely um and then as I understand your worldview as this transport layer you let the application specify their security properties um but it's still difficult to send that message quickly cheaply and securely this is a hard thing to do all at once so functionally what you're actually doing like you can describe it as a message but really what's happening is some state is written to chain a and its intent is on chain B you need to prove that state right like if you're going to mutate State on chain B based on something that has happened on chain a uh you need to have a couple of things you need to have be reasonably certain that chain a is

like in a in a final state or like has reached a economic uh position and problemistic finality such that it would cost more to basically unwind the r every everything is basically based in that so uh if you you know uh issuing Credit in a lending protocol and somebody has deposited X here and they want to remove uh you know $50 million on the other side because you deposit $100 million here you have to be pretty certain that the deposit actually went through and that that's not going to get reordered out because once you give the money on the destination chain unlike a layer one which is the ability to roll it back you have no ability to roll back the state right um and so really what you're trying to do is prove state to the best possible degree on for a destination mutation based on some state of of

source and so whether you want to do that by somebody just optimistically do doing a bunch of X on this chain and then saying hey we're we're going to roll back and if I if I did it right I can prove it in the opposite direction or how you if you want to batch it a ton to reduce the the sort of friction or cost of moving that all of those are like viable methods uh to basically reducing the number of times you do that because there is latency and finality on a given Source chain and there is cost because you're you're buying block space to basically uh write the transaction right both of those things are are very um you know batchable or avoidable like it's everything with async systems it's basically a property of the source chain itself The Source chain was um you know

100 milliseconds and one 1,000th of a penny then you would have zero reason to ever do that you just send messages immediately right uh the the property that you're describing as a property with the underlying Source chain not with the system as an intermediate system yeah I think that's right so it's like your Source chain does not have instant finality um it has some sometimes to finality and some maybe probalistic Decay uh around how that finality looks um and there is also some cost to sending a message right so these are the the this is the the environments you're you're constrained by so um the path that I'm super bulled up on is this intent-based model where you introduce

this third party um and so so this third party we call them a relayer call it a filler call it a solver it doesn't really matter what it is and and the thing that makes this so interesting you can't do this in the tcpip world but because we're moving value at least most of the time and we can also talk about whether messaging is going to be moving value or moving um information in the future and there is use cases for both but because you're moving value most of the time you can actually have this third party this relayer go and front uh the user uh out of their own balance sheet uh their their fill on the destination chain um and you can have this happen in this intent model where the user is filled quickly and I want to go into why I think this is cheap and

secure users filled quickly um and then what the protocol is doing is escrowing their user funds and releasing them back to the user only if that intent is verified as fulfilled um and so go a little bit deeper but I think we can talk about how this gets around some of your finality constraints um and I think what laru is doing you interesting ideas here too but I think this sidesteps your finality constraints and can also make things really really cheap too well maximizing security it does not avoid messaging you still need the ability to message but at the extreme case I think the amount of messaging you do might be like three or four orders of magnitude

lower um than if you send a message per transaction um so yeah I don't know I I can kind of want to go deeper in this but what's your like what have you guys been thinking at later zero around this intent like framework I think the way that we view it is an intense system can be built on any messaging layer functionally like functionally intense can't solve interrupt intents need interrupt right so intense is a optimization on top of a yeah if you're if you're a it's like Atomic swap Plus+ basically right like you're basically saying here is this function for somebody optimistically settling something and then on the back end we need to like prove that state and LPS are basically taking the risk LPS

like the relay providers the time value of capital uh you know that's that's basically what they're pricing they're pricing needing to keep inventory on these chains or having inventory on chains and the time value of capital in terms of the settlement period to get it back in terms of when the message comes back but the difference is that interrupt like so if you believe in that world and that that's going to be like a dominant function it means that the there will not be a significant amount of messages that are that are State you'll never have applications that develop uh to have state across multiple blockchains you'll never have things that progress in terms of actual General message passing and I think like that's a thesis that we believe in like very very very strongly so uh I think like intense systems can live on top of a

great interoperability protocol but intense can't solve the interop piece and so uh intense as we view it are like an optimization uh you know a way that you can try to create an optimization in liquidity on top of of a system but you still need the underlying system and again the underlying system is like what we're trying to solve for intents are just a way to to batch or to optimistically settle this stuff but it doesn't solve the underlying problem like the transport or um yeah of the transport layer at all and so uh you can't like the best intent system can't do interop um so we just view it as as a layer that ISS above it's effectively a bridging layer and you can build any number of bridging layers on top of layers Z on top of a messaging protocol you still need to solve the messaging

protocol pie itself um because you need a way to basically prove the state hey all of this stuff happens I want to unlock the money for like the person who optimistically settled over there on source and you're still trusting that system to basically report back directly otherwise the Rel layers are all getting rugged right um functionally you still need to be able to prove State on chain B back on chain a which relies on inter solution yeah I guess where i' push on is I'm not sure I like agree with the statement it doesn't solve interop um it's like I agree with you that it still has the requirement uh that there is some underlying messaging system that exists and has to get used in certain cases but I think in this intent framework we can have a user obviously

Bridge value um and we can have them Bridge value and execute action so it's like uh this sort of a chain abstraction concept where I want my intent is go and um one example I like to use actually is poker chain this doesn't relate to your past this is just something I I I bring up where I I want to have poker chain it's my application where I'm betting poker and I want to control it from my home wallet and so I sign an intent that's like uh bet bet five bucks on this hand and I have quick inop that goes and executes that it's like moves value and executes an action um you could have intense Bridge user Ops in an account abstraction sense to go sign

something I mean the easy counter example is like I'm a I'm a lending prod call that lives on chain I live on some longtail chain right I don't have deep liquidity I don't actually know the global states in the market I want to go fetch the current price of something and process a liquidation based on that I want to uh I'm now splitting across uh my storage right my storage lives in in some other layer and I want to go fetch something like that can't be so you can't have an application that actually needs or relies on state or through an action mutate state or does any of those things in an intend based system you need a way to prove something has happened on chain a to be able to trigger it on chain b um and again even even on the intens ism like in all these

things you sign a message that does X like you need somebody some way to say that that thing that you signed is a canonical block on the source chain otherwise uh you know you can just basically rug endlessly if you're affecting if you're mutating any other part of the system so like an account abstraction um yes you can sign a message for yourself doing your own thing on the other chain which you could just like go and sign it yourself um but if I'm like signing a message that is going to mutate an overall I'm executing a trade I'm I'm doing I'm doing something in state something needs to to move that and prove that and again with with the across model it's everybody is frontloading these transaction where the users want to do a swap or Bridge like we're giving it over here we're giving it over here A bunch

of money has been locked over on the source chain I need some way to say I fulfilled my duty over here please unlock the money on the source chain right and if you don't have where that is corrupted the entire system is broken it relies on that functionally like you need that piece for it to be able to exist if you can't prove that you can never have the system yep I I agree with that and I guess there's actually two things I want to push on so one I also fully agree with the like if you want to do a cross-chain lending protocol um there I think I could probably design a weird way where you have people front capital and do this kind of like intent thing but it would be really capital and efficient and wouldn't really work um I

do agree with you that for those types of applications where you really need to observe State on other chains in almost real time or basically real time you have to lock State between chains to do things like in a cross-chain lending Market um you will need to be sending messages between chains um one point that I think we can come back to a little little bit later uh when we talk about the structure of this multi-chain endgame is how many applications look like that um this is an open question I have that I'd be just genuinely curious to get your thoughts on um versus how many applications are kind of like more rollup specific um so that's kind of one point that I'd like to come back to and then I think the other thing that's worth sharing with our our listeners is

let's let's say we agree that this intense system is an optimization for um uh types of transactions like bridging or bridging and swapping or whatever else I think it is worth thinking through what the advantages of this optimization are uh again in the in the vein of what does the end State look like like is this optimization important enough that it's actually very important layer in of itself that still requires the U the messaging layer below it and um one way that I think is kind of clarifying to do this concretely this has been examples I've using is um we talk about bridging or bridging and

swapping 30 million bucks so this is something like Stargate does pretty regularly across does too um and what are the overall costs of bridging and swapping 30 million bucks um and right now if you're going to Bridge and swap 30 million bucks over let's say there's like 10,000 transactions there's 10,000 bridges that add up to the $30 million you are going to have 30 like 10 ,000 deposit transactions on your origin chain and 10,000 fill transactions on your destination chain that is like unavoidable that's it's only in like it's only in the current method right so this is like functionally this has been it's kind of the nature of building immutable systems right ever all of

these comments that everyone has gone on on the bridging side is based on a version of Stargate that's like two full years old right immutable can never be changed like across just was upgraded again like there's all of these upgrade paths along the way to like try to improve the Stargate doesn't have that mutable from day one has never changed like the only real optimization that you're describing is batching right functionally you're saying here's a system for taking the underlying base messages and saying the cost the large bulk of Stargate cost right now comes from doing messaging it comes from like you have to write an eth transaction and write this transaction so if you're sending 50 messages you're writing fif 50 eth transactions and that's very expensive and if you can batch that into a single transaction it will be cheaper

like I 100% agree and I think like most systems are moving this you see batching like becoming much more commonplace in general um but it like that's again that's a bridging layer right that's that's a a Stargate thing it's a thing that lives above for us like layers who are very focused on like the transport layer itself not the bridging layer we think bridging is just a special case of generic messaging well that might be something that we talk about more over the episode too but just go back for a second so first of all across also until like this past week it it was 18 months old code onchain hadn't changed either all the optimizations were on the like offchain fills um that that have happened like all purely just relayer competing to fill users um so so yeah

but what I was actually talking about that I think you do agree with in the bridging or swapping use case and users want to move value between chains in any way like Baseline of any system there is going to be a a transaction a user has to make to deposit their assets On the Origin chain and and there's going to be a transaction that the protocol or a relay or somebody has to make to fill the user on the destination chain that is like unavoidable gas costs you can't actually move value between chains without them you agree with that yes there's a transfer on both sides there's a transfer on both sides right so you agree with that so then the question is okay well if we're moving 30 million bucks over the course of the day ACC through 10,000 transactions and it sounds like you also generally agree

what is the most effective cheapest way to move that $30 million over 10,000 transactions um what's the most effective way to do that was the most effective or is it the safest right there's like what what what are we optimizing for as like an overall system he everyone wanted to give a big shout out to today's title sponsor say now I want to talk to you guys a little bit about why I think say is cool from a design standpoint a big problem that it solves for eevs out there and then some cool stuff that say has coming up the reason I like say from an architecture perspective is again it's a very fast blockchain parallelization all of that stuff but say has essentially been custom building block space which is for Consumer apps and dexes now they have some very cool features which enable

that so twin turbo consensus optimistic parallelization s DB all of this stuff allows you to reduce the time to finality make for very very fast transaction so if you're building a consumer app or a DEX this is basically the blockchain for you if you've been building the evm you love the evm but there are some restrictions about it that don't support your app so maybe you can't do fast enough transactions or it's not parallelized whatever it is you can now take all that stuff that you built you don't have to start from scratch and you can build it on say now recently they've launched V2 but also public Dev net so the way that you can follow that and keep up to date is go and follow say Network on Twitter all right thanks guys I was actually going to say that's where I'm sort of shaking

out as someone who's Less in the weeds the two of you maybe to try to like sum up what I think we're describing here is I actually think we're describing a pretty holistic system where we need to move uh value and data across multiple different chains and there's ultimately at the end even though like har you and I have talked quite a bit about this intense sort of framework where maybe U sort of I don't want to say middlemen but like kind of facilitators right can like front capital and make things a lot sort of cheaper and more efficient ultimately these these things still need to be reconciled on chain and that's I think what the the sort of lower layer of the stack that that Brian is describing and I guess like honestly just hearing about I hadn't really thought about this in these terms so

always dangerous to think out loud I could be wrong but it seems like there's a a little bit of a Spectrum here where it actually might be safer to do like a purely onchain type system where we're not relying on these intermediaries to front uh Capital but what we might be sacrificing there is it's like it's not the most efficient uh thing to do especially if a lot of these transactions want to get originated on ethereum there's obviously a lot of gas costs there and so we want to figure out right like how much responsibility do we want to give to this optimization layer where maybe they could make things a little bit more user friendly a little bit cheaper but like har something you and I have also talked about is is like do we ultimately this is we're going to at some point talk about this during

this episode but like is that a desirable end state for crypto right that we have a couple large market makers sitting in the middle facilitating a whole bunch of these transactions and I think that's a conversation functionally what your sacrificing is is is loudness right like when when starg was built and again I I like the goal should be to talk about like layers are an interop not like Stargate Versa cross side I think bridging is interesting but I think it's like an application on top of messaging which is like less interesting than the messaging layer itself but like when Stargate was built there was this beauty in like Unis swap right Unis swap has been deployed in a piece of immutable code that that is like never going to change it everything was Unchained you knew how to reason about everything nobody could kill the system there was

no external party there was no like a government cannot come down and shut it shut it down there can't be like a cease and desist it was hey this was this really like Cypher Punk thing that is going to exist and will be free markets forever and like that in itself is like extremely elegant when Star was built the goal of Stargate was like these are pools of liquidity that will exist anybody can freely execute a trade there is no like Reliance uh outside of that basically it's meant to be like a closed IM mutable system that has pools with Equity it's very simple just like add here and subtract there and I think what we've seen over the last couple of years is the market for certain types of transactions whether there a bridging whether's swapping whatever it is the market is saying uh and maybe the market

will go back because it usually says this and then gets hacked and like security issues come up or livess issues come up and they say oh actually like those things mattered those things didn't matter right over time uh it just depends on like the frequency of these events but I I think what the market has shown is like they're willing if they're not sacrificing security they're willing to sacrifice liveness guarantees for a cheaper cost and I think that functionally is where like the overall Market has moved I think it is we will take some set of uh decreasing liveness guarantees in order to have a more efficient execution on a per transaction basis and and I think you see a lot of things moving that way whether it's RFQ systems there unex is like a lot of things have moved down and that's just like philosophically whether I agree

with that or not or whether or not I think is like the right thing or we're just Reinventing like another slightly different version of like existing Financial system it doesn't matter that right now that's what the market is voting for is that we will sacrifice degrees of liveness for this type of transaction if cost is cheaper and that that just like is what it is and Brian I agree with that and I think like the The Narrative Arc here is super interesting and I think even looking at Unis swap here it as just the example Unis swaps done this itself right it went from like purely onchain in a single chain version to Unis swap X and their RQ system um and you know my my mental model for why this has happened is because of this

multi-chain world where now that you have multiple chains that are asynchronous it's actually like you don't have the same liveness and execution guarantees because you have these asynchronous systems that you're trying to unite and it's messy um and you know what I think will happen is what you're saying we were going to have bunch of offchain actors that are the unifying force of these different chains um and then that's going to evolve five years from now into something else so this is this is where I think it gets like extremely dangerous and this is why I care so much about censorship resistance in a protocol so one of the biggest problems I have with most other messaging protocols today is that there's there's no form of

censorship resistance you have this Hubb and smoke model and there's no onetoone relationship between a transaction on source and a transaction on destination and what that means is like packets can be dropped which is you can see as important in like a vote going through right you have a government election or any vote happening uh there is no guarantee with any message that is written on source with any existing system that those packets will be delivered on destination chain like with the layer model everything is non sorder en Force directly in this basically transferer which means that you cannot deliver packet 17 until packet 15 has been like verified and is ready to be delivered which means you can if you're censoring a single message you stop The Wheel of Time for the for the entire uh pathway all of it and so it's like you

deliver everything and there's one person that exists in the world who's willing to deliver these or you do not deliver anything there is no intermediary of like hey we were told because of this we're going to start dropping packets like these people are no longer allowed to like transact across this we don't want their messages to go through every other system that can be allowed but I think the scarier part for me the much scarier part is like you think about hft in the current world like hft at least is reasonably meritocratic if you can even say that in the sense that is like the fastest actor gets to The Exchange and front runs the transaction and this is a cost that's like born by the consumer you think about it kind of like me but there's there's a wide sort of spread of actors

who are are trying to race for this and so they're economically incentivized is doing the first at the transport layer if you don't have censorship resistance you have the ability to reorder every single transaction which means every transaction going through a line now if I had a 100 transactions going through and they each have an me of one and I have like an extractable surface of a 100 I can reorder those transactions to create like an extractable surface of like 250 I can say hey actually I have more external incentive by filling my friend Citadel who I have this custom relationship with and I'm just going to fill all of these priority orders with they're going to sit at a higher priority of packet than any normal person coming from the source chain it doesn't matter what order they submitted to doesn't matter that they were later

in the queue I'm going to fill them on the destination side because I'm more economically incentivized to do I think you're going to end up with a world that potentially is a worse Financial system than the world that we have today because of a lot of this and so I think if you're going to go down that route you should care deeply deeply deeply about the properties of the underlying messaging layer and I don't think enough people do right now so Brian I think there's a lot to unpack there um and we should talk about all our trade-offs here I want to go back though on the censorship resistance point for a second here too because like I care a lot about this as well however I think it's a nuanced question and even in how I understand the layer zero design um in my opinion and you might find this a

spicy take but the point of censorship resistance moves off of the layer zero protocol to the application itself because the application as I understand in layer zero sets their own security stack and depending on how they upgrade or change that that can be a point where they could get messed with hey everyone this episode is brought to you by Unis swop delivering the best onchain trading experience period Bar None in all of crypto here's how I would divide the protocol today there's a web part A mobile part and an extension part you can say goodbye to the popup wallet extension was just not great from a ux perspective now what they've delivered is this nice clean sidebar which where you can just very easily track swaps sign transactions send to receive crypto

anywhere just a huge ux Improvement on the web part of it huge improvements here you can buy and sell 700 plus tokens at your price on your terms and they've got limit orders which are powered by Unis swap X and there's gasless they've also got real-time charts transaction logs pool data project information all this cool stuff that just really improves your ux trading on the web and again all this is powered by the smartest protocol in terms of Unis swap or Unis swap X click the link at the bottom Unis swap extension is in Alpha right now and we'll see what you can do about getting you moved up that list thank you very much Unis swap so is like the way that I view it is ethereum is meant to be censorship resistant usdc contract on top of ethereum can do whatever it wants

it can have a blacklist it can do whatever it wants internally that should live at the application layer I actually like I very much the protocol like if you have sensorship enabled at the transport layer it will get censored over a long enough period of time whether it's through Nation stage whether it's through regul whatever that can live at application layer and applications should comply with regulation that should be a thing that naturally happens over time you should not have it at the technology layer in the same way it yeah but I guess this is just the point that I want to push on it's just I I I find like saying layer zero solve censorship resistance here I find it a little bit too much it pushes it to the application and solving censorship resistance at that layer is extremely difficult depending on what

you are and how you do I mean that's like saying ethereum isn't censorship resistant because Circle can censor transactions or addresses like the base level protocol can still be censorship resistant with the application being able to do whatever it wants if I could jump in here and maybe just try to sum this up for the listeners who like I I think there's philosophically this is going to be a challenge that crypto ultimately ends up running into it's actually something that I've thought quite a bit about which is when crypto becomes undeniable from the standpoint of governments when do governments like ultimately fit into these systems that we're building right like because you could have this like total like yeah you know Bitcoin sticks it to the man and the governments are going to crumble and we're going to rise in our I've like

never viewed it like that uh you know I've always viewed it as like ultimately governments are important stakeholders and they're going to need to fit in these systems need to conform to laws and I think what we're asking ourselves here is at what layer of the stack do we need to conform to laws and I think there's uh an interesting debate to be had about whether or not that should happen at the very base layer that layer zero operates on versus Maybe like the application layer where people can decide based on their sort of jurisdiction which is I feel like we're we're getting at I want to do my moderator duties a little bit here and give you guys a little bit of time to kind of wrap this up and then I want to push us into uh some other areas of the conversation which is um just the amount

of like a the applications that you see in the future like how much do they need interoperability across these many different layers and um what you know are we heading to this world of we just uh spoke with Ilia uh Pushkin of near in the last episode he described this world of unbridged canonical assets to exist on the Chain where they were issued and actually minimize the amount of bridging whereas I don't know uh you know Brian in in in your view if actually we're heading towards a world where they're we're mostly going to be interacting with non-canonical or rra versions of assets but we want to make sure that those are delivered and are structured as safely as they possibly can be sorry I just front ran the next part of this

conversation but so I want to give you guys a chance to wrap up the censorship resistance part here and then I'm going to push us into the next part of conversation well I think your part on the censorship resistance was well said and I think actually Brian and I might actually agree on that you know like layer zero not being censorship resistance fine applications still have a job to figure out fine it's complicated cool like I think that that's fair um I do think we should talk about the hft points that Brian brought up too if if that's cool yeah I'm gonna I'm gonna be hard nose here and say let's let's try to get through those in the next like sort of minute or minute and a half or so and then we can get into the okay going real fast so here's my high level it's like there are some

trade-offs here and this isn't might be hard to get through in a minute we'll try um $30 million I want to bridge a day over 10,000 transactions um one approach I send a message for every one of those transactions that has a cost I have to send that at least 10,000 times Brian rightfully says hey if we can batch that together great but if you batch it together you have a slower user fill experience so the intent system what we do so what the idea is I think is you have offchain liquidity so this is the hft analogy you have offchain liquidity that is going to fill those users on a per transaction basis and those offchain actors can compete on speed or on price or on both and those offchain actors fill the user quickly

they can actually choose to take finality risk which is part of the reason why cross is able to fill users much much faster than chain finality which is pretty cool um but it's it's fully borne by these offchain actors they can fill users quickly and then we can batch together their repayments and repay them in batches every hour which means we could support those 10,000 transactions by sending a total of like a hundred messages over the system over the course of a day what are our trade-offs here our trade-offs here is that we have offchain actors that have to get paid for the loan they're making to extend to users the advantage is we send wave

fewer messages the other disadvantage to Brian's point is that there is an offchain component here and if that offchain component doesn't show up users don't get filled so to put numbers around this right if I want to like the loan if there's a $30 million of one hour loans that cost cost you $300 uh if you have a 10% assumed interest rate so I'm spending 10 $300 a day on like my interest component of my loans and I'm saving from going from 10,000 messages to 150 messages a day you can also secure those messages like slowly which maybe makes it safer

and cheaper and that's that's kind of what across does and you kind of have this trade-off here so this trade-off space where I'm able to very quickly fill users with offchain liquidity cheaply and quickly and securely and I get to send way fewer messages I think that's a pretty good trade-off but I think it is worth like highlighting Brian's points Zer around like wait is this going to be like a worse version of like Citadel and jump just front running users that's again that's that's purely at the bridging layer which is which separate so there's a couple things one you said layer zero is not sensorship resistant but the application is layers is sensorship resistant but the application basically

can add layers of sensorship to it if it wants did I just missp speak though yeah you just sorry I yeah sorry the way you're basically shifting messaging to say rather than starting with the user and ending over there uh you do it and offchain entity fills and then you basically message back which again is a fine optimization of Mac uh to make what it relies on is that the component coming back still meets all of the those things if the component coming back which in your case is Uma doesn't fulfill its needs or something happens then basically the relays all get rugged they Bear all risk of the messaging system and of the source chain finality right then they're just going to price

that accordingly that's fine I think you can again walk down the censorship resistant route if there were avenues for censorship within Uma and Uma was mandated in one way or another to do X or not process orders from these people like a lot lot less people will be able to the amount that that system can withhold external pressures strictly determines the longevity of the system right should it be a world or a system that says people from Nigeria are not able to basically interact with this at all people like X people like y right where does that actually live and if a cross was saying fine we actually want these things to happen but the underlying transport layer was now being censored or

corrupted then that would be like a very big problem for the application on top right and in this case you guys are are both of them so you have like some levers of control or let's say like a mirrored um probably set of regulatory jurisdictions or restrictions based on where you are but for every other application building on top of these systems like they should care the whole point should be decoupling the transport layer and the constraints on the transport layer really minimizing the underlying constraints so that the application can do those and I think ethereum again does this extremely extremely well um right like maybe vitalic is going to fill you for your whatever tornado cash transaction or maybe uh some of these validators are not processing uh sanctioned uh you know

sanctioned addresses but like others are going to as long as there is a part of the system who's willing to fulfill it and feels that they're meeting their sort of regulatory needs then the system as a whole processes and that's not what you're getting with most other transport layers and I think you and other people should like very much care about that and I don't think people do um um and I think that is very likely to play out in a very negative Sense on a longtime Horizon so like I very much agree with like this hey if you can get censored at some level like if you can get censored you're screwed right it's not good um and I I fully agree with you uh I think that there's like a longer debate that we don't have here about whether Uma can be censored or not which I don't want to

do now and Mike you look like you're about to cut us off gentlemen I'm going to I'm going to push us I want to push us into the next uh question into the next statement here um everyone should go back and listen to that I actually feel really lucky to be on this there was so much like there are very interesting tradeoffs have just sort of risk security uh cost like very very interesting conversation but I want to move on to this I want to move on to this question where like when we think about the end state for applications like what is what are going to be the requirements for interop and this is a really difficult part of the conversation because it's so critical ultimately the underlying infrastructure needs to support the application one of the the challenges with crypto as it exists today is that we don't have many

applications with traditional definition of product Market fit and you know Brian to use your maybe to pick up on something that you brought up with why a censorship resistance is so important is like you could imagine let's just I'm picking a an app out of like a which is deployed across multiple different uh multiple different chains you know if you if you have a very important vote that's happening on eth main chain chain and votes are getting passed from different chains like what you wouldn't want somebody to be able to do is like hold that up I guess the the counter to that might be in the future maybe we imagine that um most there we don't think about uh deployments across different chains in the same way that we do today and maybe like a exists primarily in this like ethereum sort of

trust Zone which has largely solved interoperability within one another do you see what I'm saying so like I I I would be very curious to get um a sense from the two of you about how you think about the needs for interoperability from the application layer in a in a fully mature environment for blockchains who do you want to start with and where where where do you want to start with that whoever feels the bravest whoever feels the bravest so okay like the a example is pretty interesting um because I think a is an example where what can't you do on a and Brian brought this up earlier like what can't you do that I want to do I want to deposit like eth on a on ethereum and borrow op tokens on a on optimism like

this cross-chain thing here so cool there's a need for I think messaging to make that happen this would be like a cross-chain lending Market what I really don't know is how many applications there are like that um and I think the alternative world and I think it's almost clearest to think about this with like games or something like that where if I'm running uh like uh onchange game I probably want to have my own blockchain my own rollup that I'm just controlling uh to I control all the performance functions and I'm just doing stuff within my world like I don't really need to talk to other blockchains to progress my game at least vast vast majority of the time what I do need to

do is have people be able to move gas and move assets on my chain um and I need that to be cheap and smooth and seamless I spent a bu time talking to like you know almost every large game gaming studio on this space it's it's something we're like pretty active in uh yeah most like they want their own execution environment that's extremely low fee and they basically have control over block space effectively that's what they want but the other thing that they very much want every single one of them is access to liquidity across all of the broader markets so they want to be able to take their assets and access ethereum liquidity they want to be able to list on blur on openc uh they want to be able different people buy and sell from other markets um these nfts themselves uh most

of them want to be able to come out in-game assets want to be able to go where they are the highest source of liquidity uh or have a process for dealing between them and so I I think on the gaming side there is um like it's a very clear example where the majority of execution happens on a on a single chain I actually argue with this a lot my co-founder um his argument my argument is usually like what more can we do in in in sort of single chain Atomic environment uh how can we verticalized and his argument is that 99 99% of all the world's compute today still happens on single machines but the internet is still the internet right like functionally like you can argue as much as you want that this chain or the

majority of execution is going to happen in a single chain I will very much agree with you uh there still needs to be this layer uh of the internet for anything of any measure of complexity to be built yeah so but it still gets interesting in Nuance right okay so we agree that you're going to have the vast majority of like compute or execution happening on a single chain um and you're going to have some number of single chains uh and they exist and that number could be 10 it could be 10 million don't know right um we also agree that there are some applications that require like locking State between these chains and this would be like your cross chain lending Market or something like that where I really can't do my

thing on this other chain until I verify uh that something happened on my origin chain um and then if we also so then going to the gaming example so we have then this concept of there's something of value that we want to be able to trade on other chains um and I guess really two forms fungible value and non-fungible value right and so then the the difference here is that's not like a synchronous thing that's like an asynchronous thing like as long as I can trade my asset I'm happy and I think like this is where I get open-minded about what does that end State look like like can you do you how much messaging is there involved in that kind of end

State and where I look at from at least the fungible perspective I think the intent perspective works very well I'm much less clear on the non-p punchable side I mean basically every every application that exists on the Internet is effectively async right so like what like if you look at early any early application built was just like this monolithic structure it sat in a single single server you would host the server you put the server in a data center whatever and everything would run on this one machine but as applications get more and more complex uh as user needs change as single environments or not look at any modern application being built today you have all of these web services you have all of these like microservices you're using whether it's different databases whether

it's things for compute whether it's things for you know all of these different inst structures storage all of this doesn't matter everything moves out to the bounds of basically optimizing the specific thing and because using that for the thing that it's good at is much better than using a generic system right for for almost everything every web app is built this way today every major application you ever use it built this way today and so uh if the thesis is that this application will live on ethereum and ethereum will be able to do everything it needs to do extremely well uh then that you know that's a thesis that you could could have or I guess it's a belief that you could have I think the most likely scenario is that you have applications that increase in orders of magnitude of complexity and that you have

Storage liveing R weave or ipest or filecoin or wherever it is and they take that storage uh whatever it is Data from that bundle it up do a really complex computation on salana and roll the end state of value whatever the result of that is to ethereum and it lives there right you can see this world evolving in a bunch of every we separate data and settlement and execution and all of these things start to basically bifurcate uh and I think as you have applications move down that rout we're going to move from a model that is like a single very naive structure that exists today to something that is measures of complexity I think if you uh if you believe that what we're doing blockchains themselves will scale real

Financial applications on chain that can service billions of people like that will have to happen it will just have to there's no way you have a single monolithic structure that that deals with all of this there's not any any chance of that happens and so uh it really is a bet on do you think applications will stay at the same level of complex they are now or do you think that they will be forced like bridging is moving towards to search for optimizations within the application execution will you be able to pay you know $30 or $100 for a trade or a loan or everything forever for every single one or will logic and and execution all these things be shifted elsewhere I think Brian that like that is the core question here so um uh we're

saying that we are going connect blockchains and there is some latency and cost to connecting blockchains for sending messages we agree if though we are primarily moving value then there is an approach here where you actually fill users with offchain liquidity right I don't think we'll only be using moving value though but I I agree with you I agree with you that we won't only be moving value but it's kind of like one of those questions where is it like 80% of messaging or is it like 10% of messaging right uh or 80% of I guess what is the result of the question right let's say it's 80% or 10% what does the end State look like you still need a messenger

layer that maintains the properties that you want or else the layer above just gets corrupted so like what what do you what in either bound of that answer what changes in the world I'm saying EI in either bound of the answer what changes with the messaging layer underneath oh sure you have a message you layer underneath but like Point properties of the change is Aire like Brian the point of this is like what does interop look like right so I think interop looks very different you still have like what does bridging look like or what does messaging look like can can I can I hop in here I one thing that I I actually find myself asking and maybe maybe the user is thinking as well is there's like there there's also a question here about I think maybe

something that we haven't directly asked but um is ultimately are most of the financial uh are most of the use cases of blockchain is going to be primarily financial and if that is the case like is moving value like and and all of the complicated abstractions and entities that like move that value ultimately going to be where a lot of the value ends up getting extracted from these systems or is it this sort of message passing system which we're agreeing I think everyone agrees here that we're going to need these General message passing systems but I where I sense a little bit of debate is like where that ultimately gets abstract or like uh maybe to skip ahead like gets to extract that value so even if it's entirely

Financial though the question is will the applica will the financial applications that live and exist be only on a single chain or will they leverage sort of the orthogonal benefits or trade-offs or optimizations of many chains like that's the ultimate structure if they're all going to be monolithic and it's all going to be value transfer then you have one world but I I view again I view these things as as like optimization down a path right you're reducing the cost for like the end consumer is is is like Robin Hood shifting uh Zero from from $10 trades to free trades but you get front run like net good does the majority of the world's volume like route through Robin Hood or or why doesn't it right

why isn't that like the end thing right you can have you can shift this around a lot of different ways and I think there are totally fine tradeoffs at the bridging layer at the swapping layer all of that you can make a abundant set of tradeoffs for where cost lives where values extracted how to drive down the end cost to the consumer but that's all reliant on having the underlying piece right and so like yeah so Brian I I wantan I want to agree with you like like I just want to like frame this as like what what's the most interesting conversation we can have here there is going to be a messaging layer layer zero is going to build a valuable business I agree with you right um but I think it's really interesting to try to hypothesize

how big that business is and I guess my question here is like okay so you guys are doing generalized messaging it's going to happen great and when I talk about this optimization layer I'm like well the business that I'm trying to build with AC cross is this if you want to call it an optimization layer that's trying to very efficiently move value between blockchains and I want to know how big this business is I want to know how much value is running through this business that's optimized to move value and um like where there is potential tension is if my business of optimizing moving value between blockchains um is like captures a lot of value and then doesn't produce that many messages It ultimately makes the layer zero business

like less big doesn't mean it's not huge and that's where we're maybe getting tense here but I want to understand like what this NC looks like I don't care about that it's not that it's not even me saying like layer zero has to be the end transport layer it's that we're talking about two different layers like you're like you're want to expose like why across is great and at the layer above and I'm saying across is just like a thing that can be built on top of layer zero like how that layer is defined I think is is much more overall in like down the path if you if you go down the path of okay bridging is like the number one use case and moving values the number one use case the question is will you need external liquidity to basically move that right

and that's that's functionally what it is it's if we're going and we say everything is value transfer then all the native all of the assets are just going to go to Native issues anyway right like Circle ccdp layer Zer of all this says you don't need a bridge at all now you can move again a hundred million a billion dollar whatever you want for just the cost of gas and that's it and you can batch that as much as you want you can add any amount of latency you can have Circle fill on the other side and then batch it back into a single message all of these things can be done and there is no need for any external fill or bridging layer if it's all done at Native the native asset layer anyways well hold on I actually that's not true like a cross is cheaper and faster than

Circle than Circle C Circle has its own constraints that it's built in it could fill on the other side no problem all right I want to try to structure this here because uh you know for the audience here right uh one of my favorite movies of all time is thank you for not smoking and there's a guy who's a he's a um he's can't blank on the word he's one of those good people that goes to he's a lobbyist and he he has this debate with his son and uh you know it's chocolate or vanilla ice cream better and he's like oh you didn't me he's like I'm not trying to convince you I'm trying to convince them so har and Bri I want to give you guys each a couple minutes here make your pitch to the audience I think um part of the reason

why there's so much like I'm going to call Friendly competition between the two of you is like both of you guys are building monster businesses both of you guys sit at sort of a different layer in the stack and there's this sort of natural competition right feel like hey like I think my layer is super important and like creating a ton of value here so I want to give you guys a chance to like instead of going over some of the technical specifics like kind of make the pitch to the audience for like where why you guys sit is like valuable and going to uh do a good job of like kind of capturing some of the value that you're creating and then I've still got two important questions that I want to get to here so we could just do um Bri if you want if you want to take the lead

here heart you can go and then uh I've got two more questions and we can wrap up yeah yeah I mean sure so again I'm not I'm not actually trying to convince you or anybody that what we're building will be like the most valuable like again the the Brian you'd never make it as a lobbyist my friend only that is the game is how do you build the thing that has the highest maximum utility and I if I build my piece of the stack right pieces like a cross can exist again a cross intense all of it cannot solve the actual messaging layer and you need that layer back and so if we want to move up the stack and say Here's how bridging will look here's how cross chain dexes will look here's how cross chain lending

protocols can look like we can talk the application layer but I actually think it's like there's a huge amount of indecision around what the messaging layer looks like and if that's done wrong everything else is broken and corrupted anyways right and like you never scale if that's not wrong you never scale to a million blockchains and you never uh have censorship rails and immutable rails and like all of those things that matter uh my point is if you don't get that right you can't get any of the other stuff right anyways because it can all be corrupted it can all be basically exploited or you know five plus billion dollars of hacks in like all of this stuff happens it happens not actually because of validation layer there's not a single of the last 15 hacks not one of them has been because

of a militia's validator set not one every single one of them has been because of contract upgrades for the large part Wormhole $350 million hack contract upgrade $1.8 billion critical bug contract upgrade another $1.8 billion critical bug contract upgrade nomad's hack contract upgrade sockets hack contract upgrade and let half of the people in the industry are still like these are the systems we should be you people just don't care I I hear you know Hart has spent a lot of time criticizing uh layer zero for many different things but I I've never once heard any criticism around that which has been historically the biggest vector by far for every other protocol uh and so as we like walk down I I think it's a much more interesting debate in naturally it's because the thing that I

work on and actually talking about that layer and why it matters because if that layer is wrong Nothing Else Matters like if we all move there was this you know I I don't to pick on B&B chain too much but there's a joke in the early days of of BSC that was like the cap show with all nine of the validators being CZ right and it was just like every face with CZ right and it's like uh like are we okay with a world that is a chain that looks like that versus a chain that looks like uh you know like ethereum or like many of these other do we care about Bitcoin do we care about the underlying layer do we care about the application so like are we here to talk about how the application layer gets built are we here to talk about like how

the actual interrupt system gets built underneath how does that actually work and I think that like those rails if that's not the conversation just want to talk about what is the best way to like uh have a bridge that lives on top of that we can have that conversation but I actually think it's has not been decided at all and I would be much more interested in hearing across uh or heart in general say here's why this is underlying layer here's why I think application should build on this or that here is why we support uh xc20 as like a standard and X and Y and Z right here is why we're trying to move to those things I think that's a far more uh impactful conversation on like a two-year time Horizon and everything else gets solved on like a five-year time Horizon because

if you get that part wrong you end up with a very different world you end up with not ethereum you end up with insert longtail chain here that has six validators and people saying okay you know it's it's doing what we want it's doing the compute um all right so yeah so I'll try to be a lobbyist here I am trying to convince you that like this layer um is important because I actually don't think Brian people think this way that's that's my difference here I actually think most of our ecosystem is thinking that the transport layer where you're working is the thing that we've all got to figure out and spend our time thinking about which is true I agree but I I think the this layer of how we send value between blockchains is going to be where 70 80

90% of all the messaging happens or all the the interop happens I shouldn't use the word messaging and so I am trying to make the point that the way we are going to connect blockchains 80 90% of that is going to be moving value and figuring out how to move value the most efficient cheapest secure way fastest way possible is a big deal and I don't think people are thinking this way and I'm very very convinced that this intent framework that we've been working on for a while is the answer to how you are going to move value cheaply and quickly and efficiently between blockchains and I agree that it does require underlying some messaging layer to exist um and

that's where we're at um so yeah Mike that's my summary no I really appreciate it from both of you guys like the the passion just like comes across U from both of you and that's exactly who we want building um these systems so I just really appreciate both your guys' perspectives there I've got one I've got two maybe we can end on these two questions which is one maybe a slightly related question and frankly we could have done a whole episode on kind of like monolithic and modular and ultimately like what the even that those two words for me are starting to they're starting to blend a little bit and I'm sort of thinking about it from the perspective like how tightly are distinct uh systems sort of coupled but um one thing that I would love to get

your take on Brian because we've been talking about this a good amount this season is what is your view on you know when we zoom out into this end state where applications have some amount right some amount we might disagree on some of the particulars but of demand to interoperate a lot of what people want in terms of interoperability today is like I want to trade these assets right like I want to trade uh ethereum but I want to trade it on optimism where it's cheap for instance and so as soon as you kind of move into that world you're dealing with mostly bridged versions of assets and wanted to get your perspective here on when you think about the future if there's 10 chains or 100 or a thousand it sounds like you have a very modular vision of the future where like even more so than today we're

decoupling the you know traditional constructions of blockchain so in this world are we mostly dealing with wrapped versions of assets or are we mostly going to try to minimize the amount of bridging and deal with um assets on their Source chains that I guess improve the uh the capabilities to the underlying protocols does that make sense yeah I mean I I think one of the earliest articles that I wrote is like why wrapped assets suck right wrapped assets basically shift the risk from a user basically says we're taking the risk from the LPS right you swap in Unis swap LPS bear the risk of the Unis swap protocol a user executes a swap they get their ass on the other side they never think about it again a wrapped asset basically says we're shifting risk from

LPS to the end user and the end user is holding that asset carries risk in perpetu as long as the asset exists is the underlying is broken everything's rugged the asset's now worth zero right and this has happened like many many many times so I think it's a purely worse model of the underlying asset I don't think Native assets have proliferated nearly as quickly as I thought because it is a risk for ccle to extend to some us some untrusted State machine that it doesn't know that can have some issue in tether and all these others and have that basically impact the overall liquidity and so uh I have always been a huge proponent of native assets I think it's why of is so compelling I think that has been why a lot of these other things have been I think naturally you will see you will

not deal with wrapped assets you will have uh a single issuer or protocol who sort of owns where its token goes and owns how that token is moved and takes away way the cost of moving those assets such that uh you don't need to rely on sort of third party bridging Solutions you don't have to inherit other risk you don't have to have wrapped assets or you have this other uh sort of structure that lives under and is carrying risk Brian this this this is actually something I wanted to ask you because one of the things we very much agreed on like the first time we met is that wrapped assets are bad and like I I'm like a canonical asset maximalist um here but but this is where I get I I want to understand why you think ofs are

not like a version of wrapped assets because like within the of protocol you can mint you can use the brid if you control the bridge the verification logic you can mint unrestricted assets so I kind of curious go ahead yeah the I mean the token that's just a parameter that the token sets right the token can basically R limit and do you can do anything that you want it's just saying existing structure is I have a token on chain a it's going to get moved to chain B you have two options right you can use a third party who basically wraps it Bridges it so you you have your transport layer and you have another bridging structure built on top of it and typically you don't even own that

contract on the destination side it's a wrapped asset it can't it's like vanilla erc20 that's coming out on the other side you don't have any controls over what it is and you're basically reliant on this external system and they're charging you rent they're charging you 6 8 10 12 bips typically those have been like historic bridging rates right uh now you have the ability to say I'm still you're going between two chains there's some element of messaging and there's going to be some risk within that messaging some there's some security property there that you're inheriting uh but the difference is you get to own it you get to set it yourself entirely you own the end contract you can set any of the rate limits you want you can set any parameter that you want with that and you pay zero rent to anybody else now the cost of moving the

asset goes from 6 8 10 12 bips on $30 million a day like you're saying down to zero bips uh for unlimited amounts or for whatever you said is a limited amount I've got a a followup I've got a followup course sorry I just want I've got a followup question for you Brian which is the implications though of uh Bridge minimum the uh Canon being a canonical asset Maxi or Bridge minimalism is that you are then if you want to interact with an asset that has Network effects like ethereum you are now now subject to the constraints of ethereum the protocol as it exists today namely paying high gas native assets you will always always always have to have wrapped assets you don't have a single entity who can issue it there's that's

totally different for most other assets that exist whether it be stable coins whether it be any protocol that's you know look at um any any decks anything that has a token that exists and as its chain moving around even even like uh Lio rap stake de there's you know all these different versions have ways to move these things to different Chains Got It final question Brian I'm going to ask it to you I'm going to ask it to you as well because uh we we've talked about this before but I want to get your three episodes in a sort of updated view um Brian could you give us a you know one of the things that part and I have been asking this season is like as we move forward into uh this sort of more multi-chain endgame it's it's sort of apparent that we're going to have to

make compromises in in some way shape or form right I think that's just like kind of pragmatic realistic approach maybe you could even like look at the example of Unis swap the exit like the Unis swap sort of V1 was this cyppher Punk sort of of dream where every the whole logic existed on chain and there was like well actually as we need to scale ultimately we have to make some tough decisions about which are the important parts of the logic to live on chain or components of the protocol to keep the underlying core premise but if we want to scale this to being able to support you know however many millions of Trades with then we have to move some stuff offchain or whatever that is so I guess one of the questions of heart and I have been asking ourselves this season is like

where does the line kind of end like what would we Define as a successful end to this whole crypto experiment curious to get your closing thoughts there I mean what does the end State look like that's like good what is a good outcome is basically the question I mean a good outcome is immutable rails that can be used by anybody effectively I think that's why anybody got into the space like a decade plus ago we all got in for this like very strong ethos driven right it should be uh money that cannot be censored and controlled by some Central actor with be government whether anybody else it should be you should own right this this concept of like self- sovereignty uh and the ability of like actually having control of real ownership of the asset

and the you know these uh permissionless structures anybody can run anybody can be a part of true to centralization like all of these things matter a lot I think uh you're right that likely more and more and more compromises are going to get taken and I think we're seeing a huge shift towards higher and higher degrees of centralization I think liveness is being traded off very consistently and I think like there's there's a different View and this is this is the problem that I have is is there's two views there's one which is the current topology and then there's two what is the actual architecture and this is like in the very early days of Bitcoin or ethereum and anyone else like reasonably small validator set like heavily you know not greatly

decentralized yet early on many times like known actors are running these all of these and they develop into what it is today and the difference is how it was actually built it was built and you could criticize current topology for it exists right now but architecture you couldn't criticize because the architecture is actually built to be able to scale into something that is and this is like the evolution I think the problem is that people are taking things that have similar topology right now and equating them and saying these are the same where this one's even better because current topology is better when the underlying architecture is infinitely more corruptible and way worse for endstate I think those compromises that we're taking along the way that many people and it's not not just in messages it's in all systems are that higher and higher degrees of

centralization uh in ways that actually matter and being treated as as equivalent to a system that is like secure uh but sacrifices like liveness right so uh let's say equally centralized systems one is uh secure but has risk of liveness I would say like a cross falls into this right is like it's it's sacrificing some element of liveness but it basically has the same measure of of security as the underlying messaging system so in this case Uma like matches uma's security sacrifices some liveness issues if if any of these things were happen and so like that's the design decision I think more and more things are going down that path but they're being equated to systems that are like equally uh or let's say two separate centralized systems one is

secure and sacrifices liveness the one is equally centralized but is not secure at all and I think people are treating those as the exact same system and I think that is like the fundamental broken part and this is why I argue against like uh intellectual dishonesty and like bad faith arguing a lot of times I think those things are being treated as completely equal in a way when they're just nut and I think it's happening sort of across the whole industry and I think it's leading us uh and people will care more again post more hacks post more bad things like you can only sacrifice security for a certain period of time until that like comes back to haunt you um and I think we've seen that just many many many times in the space over the last you know 11 years um so yeah that to me that is the

biggest thing that is like in issue right now and I wish uh we could talk another hour on because I think it matters a lot aome thanks Brian Hart what do you think yeah I mean first of all I love Brian's passion around that point like truly man um and I I actually think we probably agree more than more than you might think around all that stuff um around the centralization risks I I do think the debates here though are pretty nuanced like one of the things I worry about for from later Zero's perspective is by pushing these decisions to the application layer you're going to have some applications that make what are going to be in my opinion really bad security tradeoffs

and you know you as an un unopinionated layer like you can't actually have an opinion about their security trade-offs but like you're going to see people using your stack in shitty ways and that's that that brings up another really interesting debate is like should you almost be paternalistic about what you can and can't do with your stack I don't know and that's a whole other conversation right but I think it's an important conversation though I think we've taken the stance of ethereum side like do you think vitalic loved Ico season or dog coins or any of these other things being like the primary driver of ethereum for a very long time right is it his place to say what is a good application what is a bad

application what is like a good use of block space and what I mean maybe block space you can have an opinion on but but I think the point is if you build something that's a low less low enough level primitive and expressive enough it will be used in bad ways but like that that that's actually a feature not a bug that's a good thing that's a good property of a system you know man I guess like I very much agree with that but I think V vitalic is like super special this way where he does offer an opinion in a very measured way right he's sort of does Shepherd it's like this benevolent dictator type stuff that he does kind of Shepherd the the the space in a certain direction um without being controlling and yeah I think that is a very interesting thing that I think

you're going to face um as Slayer zero grows where there's going to be use cases and you might be in a position where you have to sort of offer your personal opinion about how shitty or not shitty they are but's that I have personal opinions on all of it the protocol should not be opinionated right the design of of the protocol and the way you architected and the way of what you allow and disallow should not be opinionated at all look I respect you for making that decision there are like some of uh I your competitors at the messaging layer the ones that I I think this model of the of uh ccip and axel art like shipping their validator set or like selling their validator set is a kind of interesting model but they're actually making in that in that decision

or in their whole strategy they're making a different decision where they are opinionated about what security should be and that's literally what they're selling and so I think that's just like truly a very interesting strategic decision that these projects are taking like you versus them and I don't necessarily think there's a right answer but Mike I want to answer your question because you're like hey what's yeah well you should think that like that's that's very clear like back to you being a lobbyist you should absolutely think that being an opinion it is the right answer and yeah I I can see the argument on other side um Mike back to your question though end state for interop I just think we want fast cheap and secure interoperability um that's it a lot of

Brian's concerns around censorship resistance I think Loop into security you're ultimately not secure if you can be censored uh if yeah that that will break you in the end um fast cheap and secure interrupt is what we want um again the thing that I still want to keep using the rest of our season of the podcast to to dig into is how much of interop is going to be value transfer versus messaging because I think it does um it does influence or kind of like color what the Nate looks like um so yeah and you know what guys it's a really uh I almost wish it we're getting to something really interesting right at

the end there which I think is ultimately coming for all protocols like Brian I I also really love the ideals of like the protocol I have opinions but the protocol necessarily shouldn't and I think if you look around like one of the big takeaways at least from my perspective and the deck season I did with uh Dan Dan Robinson you know he described Unis swap as having a very similar approach actually they sort of viewed themselves as a platform since day one but probably you know at some point um it it is difficult to escape an opinion as well and like maybe instead of like interjecting my own opinion like it would be really interesting and informative for play everyone to look at the evolution of something like Google right like where Google the original sort of value proposition of Google is to like Brian exactly what you just

described um it's like how can we just we don't want to have an opinion we're going to index the our our data in such a way that it's really easy to be accessible to anyone over time you know not to makeig political commentary but we've all seen the search results from Gemini which they've clearly deviated from their mission and we could have a debate about whether or not that's a pro or a con but it's probably worth it for all protocol designers like I think part of the reason why we're in crypto today is that we agree with that idea that like hey we don't actually like this and these there are certain things that be it Financial rails or passage for communication that they should just be these kind of neutral systems that are not opinionated so it's a worthwhile question to ask maybe for the rest of

the season about how can we maintain those values with increasing external pressure so I I completely agree and I think from day one like Ryan and I are extremely adversarial in terms of how we think about design from day one every ision is through the lens of if we layers or labs are maximally malicious or disappear for whatever reason one reason or another how does a protocol withstand those things and I think you have to and I think you're absolutely right down the trajectory of Google like if there was this IM mutable version of Google that everybody could return to and use and mutable version of what you know whatever AI system early on that everybody could return to and use then that will exist forever and there's no

way to basically change and corrupt that I think that should be the goal for everybody if you're building something with in there you want to give people using it and building on top an ability to reason about security and liveness again you can make trade offs on either of those sections but you need to give them the ability to reason about that they cannot reason about the long-term security and livess of a protocol uh there it's it's going to end very very badly one way or another Protocols are going to go out of business and these centralized uh you know systems that are there now these service providers are going to go down and every application built on top will just be dead and if they built their contracts immutable they will have no way to change that or do anything or basically modify complete

redeploy if the security property can be changed you see it with every one of these contract upgrades they will suffer and so like ultimately I think everybody should focus on if you're building infrastructure at all through that lens at least of that nobody should need to rely on us one way or another you don't need to rely on the Unis swap team you rely on their front end right Unis swap Labs does something it builds something around the protocol that is a creative to the protocol uh but the protocol itself doesn't care it's agnostic Brian I I want to like pay you a compliment in the sense that I agree with you that the way the layer zero architecture is is designed this Google problem that Mike's describing like I don't think it's

possible like they just you guys can't it's an un unopinionated and it can't be captured I I do think though that what is really fascinating and what's been clarifying on this conversation for me is now it really is up to the application so unlike your Unis swap example where in Unis swap V1 V2 V3 deployed to ethereum blockchain like can't be changed and it just works here the application sets their security concerns for interop and in any reasonable application there needs to be a way to change them because the interop landscape isn't settled and so we now have like a lot of trust we're putting in the application which is a reasonable trade-off to make but also kind of sucks

because like there's going to be some shitty decisions made it maybe just to because I know we got to wind down here what we're describing is you know we did a season of the show on Cosmos and the app chain stack and it is a pronounced difference in between how the cosmos's ecosystem approaches decision- making versus how ethereum approaches decision-making in that ethereum views itself as this IM mutable protocol which tries to push complexity away to the edges whereas Cosmos will just have an opinion like this type of meev is good or bad and you can look at many different examples of this from osmosis saying front running is bad back running is good we are going to internalize back running revenue for the protocol uh or you know ethereum will kind of debate

there's this there's this debate around issuance on the ethereum protocol is it good should we cap it whatever and in stride you had the liquid staking module where they were like yep 25% feels good it it there ultimately like someone is going to have to make decisions right uh so I think that's a good question to keep asking about the the rest of this season about where does the decision making sit in the stack so I think guys um I know we got to wrap up here I really appreciate uh both of you doing this this was a ton of fun Brian thank you so much man for coming on the show if folks want to find out more about you the work that you're doing at layer zero What's the best way for that to happen Twitter for sure I think the majority of

the space lives on Twitter we live on Twitter pretty frequently uh I think uh yeah definitely definitely the best place to follow anything new happen awesome all right guys um appreciate you both Brian this is ton of fun har and I will see you uh next week for next episode cheers guys hey both hey everyone want to give a final shout out to this episode's title sponsor say now there a whole bunch of really exciting reasons to be building on say V2 outside of just parallelization I want you to head over to say. to looking into building on their public devet again click the link at the bottom of this episode and head over to say. start building something [Music]

today