Modular March EP 12: The Business Logic Of RaaS & Shared Sequencers with Astria

Transcript

welcome to modular March I'm handy this is Robbie and we are the rollup modular March is an educational experience about all things Moder [Music] enjoy [Music] [Applause] [Music] welcome back to modular march today we

have a presentation from Josh Bowen from asria about the shared sequencer Network that they're building amongst other cool topics in the maer space welcome Josh hey thanks for having me guys pleasure let's Dive Right In yeah okay let's just jump right into the present so um as you said I'm I'm Josh I'm the founder over at Astria um we're building the shared sequencer Network and you know I know you guys have already had some presentations on kind of shared sequencers and a lot of the different components of the modular stack so I'm going to try to present you know really briefly kind of like why we're building a shared sequencer Network what we view that to mean not just in I'd say like uh like a specific kind of technical you know implementation detail of like you know what it's built up on you know happy to

touch on details there but I think um you know know at um you know you noit already gave some kind of good context there but but generally kind of why we built a shared sequencer um really is this kind of idea of we we think it's important that people can deploy a decentralized rollup as easily as a smart contract um so a lot of kind of how we've thought about you know the the business and the product justification for building a shared sequencer came out of this motivation to say like all right if we you know um the origin kind of came out of my time at Celestia as like an early employee there and kind of saying okay there's this new abundant block space we want rollups to be able to take advantage of that how do we think we're going to do that and the

things that came out of that is like you know for kind of like scaling reasons you know these data layers you know they scale better with like larger slower blocks just due to some kind of like technical details there but as we saw with kind of you know the discussion around like based rollups that's kind of been revived but you know the original rollups were based rollups but users end up preferring these kind of faster user experience and that's why sequencers generally kind of came about and so we kind of accept the thesis of you know we have new block space we think that's going to be used by you know rollups and we think those rollups are going to run you know sequencers and generally we we view it as important that these sequencers are decentralized you know and so then the big thing is what is the

tooling necessary Beyond you know the Celestia base layer itself to make this accessible and usable by these rollups and again coming back to this like deploy a decent r as easily as a smart contract because a lot of our view is if we're going to see you know 100 or a thousand or 10,000 or whatever pick your number of rollups our view is that you know it's not like every single rollup that is like built or deployed is going to find product Market fit and so if you want to have 10,000 of these you might have to deploy a 100,00 or a million of these to find the 10,000 that actually have um some amount of product Market fit and that means you have to reduce the barrier to entry for developers to be able to build these things you know what we need to move away from is you

know a lot of the great work that you know the kind of Flagship rollups you know your arbit your optimism your ZK sync your scroll you know etc etc they spent you know multiple years and raised multiple rounds to kind of build this you know foundational technology but if you want to have like a hundred of these or a thousand again you can't have the barrier to entry be spending three years of R&D and building aund person team we need to make this something that's really accessible to like a broader audience so that we can weed out the ones that don't work and find the kind of like uh survivors that remain that are actually kind of providing value to this overall e system and so at some level you know I think it is it is useful as a thought experiment to think of a shared sequencer as a decentralized

cloud and some of the big things that you know we think are kind of um you know they might sound minor but we think are like extremely core to like the kind of um the the mindset we kind of take when building the shared sequencer from again a technical perspective but also from like the product experience that people use going back to this deploying as easily as a smart smart contract um is the permissionless nature of using the shared sequencer so you know a shared sequencer is a form of like shared security and that you have multiple rollups that are all kind of using the single kind of you know sequencer Network that is essentially just like a new L1 coming to consensus over an ordered set of blocks where those blocks contain transactions from

multiple rollups um in that design we think it's very important that there's no requirement to kind of do a registration or governance process to kind of use the shared sequencer I think we've seen historically in the other kind of primary shared security designs notably right like Cosmos with their interchange security Now called replicated security and polka dot with their par chain you know polka dot had these par chain auctions and what we saw in the Dynamics of that with that if you were a project that wanted to kind of be a par chain you need to go kind of recruit through the market a very very significant amount of capital you know hundreds of millions of dollars in some cases um to win a par chain auction for the right to be one of these relatively

small number of par chains that could be allowed and then in the cosmos land you know to be like a replicated security chain of which we have like um like Neutron and stride right you know you needed to kind of win a governance fight kind of with the atom kind of like token holders and our view is that you know what we don't want to do is take this kind of like relatively well-known um you know business and Technical challenge of establishing your own validator set to run a sovereign chain and con convert that into this kind of Bey and governance problem of saying well okay I don't need to go get my own validators I don't need to kind of stand up that network but I need to convince an existing set of validators to gives me permission to kind of bless my new

chain and maybe that's a reduction in absolute effort but it's still kind of like non-trivial and is going to lead to a relatively small number of chains that are going to kind of like be willing or able to kind of experiment in this space whereas if you compare it to kind of like deploying a smart contract on like ethereum L1 one of the existing l2s right you know no one has to ask you just need to have some funds to kind of submit your transaction to like you know deploy your contract onto the chain and and that's all there is to it you don't have to talk to anyone right it's very permissionless and so that's something that we think about a lot in the kind of shared sequencer on a technical level that leads into kind of the product experience that's available here you

know again similar to this kind of cloud vers say like you know co-cation um kind of like history if you want to deploy a server you know you can go to a CO facility it's not that hard you know you you buy a rack you buy that from A supplier Dell or whoever right and then you go talk to your Colo facility and deploy that but it's relative to the effort of I go have you know an account I set up with oos on AWS or Azure or gcp and I click a button and I have a server that Gap and effort is like dramatically larger and we think we see a lot of these kind of like longer tail projects that are able to kind of like you know get their feet wet and see if what they have is some idea that they're willing to like lean into prior

to raising kind of all of this Capital up front to kind of try this larger effort so that's one of the big kind of Theses um that that that we over at as kind of think is like core to what we're building um and kind of touching really briefly on just like you know what is a sequencer at like a very very high level and we can kind of skip through this quickly you know fundamentally our view is that you know rollups or blockchains generally are in this space of like this transaction order flow and there's this like transformation of kind of like information right you have these users submitting these unordered transactions and the role of the sequence is to create an ordered block and then the role in our view of like this rollup full node and this like execution layer is to execute the block and then you

know it's like strictly optional but in the kind of rollup landscape you know we see people generating these succinct proofs of these blocks and so this is just like a high level SE so we can come back as like a reference if it's useful again touching on a very very high level of kind of like where shared sequencers and like decentralized sequencers kind of fall in the stack there's this kind of Matrix here of like you know in the top left you have a centralized independent sequencer and and this is the state of most rollups today right you're your optimism they're running one sequencer it is owned maintained Whatever by the foundation or the labs or whatever and they are only sequencing their own chain you know we're looking much more in this kind of bottom right category where you have you know a set

of sequencers right um and and each of those decentralized sequencers is doing the job of sequencing again going back to the transaction order flow for multiple rollups um again this is a very very high level kind of like picture of like how we see this right a user is going to submit a transaction to this consensus a set of rollup sequencers who are going to give a soft commitment and then we're assume that the sequencer is going to batch um these blocks you know of transactions to a data availability layer which is going to give a hard commitment so this is just kind of like you know high level kind of like um slides that might be useful you know for like Q&A section to kind of uses like reference points um again kind of you know trying to go through this presentation relatively quickly so we can get some other things

is um you know this is uh to the kind of like permissionless thing and one of the big things about this kind of like Cloud thing and a lot of my background is in kind of Cloudland is thinking about developer tooling and developer experience as like a core differentiation in you know businessto business kind of like software products you fundamentally a shared sequencer is a piece of middleware and so our customers are assumed to be software Engineers right these are people who are developers and so what we need to provide is a highquality developer experience that is you know again hearkening back to this kind of line about like you know as easily as deploying a smart contract um we are in the business of making it an easy user experience for people to kind of build things and so this is kind of our

component architecture and the reason we've built it this way is fundamentally that we want to provide an outof theeb experience um that gives someone that is maybe um a less sophisticated kind of developer the capacity to get their foot in the door but it is modular in its kind of like software architecture in a way that allows a more sophisticated user whether they became sophisticated after using our prodct product or they showed up kind of in a more sophisticated capacity to customize something that is like a better fit for their specific needs and so you know generally we're within like the Celestia ecosystem and so what we see is on the right side you know we have Celestia and that's kind of like not our business to kind of maintain that's a separate network but that's what we use for data availability and we have the sequencer

and its own relayer component and these are somewhat tightly coupled um that is kind of it's in a separate box because fundamentally this is something that we assume the end user you know our customer is not respons responsible for you know they're not assumed to be like a validator a sequencer kind of participant in the shared sequencer Network they're just kind of consuming this as a product for them and so this is this whole right side of this kind of diagram is assumed to be something that the the customer of Aster is outsourcing and then this rollup side is very much where we see um our users kind of working you know this rollup box you know the easy example would be like this is essentially just like a guest node right this is your rollup State machine this would also be like a cosmos SDK

and we provide you know these blue boxes are like components that like Astera develops and like you know um owns is um fundamentally sidecar components and pieces of software and a lot of this is taken from kind of like modern cloud microservice architectures in that we want users to be able to integrate a rollup kind of of their choosing whether it's a guest no whether it's a cosmos application whether it's something from like Sovereign Labs with the Sovereign SDK or something kind of like you know truly novel um to be able to integrate with the sequencer very easily in kind of a robust fashion so we we provide this composer as a sidecar component and what it does is it's just responsible for kind of essentially acting as a gas station right now for abstracting the transaction submission flow between a rollup and the sequencer and this is

very useful because it provides a very well-defined interface and an outof the-box component that's like a you know permissively licensed MIT Apache 2 dual licens kind of piece of code that runs in our net you know not in our Network it runs you know as like a side card to the the users um you know develop customers whatever um kind of stuff and and we invite developers to modify this essentially what we're saying is like you can get started with this battery included model and then if you want to change how you're interacting here is a template of like permissively licensed code that allows you to um customize it such that you know you may have specific use cases but we we borrow a lot of the lessons from what we saw I think in the meev design space um around kind of like

post merge with like Mev boost you know we saw these kind of like build and these relays coming about and when the merge first happened um flashboot was kind of the primary entity running a builder and they were signaling very strongly to the kind of like broader ethereum Community saying like look we're aware that like we are like a centralized point of failure within like block building right now we highly encourage everyone to create competitive Builders we have defined these apis like please kind of like come and build your own software and like compete with us and like reduce the kind of percent of the market we have but the reality you know of the situation was it was too large of a hurdle for a lot of these developers to get over to bootstrap one of these Builders kind of like themselves there was just like the the

initial effort to kind of get their foot in the door with too much and so the reality is we really saw kind of Builder percentage kind of drop away from flash bonds when they provided their own builder software as like an open source MIT license component a lot of people could just copy and paste it and redeploy it on their server and compete directly and then start to make modifications and so that's a lot of the inspiration for kind of like why we've provided the composer here is if we want people to be able to modify this you need to provide them a starting point that works out of the box that's flexible enough that they can modify it later but the bootstrapping requirement of them having to kind of you know Genesis an entirely new component kind of like with no template was kind of too

much and then this conductor is another thing this is kind of again we want to provide this like single kind of clean interface for how rollups you know again of whatever state machine you know whether it's the state machines that are kind of dominant right now you know generally in the kind of East space or the existing rollup you know or kind of like blockchain Frameworks like Cosmos SDK or something in the future we're going to provide a single well-defined you know API um and this conductor also runs a sidecar component into this rollup and so I I can touch on that in more detail but you know I'll just kind of move forward here um and again this kind of like transaction flow right this is you know again like more of like a useful kind of reference document we have more words in our blogs and

whatever but this is how we see the kind of flow of the kind of shared sequence they working here right where a user um of a given role up right so like a retail user whatever is actually interact with the rollup node and this composer component is again provided as this like out ofthe boox abstraction layer between any given rollup node and our sequencer the relayer is something again the sequencer and relayer box are kind of like within the scope of like asria it's not a thing that we assume our kind of like rollup developers are going to use and then this conductor is provided again as this kind of outof the-box component that's responsible for retrieving blocks both from the sequencer as well as from the data availability layer and providing these filtered blocks to the rollup node and

so you know we provide this battery included kind of like platform that people can kind of develop and iterate upon um you know just kind of like wrapping up I think this is my last slide yeah so so the last slide and this Loops back to kind of the conductor and the composer fundamentally our view on a kind of like rollup like full node architecture and this kind of idea of rollups as a derivation of a kind of ordered list of data that may be we'll say low context and that you know it's just bites it a series I stream of bites from the perspective of say the L1 the rollups data in like an ethereum case may just be called at or in like Celestia case it's just like blob data and like the L1 itself is intentionally kind of ignorant of the data and and

what that data might mean to another layer and fundamentally the the work that a rollup full note is doing is running a derivation kind of function right a derivative function over this like subset of data that is stored on the L1 and it is processing that through an execution engine to generate this rollup block and if we like look at all the other diagrams these are all saying the same thing but this is kind of the r view and this is kind of fundamentally what like the conductor is like enabling through like well- defined API that is essentially responsible for saying I'm going to go read the data off the shared sequencer and that's going to give me my soft commitment I'm going to go read the batch data off of the data availability layer and that's going to give me my firm commitment but it's going to

provide a single well-defined um API this interface that we call an execution API that's going to provide this list of rollup transactions that is um specific to a given rollup node that's saying you know okay we can have 100 rollups running against this shared sequencer in this da layer and you as rollup you know one might only want you know a given subset of those transactions in you know namespace one or whatever and that list of rollup transactions going to be provided to your rollup node via this well-defined engine API and then you just do the execution of the rollup block and so sorry the point I'm trying to get at here is like a lot of kind of asria as a project and how we think about ourselves

in the broader kind of blockchain spaces fundamentally yes we're building a shared sequencer um we are you know asri itself is a shared sequencer network but fundamentally the problem we're trying to solve is to make it kind of easy for developers to be able to build rollups and specifically we're looking kind of farther out we're not just trying to kind of satisfy existing developers who might be saying oh I'm gonna go take you know the op stack and I'm gonna Fork that and I'm gonna go deploy that on like AWS or I'm gonna go take like a polygon cdk and I'm gonna Fork that and I'm gonna go deploy that on like gcp or I'm going to go take an arbitrum orbit chain and maybe I actually do some modifications with orbit um you know using um stylus and have like at least a slightly different kind of state machine

um and I'm gonna go redeploy that you know we're interested in providing a welldefined but generalized interface such that we can see this low barrior entry and Innovation on the state machine so that the world of many roll-ups takes advantage of a lot of the properties provided by like a modular architecture that is not just saying oh we're just going to have like 500 different evm nodes that's fundamentally like an undifferentiated and to us an uninteresting future if you just say yeah cool I'm like the 75th L3 running on base using XYZ for data availability it's like great but like why would someone use that instead of the 74th L3 running on you know Basse or whatever right we think we need to provide a platform that lowers the barrier entry

for people to do interesting novel Innovative things on the state machine and that's a lot of kind of how we structure our project so I know that was like a long thing I'll pause there um but that was kind of like just giving a high level view of like our vision at astri and how we kind of approach this as like a like a business unit and where we see this over you know whether three months six months 12 months and kind of like what our motivation is kind of in this broader modular space awesome I love the vision um before we get too deep into the technical details um I'm curious about the business logic and and especially because you mention you brought up flash Bots how they had a significant market share of uh the Builder the Builder market and how they

open source their Builder uh and then and then they saw some of that market share dwindle as a result of people coming in uh kind of forking their Builder and then modifying it and optimizing it um and and how you're you're aiming to almost replicate a similar trajectory for the shared sequencer like what's what's the business logic there um for you know behind open sourcing it and not just kind of like taking that marking share and running with it yeah so I think a lot of it a question of kind of like where you're kind of like capturing Revenue right like the the like try it I'm pretty exhausted by this question now it's been asked of me like you know a couple dozen times over like the last two and a half years it's like where does value ACW in the modular stack right it's just a glow

hanging fruit thing this question has been asked like so many times but like fundamentally you do have to think about like where in the stack him like I as like a project as a protocol actually like capturing value and so you know the things that we are providing you know all of our code is open source because I think it would be insane for like any of these people to say like oh I'm going to go use like you know like a non-custodial cryptocurrency Network where like I don't have the ability to like look into the code like that's just like a good way to like lose all your funds um you know whether or not these customers are like actually going to like review the code themselves you know you want to make sure that like someone can but um fundamentally these components that we're providing is like

an open source thing right and these things to be modified are fundamentally you know hearkening to this um you know it's almost like kind of like TR at this point like Joel spolsky has a really really good blog post that every been references from like whatever 20012 2002 and he's a former uh early kind of person at like Microsoft who did like Excel and he did a bunch of interesting things after that but he had this post on kind of commoditize the compliment and what that means is like you are like somewhere in some like vertical stack of like software that is being developed and so in his case it was like you know he's like a Microsoft or like Excel or whatever and so you start with like kind of their example but like if you're Microsoft right you're in the business of selling uh primarily will St like

operating systems right on top of computers and what was very valuable to Microsoft is saying they want to commoditize the compliment so to have someone be able to buy an operating system it implies that they have to have already purchased like a computer like a desktop so it is it beneficial for Microsoft to reduce the barried entry for people to acquire computers right if computers are cheaper they have a larger customer base from which they can sell operating systems into right so it is beneficial for them to commoditize the act of kind of being a computer and what we really saw you know this is a little bit like have like you know 20 years ago or whatever right we saw this very large kind of divide between you know a PC and a Mac and you saw this very aggressive marketing campaign from um you know

Apple to say like we are like a Mac we're not a PC right they very much did not want to be commoditized because they knew that they are in the business of making money by selling computers right and to this day right like they sell very expensive kind of Premium relatively high margin as Hardware goes kind of computers and it was important for them to not be commoditized but from Microsoft perspective they're like look maybe I have a licensing agreement whatever but like I don't really care if people are buying like a computer that is like $2,000 or like $1,000 I just want them to buy a computer that they are then going to buy like a Microsoft thing and so from as's view like we are in the business of running a shared sequencer network to use the shared sequencer Network you have to submit

transactions into it and it has like an implied cost right it's a decentralized network you need like your dos prevention it cost money to like submit a transaction to like use the resources that network is providing what we want to do is we want to commoditize the um all of the kind of components that are necessary to use the shared sequence or network right and by and large there going to be like rollup Frameworks and to some degree and I'll touch on this a little more like you know roll up as a service things right fundamentally these components required for a developer to start using the shared sequencer Network we want to drive the cost of those as close to zero as possible because that's not where we're capturing value right I am not trying to like have a differentiated value by creating the best rollup framework from like a

polygon cdk or an orbit or an OP stack what I want is that all of those things are easily available to users to build on top of our shared sequencer platform and that there is not a lot of like rent capture right so you know there's been this like meme that I think it was like the bank click guys that put around right it's like the Spider-Man everyone shooting each other me whatever like and it was like oh the rollup of the service and the shared sequencers and the uh rollup Frameworks all kind of competing with each other right and to some degree yes right all of these people are trying to say okay we have a stack and you need all of these components in the stack to create one kind of working product that a retail end user might interact with which of these layers in the stack is

actually going to be extracting the most value and kind of a stance we have taken is we say look we are deciding that our middleware our shared sequencer that is where we are choosing to kind of capture value we are the ones that are going to have a network that is going to like AC crew Revenue through that kind of part of the overall kind of transaction flow and therefore we want to make sure that no one is able to kind of gatekeep via kind of like charging a premium for either the rollup framework itself right we've seen a lot of this right in the difference in adoption between say the op stack and like the arbitrum like orbit stack right like fundamentally right um there are some people now right and the arbitrum people are gonna get pissed at this but like the reality is

when arbitrum came out they said our stuff is BSL right it's like a business source license and the general view of a lot of software Engineers is not absolute but like coming from my background generally the view is I go to a repo and I see it's BSL or lacking a license and I just leave I'm just like I don't give a [ __ ] what's going on there it might be interesting it might be good code it is not permissively licensed it might as well not exist to me because it is just not worth my time to even investigate right I'm not going to go get into a conversation where I'm going to go ping my legal counsel and say hey can you review this BSL so that I can go have a discussion with like pachy over at the arbitrum Dow to figure out what

is their kind of ideological worldview that they feel is myself being a sufficient member of the arbitr ecosystem to have permission to use their code I just say ah there's the orbit train happening over there I don't care that functionally doesn't exist I'm going to go look at the op stack because that is an MIT license codebase and I can use that permissively and I can modify that how I want I don't have to have a conversation with anyone to figure out if I'm going to be able to use that right and what we've seen is a lot more people are deploying op stack chains than they are orbit chains right and recently we saw the arbitrum Dow kind of specify more explicitly here is the default out of the box kind of like license agreement that you can have for using arbitrum codebase right but

fundamentally they kind of caved because no one was using their damn code because it doesn't matter if your code is good if everyone just ah it's like too much effort for me to consider right like I'm like a Dev I'm like one guy I don't have a company not gonna get [ __ ] like a lawyer on the phone to decide whether I can use this GitHub repo or not right it's like not worth my time I don't care if it's better or not it's not worth my time right and so from our view it's kind of important that we don't see kind of people say oh the only way I can build a rollup is via code that requires me to go get a business agreement that requires me to have Le lawyers in the mix and therefore my starting point is a couple thousand dollars of legal fees

before I even start you know building a prototype to decide if something's gonna work right and then going one step further you know you know asri fundamentally right has done like a relatively poor job of kind of like marketing itself broadly about what exactly it is we're building right what is our kind of like Viewpoint kind of like in this broader space you know we've been a little confusing right we're building a shared sequence we're building our own evm we're building our own role up as a service and why are we building all of these things it's all in this vein of this commoditized compl we built their own rollup as a service because around you know mid last summer right there's a lot of hype right you have Caldera you have conduit right and the gelato is pivoting into this there's like 17 other goddamn rollup of services

right and like what are we hearing from kind of people who are interested in building a rollup is one they're kind of learning what is the kind of design space and capability of the role they can build from the rollup of the service providers right that's their entry point in the market so the customer relation is starting at the rollup of the service providers and then what we're also hearing is fundamentally these rollup of service providers are generally VC funded entities that raised relatively good valuations whether in the bull market or the bare Market or whatever right and they have to return VC style returns and they're enterprise software right there's software selling to other software developers the way you're going to want to structure your contracts is you don't want someone to rock Up Click a button and pay you a monthly fee with

the option to cancel that subscription immediately your investors aren't going to want to see that what they're going to want is you go to their website and you click a button and you schedule a call and you talk to the team and that team gets on and their sales guy is there and if you want to use their stuff you have a sales call and they're not going to try to get you on a monthly pay as you go thing they're going to try to lock you into a 12 month contract so that they can mark in their investor update next month that we have these many customers signed for 12 months and we can guarantee all that revenue and again this is another thing that is a gatee kept kind of like access to this people think the way they build a rollup or deploy a rollup requires them to go

with a rollup service provider and when they talk to those people they're going to say well I have Partnerships with XYZ rollup framework I'm going to be reselling those and you're going to have to commit to a a 12- month thing and then going one step further right you know we had like 48 for four like this last week right but prior to that you're seeing this very very high call data cost right 10 grand 20 grand a month to post you know data to like ethereum L1 call data so now what you're seeing is you're a user and you're saying I want to deploy an arbitr orbit chain I don't know how much I have to pay arbitrum down I'm gonna have to get a conversation with them I'm gonna have to get lawyers I'm gonna go to roll up a service provider and they're gonna say

I'm gonna try to lock you in a 12 Monon thing and the numbers we are hearing were ranging from $5,000 to $155,000 a month on a 12 Monon commitment so 150 Grand and then they're going to say I'm going to use you know you op stack arrum orbit whatever and maybe it writes empty blocks is called that or regardless of usage my floor cost of L1 called that is gonna be $10,000 a month and so now you're like cool I want to do a rollup and I'm faced with a whatever call it $20,000 a month bill with a 12 month commitment before I've even like written any code before I've deployed anything and now I have kind of contractual relations that commit me to like a 12- month effort you say well maybe I don't have like $500,000 in the bank I'm not going to explore this option so we built

our own wrath because again we need to commoditize this we need to understand what is the effort that is actually required to deploy one of these rollups um in a way that uses our shared sequencer and so we can understand from like a dog fooding perspective right what is the complexity actually use our Network and then also we can say look what does it cost in Cloud compute right just naively how much money does it cost to run a GU node and a block Scout instance and you know an indexer for that block Scout and how much band Wich is that going to cost me and how much storage is that going to cost me over what time frame at you know know even measure it at the standard retail no time commitment to like a cloud provider and then understand what are the margins

that we think these wrath providers are charging and how do we drive those margins down really really aggressively and you know we haven't kind of like doubled down on our wrath and I can give a demo of it you know we have it in kind of like a private kind of like you know internally managed capacity but the reality is because we just saw the market kind of take away a lot of this kind of gatekeeping ability to kind of like you know have the only one option and and these prices kind of came down very aggressively on their own so we've kind of developed this and we dog food and we understand a lot of the kind of metrics and the cost of these things but you know we kind of realized that like okay based on where we're at right now that's maybe not the best use of our

time but I know there's like a longwinded answer but generally kind of like our view is this thing like we're making money on sequencing I'm not making money on charging licensing fees on the rollup framework that's someone else's job and it's my job to make sure they're not getting as much money as possible for that so that my users have the cheapest kind of like initial buyin cost and and then therefore we can have the largest qu of customers I like the the business philosophy of showing a a a category hey it's not that hard what you're doing actually it's this easy that we can do it look how easy it is and then and then by showing someone how easy it is uh you show they're not providing that much value and therefore Drive their prices down they can't charge that much if it's that easy to do

what they do um and then and then the other point is like no argument from open sourcing software on my front big fan of Open Source in general Andy yeah no I mean I think we've talked about Ras being commoditized on the channel before Josh would you mind just pushing back a couple slides um I think to um one of the first slides where you were talking about your design I think before this one maybe or yeah I think basically on the on the decentralized cloud kind of conversation whether it's on the slide or not um and based on this on on your long-winded answer that you gave one of the key things was like this removal of like basically BD something

that we're like getting really really passionate about in the modular expansion like we can't have thousands of app chains or rollups or isolated e execution environments and need interoperability providers to talk to Every Chain every BD we can't have rpcs needing to do a grants just to get on their chain we can't have Bridges sequencers like all this has to be like like you say out of the box so the question that I have for you is like how is asria solution that you're providing um supposed to be that much different from the uh kind of antithesis of permissioned uh deployment of rollups and then eventual connection to Shared sequencers right so so so I guess just

like on like the permissions nature and like the gate kept nature right if like we want to say like who is like the primary Gatekeepers right and like people are going to get like mad but I think quite frankly like a lot of this industry is like unwilling to just like say the quiet part out loud thing that like everyone discuss right like like who are like the primary Gatekeepers in a lot of like the broader kind of crypto industry when it comes to new Chains It's going to be you have chain link you have layer zero right and fundamentally this is going to be chain link is kind of the like dominant like Oracle feed right but you think like what an oracle is right like the green frogs are gonna get all mad on Twitter but like [ __ ] right um fundamentally right what you're

doing is you're gonna go get some like price feeds from like Tower research or like whatever the hell right or like Amber group you're get these price feeds you're going to aggregate them across like a set of nodes you're going to run some averaging kind of mechanism 12 whatever the hell right and you're just going to post a Json blob or whatever the kind of structure is to like an ethereum contract on chain you can talk to chain link and chain Link's gonna say yeah that's going to cost you $10 million a year right like why is it going to cost you $10 million a year well because chain link has a token and chain link has a lot of employees and they pay their employees well and they want to be able to pay those salaries right but if you're like wanting to have

like a 100 200 500 thousand new kind of like rollup chains and these things are going to be like by by and large right like I I'm still have the view that like crypto is primarily like a financial thing right and so these people want to do like Financial tasks you know generally the kind of like rough things we're doing in like Finance you're gonna have like a deck you know you're have like a spot deck an amm whatever you have a lending pool you have a per and a per to some degree is just like an abstraction on having leverage on like a lending pool combined with like a DEX with like some different like kind of like technical mechanisms and that lending protocol is going to be reliant on having like an external Oracle price feed so that can like run liquidations right and so you just can't have like a

defi ecosystem bootstrapped unless you go get an Oracle and the Oracle is going to say yeah we know that and we're the dominant thing and we're just going to charge a premium for that right and so you know you see that and then you know on like bridging right like liquidity you know one of the kind of trade-offs of like an amm design right is fundamentally amms are like just very very bad from like a slippage per dollar of like tvl locked right like relative to like a like an order book or whatever right they just like have like a bad ratio there so you need to get a lot of liquidity and so what you see is these Bridge networks right like layer zero being like the strongest example here who are essentially saying well we're not just selling you the tech of doing

like a mint lock escrow Bridge which is like not that technically complicated of a thing right you say like oh I have like a token and I'm G like lock in a contract here and I'm have like an oracle that's going to like post like an ad a station of a state route on like a different chain that chain is going to Mint a derivative token on that maybe you have like a third leg that's going to say yeah now when I have proof that like the derivative token is like minted or whatever I go burn it back on like the origin chain like that technical design is not that complicated it's been done like whatever like 10 20 times now but they're gonna say yeah but you also get access to our liquidity Network as like an outof thebox thing and that's actually what you're buying with like a

partners with like layer zero or whatever right and they're going to try to restrain that access right because again they want to make sure they're getting their cut and so you have this thing where there's a lot of these technical fundamentals of saying okay I can go take open source code and I can deploy it on like one of n clouds and it's just not that hard there's a lot of like good kind of deployment tools we've built a bunch ourselves um but fundamentally if you actually want to get it into a space where it can be like a robust defi ecosystem you have to go get partnership from people like chain you have to go get liquidity from someone like layer zero and they know that that is going to be the kind of like limiting factor for you to bootstrap and for a lot of these

projects they're just going to say look actually the viable way for me to do this is to do something like blast where I'm going to have some like ponic mechanism where I'm going to get a lot of tvl and that tvl is allowing me to raise a bunch of capital or start collecting like a like a premium a couple bips or whatever off of that capital I have locked then I'm going to take that Capital I'm going to use the liquidity I'm going to incentivize developer ecosystem and then I'm going to use this liquidity and this developer ecosystem to go raise external capital and to give me negotiating power I go talk to layer zero and chain link so that I'm able to get a BD integration with these people at a lower cost because I've already done a bunch of bootstrapping work to get this right

what you end up with is a lot of this kind of like magic madeup yield because you need to have some mechanism of creating this yield to get this liquidity to allow you to get these BD things because you don't have any of this where the hell are you getting $10 million to give to chain link to get an oracle to bootstrap the fact that you can even have like a lending pool on your chain so that's a lot of the things we look at a lot of the kind of newcomers in the space I think hyperlanes a very good example of them just kind of biting the bullet and saying look what are we doing on like the bridging this stuff's not that hard like you just go make an escrow contract and you say well this is how you build like an esro mint lock whatever contract

on an evm this is how you go do that on a CM wam chain right this is how you go do that on fuel all of that can just be open source you can have a bash script that just says get cloned this repo put in your kind of like you know private key into like this field for your your deployment address right run a script and it's going to go deploy the contract there running a relayer is just like an outof thee boox kind of component go run it on any Cloud it's a fancy KRON job it's going to go relay packets from like point A to point B and and they can just provide that as like an outof the-box thing there's no need to kind of tie into layers here and there's difficulty like you don't get a liquidity Network kind of for free out of that but the

technical thing is kind of taken away from this kind of like gate cep space and then in the the kind of Oracle space I think we're starting to see a lot of new entrance right most notably I think I would point out like you have like the chronicle guys right it's like a spin out of like maker Dow saying like look we're going to get in the business saying we had an oracle the fine Oracle it works well it works at scale right supporting a lot of the maker activity there's no reason we can't just deploy again those contracts on any1 we can go run another instance of like the nodes that actually go read the price feeds from wherever and write that data on chain and we just can take a differentiated kind of Market position than chain link where we're say no we're

going to be in the business of supporting a large number of chains and we are just going to cut down on like chain link ability to charge these margins because very much A lot of these price feeds are like a good enough thing right like where is chain link getting like magical data that like no one else is kind of buying off the market right there's not that many centralized exchanges most of these people are going to have like an API with a quota that's just going to be cost that you just like pay them they're open to selling their data for money right um and so we're starting to see that and the last thing and that you know kind of as is thinking a lot about from a design and you know going back to this kind of like last slide in my presentation on the um the

use of kind of a you know a shared sequencer as this kind of Base that provides the derivation function that is the actual rollup State you know the way we see this and and where I'm kind of pulling a lot of the terminology from the Der ation function is out of like the the optimism guys really really good docks on like their A2 L2 derivation path that's how they do bridge transactions right they just look at the L1 they say ah there's a deposit transaction on this L1 I'm just going to derive at the top of the Block in my like L2 State machine that like Bridge transaction there's no reason we can't also use like an oracle feed for that right you can just say hey we posted an oracle update and now was posted on like the L1 or in our case on the shared

sequencer layer and this rollup or many rollups can all choose to just say oh as part of my rollups derivation function I'm going to just pull inherently at every single block timing and stack at the top of my block an oracle update that is shared among the shared sequencer and now all of the rollups that use the single shared sequencer all get the exact same Oracle feed and specifically they don't have to go do an integration with like an external partner right and fundamentally again going back to this kind of like business positioning right a shared sequencer is middleware and if you look in like a B2B SAS kind of like Enterprise case how middleware works right is a middleware is going to be kind of like a reseller of like plugins or like you know whether it's down stream or whether it's parallel you know these plugins and

ecosystems that like work into the shared sequencer and then the shared sequ the middleware generally and then you're going to have customers of the middleware right and what you're going to get is the more customers you get of the middleware the easier time as a piece of middleware you're going to have convincing plugins or or kind of ecosystems to tie in and use you as a channel sales partner to get to your existing customers and then the more Channel partners and plugins you get the easier it is for you to get kind of more customers and this is why you see a lot of people when they they talk about shared sequencers and you know when I you know do VC calls or whatever and they ask about you know us versus espresso or whatever right and they say you know this seems like an ecosystem

that is a win or take most and there is like a reality to that right middleware is something that has a flywheel effect in compounding games to integrate Gams to Integrations right um but I think we're just like early enough in the space but that is generally how we view this right can we provide you know bridging as something that is kind of shared among a lot of these chains right can we provide Oracle updates is something that is integrated with the shared sequence or that comes out of the box to these rollups and again going to this commoditize the compliment what are all of the things that we can provide as this batteries included experience to a developer well I want to build a rollup and maybe I'm going to start with this you know you know op stack or arbitrum or polygon cdk or whatever to develop my

like rollup framework but I can also use the shared sequencer and now I get decentralized sequencing and now I get kind of an abstraction around the da layer so I have to do that integration myself and then I also get an oracle feed out of that for free and maybe I get some liquidity and bridging out of that as well right and so they're able to bootstrap and benefit from a lot of this kind of flywheel Network effect stuff yeah so first of all man you've got a way of like crafting a visual for an answer which is it really paints a picture and I think that the idea behind actually going out and um commoditizing the the the compliment commoditizing and then offering the compliments is powerful and then in the crypto ethos it's powerful

from a whole different angle of like libertarianism of like rent seeking of like the reason why we're here you know but I'm still not sure you're going to be able to to like get over all those humps all at once like are people going to come to you and be you'll be able to provide them a 10x cheaper solution to launch a rollup compared to a Ras provider with that Tooling in the next year like is that a reality that you think is is actually happening yeah so I think there like a big question around like the bootstrapping right like fundamentally like a shared sequencer and it's like why would someone use you know X over like Y and like fundamentally from like a business a startup kind of like just like um like process we went about like with Astria

we kind of like skewed a lot of the kind of more conventional kind of like startup y combinator accelerator kind of style mechanism where you say go build the minimum viable product get it in front of users and then iterate on their feedback very rapidly and I think if you look at like these wrath providers right they took kind of like the lwh hanging fruit and they said I'm going to go take an open- Source rollup framework off the shelf I'm not going to do anything building my own rollup framework I'm going to go take their bass scripts or whatever the hell right and I'm just going to like straight just copy that code and I'm just going to go deploy that on like gcp or AWS or Azure or like on-prem or whatever right and then I'm going to go resell that and I'm going to

charge like a margin on top of that right and I'm going to clear that button and essentially this is like a classic like devops like platform of a service thing right you're making money on the margin of you have a bunch of like devops guys on your side and they are allowing these servers to kind of run with like high up time and you resell that at like a premium to like your end user for like a specific use case right in like General like Cloud you see people kind of trying to go up Market from like infrastructure of a service to platform of a service to like software as a service that is these more kind of like TurnKey things right and we took this different approach of saying we're going to look at this ecosystem kind of more holistically and I'm gonna say I'm

not going to go pick the thing that is like the off-the-shelf kind of like standard thing and just go redeploy that and like a very like unmo kind of like undifferentiated thing and that means that we're going to have like a slower time before we get in front of kind of like end users and customers and one of the big limiting factors of like asria right now from like an adoption perspective is a shared sequencer is like another L1 right it's a decentralized network and the reality of a decentralized network is that there's a somewhat arbitrary distinction between like a test net and a main net right like we saw this again was like layer zero right when they made like a bridge between like whatever like sepoi or whatever whatever test net it was right and there's like a market and like the reality is right like oh that's a test

net but like now there's like a market to like buy the test net coins so the test net coins are like real right they have a value because they're just Dem Market that's just how money Works um but from from our view right like we haven't like labeled like our Shar sequenc inter net we is like a main net thing right we're going through an audit right now right fundamentally we're building like a novel State machine for the system and so people can't build on top of the shared sequence they're in a main net capacity because they have a dependency kind of under them that isn't kind of like main net ready by a somewhat kind of arbitrary kind of like you know signaling kind of decision and so we fundamentally are going like lat to Market relative to these people taking something kind of more out of the

box kind of off the shelf but fundamentally our view is that you know we have what we believe to be the correct mid long-term design and we were willing to kind of accept the tradeoff of saying no I'm not going to do the ex absolute minimum viable thing and get that in front of customers right now because to some degree I think the market is not quite ready for it right we're seeing a bunch of l3s but you know my view is like we see a lot of these like l2s l3s whatever they're like op stack super Shar whatever but like to some degree they're kind of so undifferentiated they're like a marketing exercise right you like what is the difference between like mode Network and like optimism main net you're like well optimism main net is red and mode network is yellow right and

then there's like some like marketing things to it and there's like some incentive structures around like whatever the coins out of like mode Network and like their structure thing but like fundamentally codewise like the mode Network guys haven't made like significant changes to like the Upstream op main net and from an end user perspective they're kind of like why the hell would I choose to use the yellow network instead of like the blue network instead of the red Network right and so our view was saying we see that I don't think there's a lot of like differentiation necessarily in that space we think we can provide you know some of the functionality of what people want out of these rollups and these rollup of the service which is this job of sequencing this abstraction of running the kind of infrastructure we can provide that in a way that is

abstract from the specific State machine implementation which in this case might be again this op stack this arbitrum orbit this polygon cdk and so yeah we're going to be a little bit later on getting the immediate things right now that are ready to go and getting those main net but we are able to carry on a lot of our kind of investment from like an Institutional technical knowledge because we have a large team you know my team is like 17 people now right and of those people 14 of those are like protocol Engineers right and then it's like myself as a CEO and like Aida as our great head of strategy and like Christian is our coo right like 14 Engineers most of them do are protocol engineering right we have a lot of institutional knowledge and momentum on how to do protocol engineering we've

invested that and therefore based on how the future goes we're not completely dependent on these rollup Frameworks from saying oh well optimism and arbitrum and polygon think this is the way that a rollup should be built in three months and six months and 12 months if you're a person who just has Pur devops experience and it's in the business of deploying code that other people write if they choose to go a direction that is not desirable for you from a business ecosystem or from an ideological position you don't have the institutional protocol engineering knowledge to build your own kind of rollup framework on top of that because you didn't make that investment you may have gotten to Market sooner but you just have like a differentiated kind of amount and quality of institutional knowledge for being able to kind of move in the market and so that's like a

fundamental trade-off to your point on the kind of like specific you know are we going to be 10x cheap I don't think we are and I I gave a talk at um you know the the espresso guys kind of like beyond the Bas layer event at like East Denver that kind of touched on this in like a little short like two-minute rant like fundamentally we we do as like an industry have to kind of decide like what are the thing we're doing here right are we doing centralized things or are we doing decentralize things fundamentally decentralization at an absolute technical level is going to be more expensive because like how do you decentralize a thing you have more people run more copies of the same thing right and one of the reason I'm very excited about Celestia is like it changes a little bit of the ratio there

right you can have one person run some fractional percentage of the overall kind of like data that makes up the Ledger instead of every single person has to you know store 100% of The Ledger and therefore you have like end copies right but at the end of the day the difference is like you run a centralized service you might have like three copies of the data you might have one like us West you might have one in Europe you might have one in like Asia or whatever right if you're running like a blockchain like how the hell many copies of like The ethereum Ledger are there I don't know there's like half a million like validators and that's like an arbitrary biding mechanism like you might have like 10,000 or 100,000 or whatever like full nodes of like the St right that costs money like someone's

paying money to run computers and so similar on the sequencing thing right if you run one centralized sequencer that's going to be cheaper just to accept the fact that like okay it's like a centralized thing do you want to be in the business of doing centralized things right you know are you okay with like having the Escape patch you know I think a lot of people talk about oh the Escape patch and your funds are safe and I agree on that but fundamentally if you're like using optimism and like the government is like hey the optimism sequencer is going to be like turned off and like good luck getting that back on you know maybe that's off for like two weeks like how long before people are just like hm maybe the only transactions I'm submitting to the L1 through like the SK patch are like not transactions

to continue trading using optimism maybe I'm just going to withdraw all my funds off of like optimism because fundamentally I came here for the sequencer and I don't want to like fud the like regulatory risk like excessively no but there is this design question of if you're doing a decentral I think it's going to be more expensive and are you okay with that and what kind of fundamental like design space does that allow you to work in our view is we need to kind of commoditize everything else such that the Allin cost maybe it's not absolutely cheaper but it's providing a high quality user experience that is worth people paying this cost of it being decentralized J you know what's gonna happen is those Ras guys are gonna end up offering Astria as a shared sequencer on their Ras and then people are going to be like well [ __ ] why don't I just go

use Astria yeah and or they're going to build their own shared sequencer right like the least surprising thing in the world like I would be extremely surprised if like the conduit guys were not building their own shared sequencer right now like that would be like very very surpris in to me right um I just assume that these people are going to kind of like you know a lot that's like like showing my own bags thing right but like to some degree based on my time spent in this industry I think I have like a relatively good record of building things that I then see a large number of kind of competitors come up into the space and build a similar thing and there's a fundamental reality I think um you know um you Patrick Collins over you know the founder of like stripe you know had a really good podcast on

like DW's podcast recently and you know they talked about like Moes and how do you have a moat there is just a moat incompetency there's just like how do I have a moot like I'm going to hire better people and I'm going to execute better and if I continually do that I'm never going to give someone else an opportunity to catch up to me and everyone's like oh that's not comfy that doesn't let you quiet quit that doesn't let you retire early and you're like no but you're doing a competitive business right you can just hire better people and compete and you know if people in competitors come in like that's a good thing right you know to the point on commone Tiz the compliment right there's this very like Bezos term right it's like you know your margin is my opportunity right fundamentally if you

just say I can run a tighter ship I can run like a more efficient organization I can just have a 1% smaller profit margin than you and I'm still able to run that efficiently enough that I am still profitable enough and maybe you're not profitable at that margin anymore right and and I think one of the things in crypto that I'm hoping we see in the kind of rollup space and going forward in this modular space kind of enabled by a lot of this flexibility and different kind of economic specialization different layers in the stack is more competitive Dynamics where you know one of the biggest kind of problems I think in crypto is like these networks like don't go away right there networks no one uses that just sit there and just like have like a captive like 20 billion dollars in them right I think I actually

want to see more competitive Dynamics where yeah three people are building the same thing and maybe two of those people go away or maybe one of those people goes away right but the people who survive will be more competitive in the space they'll have a better quality organization they'll create better quality code better quality products and the end user experience will be better for that and we won't have a lot of these kind of like Gatekeepers that don't have like sufficient economic kind of competitive pressure to improve their product and aren't allowed to just kind of like sit on their hunches how do you uh prevent a compliment from coming in and commoditizing Astria I I don't I don't think I don't think you do really like to some degree right like that's just like a competitive Dynamic right you can see it from like a marketing perspective again

going back to like the apple and whatever I think we'll see this with um you know again like the arbitrum people right they tried to do this by way of like making their code like BSL right they said okay you guys just can't copy and paste unless you essentially pay a tax and that structurally looks a lot like what you see with like apple and like the App Store and like the 30% rent they charge on like you know Apple music for Spotify right you know apple is okay to have Spotify running on the iPhone because they know that Apple music definitionally will always have a 30% margin advantage on Spotify because Spotify has to pay rent to them arbitrum doesn't have to worry about other orbit chains competing with arbitrum one because those orbit chains are always going to have to pay a 10% tax to like

the arbitrum Dow so they'll always have just like a mechanical advantage if we look in the um you know optimism space right you know a lot of what they're trying to do is they're saying we're going to be like very open and kind of you know permissionless and like MIT license and a lot of props to them for that and kind of like giving their software away for free in that capacity to some degree there's like a legitimacy to it right why would someone want to use like a rollup Network that is like the like illegitimate kind of like child of this I think that's what we've seen a lot with like mettis and Boba right when they like forked like optimism B1 there's always this question like why would you use the one that's the fork of the mainstream one the mainstream one's still going why do you want to use like

the illegitimate Fork of that um and we see this in crypto networks generally right forking is fundamentally like a social consensus thing and so from an asria perspective like fundamentally again to like theote via competency our job is just to provide a better quality user experience and product and developer experience developer tooling such that if someone else chooses to say oh we're going to go get clone our repos our repos are MIT everyone can like get clone our freaking repo and go run like a separate like ask you a shared sequence or network if they want right and we just have to believe that we ourselves are a sufficiently competent organization that those people will do a worse job utilizing our code than we will be able to do utilizing our code and to some degree like if we can't if someone else can go take our open source

code and like make a better product then like maybe we shouldn't be in this business at all because like we were able to create something and we were unable to kind of like you know utilize it to its maximum ability maybe it's better if you get kind of like beat out of the market by like your code was open source and you just weren't able to capitalize on it great example of this is Craig right forking Bitcoin and trying to do his thing with Bitcoin SV while Bitcoin the the OG is still running look how look how that worked out um Josh this was this was a banger man I think we might call this one the competency tax yeah sounds good to me I mean if you guys have any other questions you know happy to chat I have a little bit of time I can go over but really really

appreciate you guys having me on the podcast yeah I'll shoot for one more which is are you worried as like we had a really interesting conversation where proof of work Bitcoin is like this proof of energy proof of electricity it's like this hard-backed money if you will and proof of stak's almost like a meme it's like if if everybody believes that ethereum has value everyone's willing to stake to believe in this method then there's a lot of social consensus so a theorem itself is almost like it's a community it's a social uh consensus more than almost more than like a monetary consensus are you worried at all about how you're GNA bootstrap the social consensus of Astria as a extra L1 for rollups like how are you going to go

and convince you know all like arbitrum op ISM or even like smaller ones that hey guys like we are a viable option and like like we have this X Y and Z to prove it and like here's who believes in us like like how are you gonna cross the chasm of like the the social consensus yeah I mean I think like fundamentally right like my view is that like a it was like the like a like the biggest mistake ethereum ever made was like moving to like proof of stake right like fundamentally giving up proof of work like they could have very easily had like a dominant position in being the only smart contract TR that ever proof of work it's going to be very hard to bootstrap one of these fundamentally it was to some degree just like the ethereum people being like a little bit

too leftist on like the environmental front and just not having sufficient confidence that like their coin was like sufficiently valuable to justify like some energy expenditure on Earth and yeah it's just like it's don't have enough self-confidence like the project as a whole but to like the prove a stake kind of like you know like mtic aspect I think to some degree the fact that it's all kind of like mtic and fake and just like social consensus thing right like makes the bootstrapping problem like very easy right it's very difficult to kind of like justify people to like get over this hurdle of like running a bunch of like you know proof of work miners like making as6 and like taping stuff out and like doing all this work to kind of bootstrap that if it's proof of stake there's just kind of like ah like people

have money and they will put that money where like they get yield and like the proof of stakes thing like generates some like yield and so people just kind of go there and from like an economic legitimacy or an economic security or whatever there's like various like forms of like scops and like how you can just like make the number higher right you know like the reality is like base put like a contract on chain and they didn't even seem to have hired Engineers yet and people put like $500 million in the Box on chain like the bar to have like these idiots like put money in like a box seems to be like the floor basically it's like ah you have some good tweets and you're backed by Paradigm um and then on like the technical mechanism of like you know what's an actual like

attack Vector on one of these proof of stake chains like one we've just like never seen it actually happen like the East classic people you know y you know now over at like Celestia has like dealt with like a lot of the forks when like East classic and that fork happened and dealing with like the real-time forks and how to like kind of reach social consensus again on that but in like a pro of stake Network we just haven't seen any of like these big divergences because again like a lot of these chains are all going to be kind of like varying levels of VC chains they're going to have like relatively low float when they launch you know and and the float's going to be owned in like you know staking contracts you know do you stake before like things are unlocked maybe it's like you know technically

impossible and that like you know two-thirds of the stake weight is like stakes and can't even be transferred via the state machine rules of the chain so like good luck attacking that and even then after it's like how many centralized exchanges are there for the volume of these right the top five exchanges by volume are probably like you know some like 80 plus % for basically every single coin out there like these people are generally good actors they don't benefit If These Chains Get Like maliciously taken over like they will find the right person to talk to if they're like hey someone seems to be trying to buy like a relevant double digigit percentage of like your overall Network you know I just don't think we have you know people who are like Carl Ian level like activist investor trying to like

maliciously attack like a proof of stake chain and even then you just kind of hard fork and you go to social consens and you just say like ah do like the majority of like the crypto Twitter Discord whatever users like not like whatever billionaire like maliciously tried to acquire like two-thirds of the stake by buying all the liquidity on like okx and binance then like sure we'll just like hard Fork it's like you know 10 lines of code and we just like zero out their account balance and like you know good luck have fun like see if the centralized exchanges will still take your coins um I just I'm not that stressed like obviously I worry about it right because fundamentally it's like necessary to the network and the Boost ding but I just I just think it's like a little bit overblown or you can do

something you know like we've seen a bunch of other projects they like ah I'm just going to like I got to go get like the social consensus that is ethereum and that just means that I have to get you know like the ethereum foundation like Priestly kind of cast to like accept that like I am Shilling their bags sufficiently so fine I'm going to go have like an AVS component of this such that I have like some igen layer restake e and that just like pumps up my Economic Security numbers by like $500 million like there there are a lot of mechanisms here um but in general it's not something that like stresses me out too much because we just haven't seen one of these attacks happen and the the bar to get to like a sufficiently high amount of capital it's like a non-trivial thing that doesn't cost like

millions of dollars of a capital B stake and would like persist Beyond like a social consensus discussion I just don't think that's like a reality that we live in right now Josh um I think this has been my favorite episode of majra March um thank you so much for uh making us laugh I think it's a Friday thing I mean not take any credit from you but Fridays The Vibes are up we just had a great morning and I mean this has been extremely extremely insightful from several perspectives in terms of business logic um onchain uh incentive incentive design game theory Rob's got a math and Philosophy degree so um I know his brain is very stimulated right now as well so thank you for for coming on um and uh

sharing some insights yeah really appreciate you guys having me cheers