Shared Sequencers in the Modular Stack | Josh Bowen, Ben Fisch

Transcript

we don't want to continually have this phased appropriately say okay I'm going to go create my new roll up I'm going to stand it up I'm going to have a centralized sequencer and then I'm going to start you know the clock and the effort on you know establishing a decentralized set of sequencers for every single roll up right we've used this is just fundamentally kind of unscalable um kind of like human coordination work um and that there is a design you know this shared sequence they're designed that allow us Roll-Ups to kind of start be you know decentralized at like the default state for when they can fit welcome back to another episode of Xerox research today is April 17th and we're running hot with the modular thesis this week

um this time we're joined by two different co-founders Josh and Ben who are the founders of two different shared sequencing protocols Austria and espresso uh they're both at the Forefront of the modular thesis so it's really great to get their ideas and their insights on really just how they think about the future evolution of blockchain but before we get to the interview we're joined by Westie and ren from the blockworks research team to discuss everything going on in the current state of the market uh Ren how about I toss things over to you to kick us off this week with a little hot seat cool thrown yeah sure I got a custom for this week my custom is the votes and team over there last week on Tuesday they launched votes V2 one part off that V2 is a better interest rate amm for the

interesting itself but the larger part which is also the Machuda part was the non-custodial Clearing House they announced for any derivative asset so basically see what you think of it is it's a cross Marketing System for any user or any wallet or even any decks to cross margin across any derivatives exchange so it can be options it can be perhaps it can be an average to change there's any derivative that you can think of it can be integrated and a user can have a single marketing account to trade across multiple exchanges so there's a lot of user experience upgrades that you can enable to that for example you can open the ease log on one perfect change and close on another

perfect change you can use any governance approved asset as collateral to open a position on an exchange and it's something that's been explored especially within the Solana ecosystem given the low cost of rebalance and collateral one protocol that did this was modify until the mango hack happened and they saw the pivoted a bit I think this is the first case I've seen a sort of caring house within the evm ecosystem and I think it's going to unlock a lot of opportunity from ux sort of user perspective but also for devs building derivative protocols what's the uh the user flow for this like if I want to take advantage of this do I go to volts

and then volts kind of handles everything on the back end for me so they kind of open the trades on the the exchange of my choosing and kind of like capture the trade that I'm putting on or what does this look like to the user yeah so the Avenue released all of the details I'm guessing here that a as a user you would go to votes you would open a margin account let's think of that as like a sub account on like a centralization right you can open multiple of these sub accounts safety you can have multiple cost marketing wallets and from there you can interact with any decks that's integrated within this carrying house and from there you can open and close positions to carry house access so they can automated liquidation engine so that it ensures that none of the connected dexes become

insolvent and so to be honest it's not so different from any trading experience you have you know if anything it may be better because you're able to trade on multiple venues at the same time for example if you had an Eaves long on one perfect change and you had the Eve show that another perfect change now the collateral is sort of aggregated aggregated together and you can use the same collateral across both Physicians right which is a decently common trade that people do so delta neutral trace and it's a very intuitive music would it be fair to kind of call this like a an aggregator for Perpetual exchanges I think that's one way of putting it I feel like that's sort of understanding it a little bit

sort of like a I think the best way I would describe it is a infrastructure layer that allows you to cross margin across any derivative exchange so you ever only ever need to trade from one location and you only ever need one set of collateral you don't need collateral splitter class multiple exchanges now my question is how do they accrue value if the users are still trading on these different purpose changes the fees are going to those exchanges where this sort of volts fit in from a value accrual perspective yeah so in the initial light paper they did mention a fee of some sort on sort of opening or closing transactions with

any connected exchange I think one may ask for example why do I want to pay that extra fee right but if you're gonna run these trades that use multiple dexes or multiple purpose or options for those because at the same time that Capital efficiency if you think of it simply is like instantly doubled right whatever like if for example you needed one million on one exchange for each long one million on another exchange for you short but actually those two positions cancel out from a Delta perspective so it's a delta neutral trade now with votes V2 you don't need 1 million in one million you just need one million right and so now you return the quote unquote like doubled from a

percentage perspective and I think that's where the value accrual comes in I read your report and I thought it was super cool and I just want to do a little self-shill for you I think that everyone should head over to blockworks research.com and give that a read if you haven't subbed yet you definitely should I wonder if this leads to a Confluence in funding rates across different perfect changes did you have uh that Capital efficiency where you're able to long and short and basically the same trade across different exchanges um because at this point I mean if you use anything like dydx or Quinta a lot of them have completely different funding rates so I wonder if sort of you can they converge on the same over time yeah it does seem like a power user uh power tool for Arbitrage Rogers so we'll

be interesting to see if it kind of gets used in that especially considering a majority of at least on chain spot volume is driven by Arbitrage and overall Mev so um we'll be interesting to see if that kind of continues on and use leveraging Volts for that some Arbitrage purposes as well but continuing with Ren's general idea of improving crypto ux I'm going to put all of crypto ux in the cool thrown this week it has uh it's been a long time coming since it's probably still even a bit of a stretch to put them there but um crypto ux has been a pain point for all of us for the past couple years and you know still has a lot to go but over the past couple days I have played with a Solano phone and downloaded the uni

swap app on an iOS device so that right there just being able to say I've done those two things is a massive improvement from where we are even just like six to 12 months ago so my experience with the Solana phone was honestly remarkable I thought it it definitely exceeded my expectations which were admittedly quite low but I still walked away feeling like it was a zero to one moment being able to hold like in a phone that it feels high quality and whatnot and right next to the Google Play Store it has something called adapt store that I was just like this cool unlock it you can click into the DAP store it has like the mango market so the Jupiter app and it's got like nft exchanges and just

like they look like you're downloading you know Uber or whatever app you would on a normal phone so that was super cool and then specifically with the uni swap experience I haven't like put any assets on there started trading with it yet but the idea that I can go to the Apple App Store download an app uh when you set up the wallet you have the option to do smart con uh smart or sorry iCloud storage with your wallet keys which you know the security the security nerds out there I'm sure having like a mental breakdown over that um but from a ux standpoint it was effortless to open a wallet there was no memorizing a seed phrase they're like the even the wallet address is somewhat like not hidden from you by any means but it's not like shoved in your face as this massive hex string of digits that's

like definitely turn some people off so they're like ux is clearly a focus and that is feels like this this you know sentiment change in the industry which I think has been a cool unlock yeah I agree there there's been a huge amount of mobile first mobile Focus ux Improvement for example out of Eve Denver you saw that project to come up the hackathon I think it was called upcave which uses to secure Enclave within your iPhone so for those that are unaware to secure Enclave in your iPhone it's used to stores user keys and sensitive info basically give this guy found a way to use it to store a private key and there you have a perfectly secure Hardware wallet on your iPhone right and for example out of East Tokyo there were a few interesting projects

one was using your voice as a form of 2fa which it does pretty interesting maybe some security concerns there if someone has a very similar voice to you or someone wanted to use like an AI voice model to recreate your voice which has happened already or for example another I think much larger ux upgrade was another project called Mina and it uses Japan government issued IDs and the nfcs the NFC included in the ID to sort of enable you to pay for transactions to your crypto whatever so rather than for example go to your wallet approving your transaction checking all the details all you have to do is tap your card on for example like a payment scanner or your

phone and there you have it a transaction goes through and that's a user experience that a lot of people in developed countries are familiar with right like Apple pay pay wave just swiping your credit card on tapping it and those are the type of user experience improvements I think that will onboard a lot of users but at the end of the day it still depends on deaths to build actual interesting use cases even if that functionality is there no one's going to use it if there's nothing to actually do right yeah strong and agree around with everything you said I also just think this is like the beginning of of a mobile push for crypto I think Dan I don't know if you remember but we talked to Antonio two three months back on this

show and I think he said that dydxv4 will be putting a focus on a mobile app at least I think so don't don't quote me on that but I believe he said that um so that's super exciting to see and I think this is just the beginning of a wider movement I mean everyone's always on their phones I guess I would just like to see some more games kind of go this route because the games obviously don't stack up to console and PC games so I think the hyper casual genre for games makes a lot of sense for crypto native content I just know there's that pesky battle going on with uh game developers and App Store and how to work nfts under there so that'll be interesting to watch play out as well Westy who you got in your hot seat or cooled around

yeah so in my hot scene this week I have Kamikaze for the past couple months has been extremely bearish on the Shanghai unlocks um basically berating anybody who thinks it was it could somehow be a bullish event with billions of dollars on locking Etc but what we saw last week with chappella was the upgrade went successfully um we saw around one billion dollars in the East be withdrawn from a combination of both partial and full withdrawals but yet the prices actually increased 10 since the event um so it was sort of a non-event if anything a bullish event because we've actually seen more people deposit their Ethan while withdraw the wreath at this

point um and so like the flows are really in the positive direction and as we go forward I think you know by the end of the year we'll definitely have more uh East Stakes than we do currently and so I think people realize that um while there was that Supply overhang it really was a super bullish event um but yeah looking at the queue right now it looks like it's sitting around 21 400 in terms of the execute so if you wanted to unstake you would take roughly two weeks around 13 to 15 days um so if you wanted to unstake it'll only take two weeks uh whereas something like Cosmos would take three weeks um but yeah it's been a massive overhang on East price over the past couple months we saw the East Bitcoin ratio

dropped to its level lowest levels in a while um so I think this has really given a lot of like thanks to eth to a lot of alts associated with eth and yeah it's great to see that was all sort of a nothing Burger yeah that that's I guess I gotta get myself a pat on the back I took a risk last week and uh did a preemptive cool throne for uh you know this ending up being a nothing burger and it kind of did turn out to be just that we even seen some nice bear capitulation on the timeline as you mentioned Westie uh but I got a question for you so right now the withdrawal queue is about you know 13 to 14 days as you mentioned do you think that this is still like the initial wave of withdrawals that's kind of pushing that

number to where like pushing up that withdrawal kill to this this time frame uh or and do you expect this to kind of trend higher or lower from here yeah it's interesting because the queue like right after Shanghai wasn't really super high but then over time sort of grew and grew whereas like a couple days ago I think it was 17 000 now we're 21 000 um so it's sort of been creeping up as people I think are just testing withdrawals as well as I think there's a lot of people waiting after the initial event expecting there to be a long queue and seeing oh oh wait it's actually not as long as I thought and starting to withdraw but I think over time like I don't think it'll get much higher than this in the short term I think over time it's going to slowly uh weed its way

down and I think it may converge on on a like a week or so after a longer period of time and maybe go through certain in waves of withdrawals here and there for certain events and whatnot but yeah I don't see the execute getting much higher than it is at the current moment I think anyone especially big entities such as Kraken who needed to withdraw I think they've already done so or if they're doing so or drinks doing so at a very steady Pace as opposed to all at once and so yeah I don't see the queue getting super high from here yeah that eth uh Bitcoin ratio too has been whole Notch stronger on Monday at least a lot can change in the next two days so I guess I'll bite my tongue but you see

Heath holding above 2K and Bitcoin having lost the 30k mark briefly so kind of surprised to see that honestly with uh all the drama around chappella yeah there's been a lot of alts ripping especially those Associated to East so a lot of liquids taking derivatives those tokens have done really well the layer twos arbitrum and optimism have done well and I think that's going to continue to be a narrative because the next upgrade is Cancun Dena which just post a common Q3 of this year and that is going to include proto-dink sharding which obviously helps the margins of these layer twos so I think that's going to be a huge narrative going forward yeah I think that's a no-brainer Westie right like we saw how hot the liquid's taking governance tokens were running into Shanghai and obviously because that was just such a huge upgrade and such a

huge like momentum pull as far as like attention and narratives work um and now we're gonna see the same thing again right we have this massive upgrade that has a huge unlock specifically for l2s I'm definitely going to be paying attention to what's going on with the our layer 2 tokens yeah that's a perfect segue as well into my cool Throne I've got actually Hot Seat sorry uh I've got and it's actually myself unfortunately so Arab airdrop sellers uh I did admittedly sell the airdrop uh back at around like 1 30 to 1 40 or something like that but it seems like they're really getting their act together with uh the governance drama that we saw a couple weeks ago uncovered pretty extensively 1.1 and 1.2 just got passed today at 2 p.m eastern time and there's a lot of cool ideas on the forum

for a grants program or our value cruel avenues that could be taken or you know pursued to provide more value cruel to that token so really bullish on the arbitrum ecosystem in general you also just see transaction activity um holding pretty strong despite the airdrop having been you know already passed so that's really strong uh that's that's a strong signal I think you also see Dex volumes pretty strong on arbitrum coming in at like number two over the last few days just behind ethereum um so yeah I guess I'm putting myself on the cool throne for this week by selling the airdrop but we saw the price run up to about 170 over the last few days it was over the weekend uh so yeah if you held hats off to you yeah I think part of this has to

do with like I said that like layer 2 rotation as well as it seems like in this recovery a lot of people are latching on to the newer shinier tokens and so ARB is one of those new tokens blur I think is another one that's doing extremely well um the market really loves especially in a rally like this to to latch on to the same few tokens that have shown to be outliers or ones that you know are new that everyone's looking at and so I think our was definitely on the back of that and yeah I'm with you Sam I should be in the hot seat as well I sold my airdrop um at like the local top I felt pretty good at the time and now it looks like a pretty terrible decision but it's good to see for arm as a whole I'm going to

need the hot seat too I sold my arm hey everyone knows on this intro I I think one thing that I'm really looking forward to is how the operation Foundation is planning to use that initial 175 million up that they're going to get vested for the first year we've seen different chains sort of incentivized different types of Partnerships like party God Rihanna's like the bde chain they do a lot of like well web to focus Partnerships and I'm interested to see what the first few big Partnerships that come out of this ecosystem Grant fund is going to be and I think that should shape the narrative going forward for our picture because they haven't really done any of like BD or like ecosystem growth initiative so

far a lot of that has just been pure organic growth perhaps on the back of everyone farming the airdrop but so far like no one's already dictated the direction that average was going to in any sense yeah we've also seen an increase on Nova which is their chain that has a data availability committee a little bit less decentralized so it's kind of more I guess optimal for social and gaming applications so I wonder if if that uh Telltale is a Telltale sign for people expecting kind of a retroactive ARB air drop similar to how op's been doing theirs yeah that's an interesting point as well Sam and this will kind of this whole conversation kind of segues right into the interview that uh is coming next and you know we really get into the weeds of why uh

decentralizing a sequencer is kind of important and really in the ethos of crypto and speaking on mechanisms to do that from a shared sequencing standpoint right and so uh in the interview we really went quickly from you know Hi how are you to what is the meaning of life and it'd be great if we could get like a a back step here Wesleyan you know you're you're super In The Weeds on this topic as well so if you could just you know kind of walk us through the role of the sequencer today and in a roll-up and maybe we use like Arbor trim since that's what we're talking about let's use that as an example so what is the role of the uh sequencer today for arbitrum and what does it mean that it's centralized yeah that's a really good question and I'm super decided to listen

to the interview I'm sure they get super in the weeds but yeah like a sequencer is essentially how transactions are ordered on a roll-up so to think of this from sort of the user's perspective and say you want to do a transaction arbitrum where you actually you send your transactions to a sequencer and they're the ones that are basically in charge of ordering transactions and they basically tell the user yeah like I promise I'm going to put these transactions in the layer one at some point in the future that's called like a soft confirmation and then eventually it does do that once it orders the transactions and the way it sees uh it won't it wants to order it for arbitrum it does like a fair ordering system so it's sort of like first come first serve others have different ways of doing that

whether like capturing Mev Etc and there are a couple reasons why you know a centralized sequencer is not good and a lot of people want to decentralize the sequencer I think like two of the more technical reasons are aliveness and censorship resistance so if you have like one single sequencer who's one node that's taking all the transactions ordering it if that goes down then the entire roll-up goes down so if you have multiple sequencers all of a sudden you just need one of the sequencers to be up for the roll-up to maintain liveness and then with censorship resistant if that one sequencer is just like I don't like you I'm not going to include your transaction well then you know either

the user can force their transaction L1 or if you have a decentralized sequencer all it takes is one honest uh sequencer for you to get your transaction through and then there's like a third reason no one really talks about but that's sort of uh regulatory Arbitrage where if you have a decentralized sequencer all of a sudden it looks a lot better from a regulation standpoint because it could be looked at as a money transmitter otherwise so I think those are the reasons why you would decentralize a sequencer I hope that was a good overview no that was perfect that was perfect so that's uh that's your crash course prerequisite to an excellent interview uh with Josh and Ben from Astra and espresso respectively all right everyone we got an awesome episode coming up with Ben from espresso and

Josh from Austria to talk all things uh decentralized sequencing in the modular stack really appreciate you guys coming on this is a Hot Topic right now I was hoping uh and maybe Josh we can start with you I just kind of want to get your overall vision for what you think the modular stack looks like and then also give an overview of your project and and how it differentiates from others yeah great so I need like kind of like a high level overview I guess it's like broadly in the sense of like we think there should be like many more Roll-Ups my view is kind of on this is part of scaling blockchains and there's really two ways you can scale blockchains you can have like larger faster blocks you know and or um or you can have layered and or like sharded chains then we view Roll-Ups as

like a design for how you can do sharded chain specifically like the benefit of like a roll-up design it's like it's more flexible and how you could do the sharding it doesn't kind of box you into like one kind of sharding mechanism as a whole um on our projects or project like Austria um we're building a shared sequencer Network I don't know with the purpose of like a shared sequencer network is that you can have this decentralized lock production a sequencing layer of a roll up but it can be utilized by one or more Roll-Ups rather than each roll of having its own kind of decentralization process how about you Ben yeah well uh I mean Josh gave a good overview of Roll-Ups as the scalability solution um which are very similar to sharding do

fragment the interoperability and the the shared state of the blockchain ecosystem and so the idea of building what espresso is building and what I understand Astra is building two is to get all these different Roll-Ups to regain interoperability and also not only share security and decentralization um you know of the uh um of of at least the availability of data as well as the ordering of data that gets processed by each of these different Roll-Ups and allowing each roll up to only focus on charting just the computation aspect of state right so every roll-up is at taking advantage of the heterogeneity between

approvers who are for a given role of proving the results of computations and Computing updated states and proving that for the rest of the you know the blockchain world to be able to verify um but not but not compromising on the decentralization of the actual you know ordering and availability of data among all these different Roll-Ups as well as the interoperability and when we think about where we are in we most people think about roles today right there the mind's usually gen generally come to ethereum and it's l2s um and so if we look at the reality of those robes today we find that they're mostly it's siled and centralized you briefly just walk us through uh the role

of a centralized sequencer in the ethereum L2 landscape today um and you know is this just a phase in the evolution of l2s or is this like a a logical edit state in some certain cases maybe a Josh I thought this will open to you first yeah so I mean so I think it's a phase in that like every existing like L2 right that like like prominent right it's obviously like publicly signal that they intend to decentralize kind of all aspects of it kind of over time right um you know I think optimism at most explicitly kind of listed like like a road map of like the different things they need to do including you know how you're going to decentralized like management of like the contract on like the L1 chain at both decentralizing the sequencer uh what are the things that

and kind of like one of our motivations here is that we think again to the idea there should be more roll up than there are now now we have like a handful you know five to ten roll up uh if we want to have more we don't want to continually have this phase to replay okay I'm gonna go create my new roll up I'm gonna stand it up I'm gonna have a centralized sequencer and then I'm going to start you know the clock and the effort on you know establishing a decentralized set of sequencers for every single roll up right we've used this is just fundamentally kind of unscalable um kind of like human coordination work and that there is a design you do this shared sequencer design um that allows Roll-Ups to kind of start be decentralized at like the default state for when they come into existence

so I would just add that it's I think it's not only about more and more Roll-Ups coming into existence and having to um you know go through this process on their own but I think even if we just focus on the Roll-Ups that exists today no one has fully decentralized their sequencers and no Roll-Ups are interoperable with each other and even if there's just two Roll-Ups in existence like that's both of those are important things to to solve um I think that when it comes to talking about uh the role of a centralized sequencer it's important to like recognize okay well where do we start with just what roles were supposed to be to begin with the robes don't need a sequencer they could just use the layer

one ethereum for ordering of data and availability of data so like the original concept of a roll-up was really just well we have a proof system and the proof system proves to to ethereum um what's the result of execution so now ethereum doesn't have to execute and they just do the ordering and and the availability um the problem is that even when it just comes to ordering data and making it available ethereum is inefficient and expensive and so uh naturally every project that was building a roll up found it more convenient and cheaper to run its own centralized sequencer so there is like this fundamental problem to solve of how do we how do we sort of have our cake and eat

it too right like how do we get this efficiency that we want but without um without relying entirely on on the L1 for ordering availability data which is still you know uh too slow and too expensive like a good point right if like not just focusing on like new you know theoretical roll up to the existing Roll-Ups and you know one of the things that like you know coming out of like the fluffy ecosystem and like using like the modular terminology very heavily we think about modularity of like the this kind of economic specialization right like different kind of like projects entities should focus on like one thing and right now you know the role if they're having to focus on many many things right and they say yeah we'll do the decentralization thing but like they have a very long list of things to do

before that is like the kind of like initial priority for them which kind of create room for like projects like myself and Ben's who are like we're gonna do that and we can provide that as a service to you uh because it's not your focus but it could be our Focus hey everyone big announcement from the blockworks podcast Network we're launching a new show called Lightspeed and hiring Two Hosts to to come build it with us lightsweet is a show for Builders focused on the use cases that will onboard the next generation of crypto users by taking learnings and inspiration from the garage days of Silicon Valley we really want to capture the perspective from Builders because that's what the ethos of crypto is content experimentation and Relentless Innovation to build products that users can't resist if hosting a show like this sounds exciting to you then head over to

the careers page on blockworks.com which we have Linked In the show notes you can also reach out to mirror Sam on Twitter to talk more about the opportunity but overall we're stoked about Lightspeed so if you think you'd be a great host please do not hesitate to reach out Josh do you mind going into the actual like kind of design choices you guys are thinking about and and how it differs from from band and espresso yeah so I think like like I don't want to like you know speak for Ben here too much or I'm like design differences and I think you know after as a project had like shared less kind of technical details probably our code is public right but we don't have like a thorough write-up like the espresso guys do on like here is like a top to bottom like what we're working on

um I think something like the the key differences might be kind of we're coming out of this like Celestial ecosystem right and so or default the some shooters that we're using Celestia as the data availability layer and we're focusing more on kind of what you know the Lefty term Sovereign Roll-Ups wherein the Roll-Ups um kind of like definitional like ordering uh is up to the roll up itself it is not you know with any dependency on this L1 smart contract or this kind of you know uh enshrined bridge between the L1 and the L2 um so I think that's like one of the kind of core differences we're looking at you know I take the view that I think was kind of promoted by like like Kelvin at optimism and then the class reported a very good follow-on like block summarizing of like the existing kind of

like first passive roll-up design kind of bundle of the bridging added the like secret thing batching block production element of it together potentially like too closely not in that like the like implementation was like incorrect edge of that like the concept for like a Layman became assumed that those are like inherently Concepts that must go together and coming from the celestial side where selectric doesn't have this like you know trust minimized bridge between it because Celestia just does the data availability portion of it requires us to frame things differently and we're doing like the block production the sequencing bit the ordering and then we batch it down to a data layer and then the kind of you know bridge between you know that roll up and any other chain whether it's another

roll up within the same d8 here whether it's another chain within a different ecosystem that is like an orthogonal component of the structure there yeah uh in some ways well in fact bridging is a form of like client verification right so uh all the bridge is is putting a light client on one chain for another but the the the requirements of of a bridge uh for like climb verification even just for Bridges between Roll-Ups to verify there's different there's two different types of bridges we can talk about right there's there's there's bridging in terms of verifying one chain needs to verify the consensus of another chain so if ethereum's smart contract needs to verify like the state of the

decentralized sequencer of a roll-up that's um that's like client verification of the consensus protocol then there's also like client verification between um uh between an ethereum smart contract and the actual state of the roll-up so verifying the result of execution verifying either fraud proofs or being able to check fraud proofs or being able to uh to verify zero knowledge proofs uh and that's something that you also may want to do between different blow-ups when you uh enable interoperability uh so I I I I certainly have a similar modular view of the role I wouldn't say like the uh the the bridging is monolithic with everything else that a sequencer is doing I see it as sort of a

special case of like client verification uh once you design a light client then you are flexible in terms of how you can Bridge whether you do Sovereign Roll-Ups or non-sovereign Roll-Ups Etc so and almost be fair to say that like sharing sequencing is basically offering this like decentralization as a service saying hey we're going to decentralize your sequencer while you go on and build all the other things that you're excited about building and then that kind of gives us the opportunity uh to only focus on this one thing and so to do that you know we're going to go out and basically create the the network of node operators if you will that are going to manage these sequencers on their behalf I'm just wondering what is the hardware requirements to participate in a shared sequencer Network like is this something

that's going to be like a high barrier to entry and almost like create um you know like I'm just trying to think about like Pro from that perspective right like is it going to be hard to participate in this network it's basically the question here very easy to like run a sequence ignored in a shared sequencing Network because it is not doing computation right so like your computational cost will be like very low research I think they're like a 2B determined question um and in bed I'm assuming you have like a like a good perspective on the space that you guys work on like hot shot and have like performance stuff you've done there and tarun I think if the only person I know that's mentioned this but like uh what are the thoughts around like is this going to be like a high bandwidth thing right what we see with

like arbitrary and like optimism sequencers the centralized right so they're kind of getting hammered by like user transaction kind of open question of if this is similar to what we thought with like priority gas options and whether you know a kind of like proper nav you know off-chain VBS or invert called PBS kind of Market playful results out of that but I do think there is like a likelihood that you know if you get a large amount of adoption of a shared sequencer Network then you know given nodes within it will have a relatively large kind of inbound of transaction through them and might have to at least you know manage like one of them cool between them um and how Side Band would you call computational requirements there uh but I think it's still like an unanswered question right now so the way I think about it is we need

to leverage heterogeneity in the network so roll ups are perfect example of this to begin with I mean when when so before we've been talking about participation and costs and running a sequencer let's recognize that producing proofs for a roll-up is like the most expensive task that anyone is going to be doing for any for any blockchain ecosystem um and the idea there is that there are there are going to be a few specialized nodes that are producing these proofs and that's fine that's perfectly fine because they're doing the work that's now enabling many other nodes in the network to do much less work in order to verify things the fact that proving is centralized does that mean the whole system is

centralized I think you need to take a more nuanced view of what it means for something to be centralized or decentralized and focus on what are the goals right can we design the system to be censorship resistant or credibly neutral right can we design the system to exhibit anti-monopolistic behavior in terms of pricing and that is possible within a heterogeneous system um if the provers are are are working really hard to produce proofs but they aren't in that they aren't playing a very privileged role or they're like the only person who can prove then if provers start to go Rogue and start censoring um or start you know jacking up prices then it creates room for other approvers to come in it's a competitive role to play in the network and it there are

similar things that you can do when it comes to bandwidth so in the espresso sequencer design we leverage optimistic responsiveness of the hot stuff or hot shot protocol that we have which is an extension of hot stuff in order to to to to to to to take advantage of a few nodes which have a lot of bandwidth available to assist other nodes that have less bandwidth available and it's a it's a very similar Dynamic to proving and verifying in a real obstetric ecosystem now a question I keep getting hung up on and I actually asked this to Nick from Celestia the other week but why would an L2 want to use an external sequencer when that's where a lot of

people think the main value accrual will come for their token and Nick seems to think that you know we ultimately just want to build tools for developers and let it be as easy as it as it possibly can be to spin up a roll-up would you say you share that same view or are you slightly worried that um you know you open source your your shared sequencer software that you work on over the next year year and a half and then maybe another L2 comes along and is like hey that's a a great idea like thanks for doing the work for us but now we're going to take that and apply it to our token yeah I mean anyone can Fork ethereum anyone can Fork Bitcoin right the the idea is that there is a strong Network effect of having all roll ups share a common sequence there's

not about the code it's not about the algorithm itself the whole point is right I mean all these protocols they build on top of other protocols the the the the value is in everyone sharing a common protocol that's doing the right thing to enable what they need right and that is uh that is the benefit of shared Secrets right that's the benefit of what we're developing that's the benefit of Josh has developing it's not as much in the the algorithm that we're publishing it's it's in the fact that we're trying to get a lot of Roll-Ups to share this common infrastructure um and there are clear benefits to Roll-Ups that run on a shared sequencer that uh that give them advantages over robes that are not

um of course there's a concern of you know uh incentives right there's a concern of well will will roll ups running on a shared sequencer be able to see as much value accrual as they would have on their own and that problem needs to be addressed through uh economic mechanism designs and allocation of uh of value that's being generated by by the sequencer as long as each roll up has the governance to be able to decide whether or not to continue using the sequencer then it has leverage in in deciding whether to opt into to a fair protocol or not yeah like like Ben said you know not not in that responsible in the previous one right mentioning like we want to minimize like monopolistic like actors within the system here right

and and so to some degree I think there's this inherent trade-off like I guess from like a venture style return portfolio where like okay if you want to be like a Google like how did you like make money you're like well you made like a network effects and then you became like the centralized point of like in Ingress for like the intra essentially right and like that at the very like either valuable monopolistic position into like oh they have like an active like lawsuit against them and like abusing that position in that option Market selected or whatever right but like to some degree like the shared sequence are right like is trying to fill this kind of like definitionally like monopolistic role of like the block proposer within like a network right someone just have to pick what is the next block that we're all going to agree

upon we take you know stake weighted round robin leader you know selection mechanism and we say all right someone's going to take a turn and when it's their turn they have a monopoly for this whatever the block time is right but like we want to minimize their kind of ability to extract rent from that Monopoly and one of the kind of values of that right why we push kind of this idea of like we don't want to take away like chain sovereignty right um like we want to make it like relatively intentionally like easy if you're saying yeah okay if the sheriff's sequence their network is like extracting rent for you if you don't think you're getting your value like you it should be relatively easy for you to leave right because fundamentally we don't want to be like capture it right we're trying to move from this like web

2 world where you say I'm going to establish a thing I am going to like honeybot users and then I'm going to extract threat right like the relatively standard phased approach we like don't want that um to the point of like why would someone use a shared sequencer right you know I think a lot of people are really like undervaluing the like that the value of like having decentralization provided for you right uh everyone's like oh why would they use the thing I'm like look like no one has decentralized their theme yeah right and like they have plenty of capital and they just have not done it which implies it is not a wholly trivial problem we take a look at the cosmos ecosystem right and see how many chains are there you know order of like you know 100 to 200 checks many of them have different

decentralized validator steps you could go into the forums of all of these and you know find out what is the profitability of all of these but fundamentally they're the reason that the cosmos ecosystem has also tried to make you know a shared security mechanism because all the chains that say Hey I want to go make my own chain but it's really annoying to go establish a decentralized value that even with you know all part a hundred examples of how exactly you do that they say I don't want to do that right I would rather buy that as an off-the-shelf service um so I do think there's like a lot of kind of good indicators of like this being a desirable service and then the economics of like what you have to pay for that service you know idea really will be resolved it's just like a

standard Market mechanism right where we say yeah you're gonna have to pay whatever the cost is for these people to run this network and that will come out of your bottom line right but presumably you are adding value and you would have to pay you know something you know essentially that cost yourself um if you had to do that you know um without buying it as like an off-the-shelf service there are some interesting tweets uh about uh Coachella is I think it's I get I've given this analogy before I think it was at East Global there there there's an analogy between a music festival um which you know has a number of different bands that are coming so we can take Coachella which just happened this past weekend right um and how people are buying tickets

directly from the uh the the festival organizer and how does the festival decide this is a festival that happens every year how does it decide how much to pay the different bands that are showing up to play right um and and so it it's very much the same kind of economic allocation problem um you know Taylor Swift can uh make 10 million a night uh on our own stadium so if you want Taylor Swift to show up to participate in Coachella you're going to have to make sure that she's incentivized to do that it would not be the same for other bands that are um you know not on the first stage are are generating much less and so the challenge mainly becomes how do you know how much a roll-up would make on its own uh and what's pretty cool about

blockchains are the it probably even more so than music festivals it's fairly uh possible to uh to look based on the transactions whether transactions are just hitting a particular roll-up what are cross-roll of transactions can we simulate what you know what makes sense in a transparent way in terms of how much each roll up would make on its own and therefore make sure that they get that and divide the Surplus among the other Roll-Ups who are now participating in the shared ecosystem of the sequencer itself experience a little bit I think they've envoc you know we've seen discussion around like unifoth and like you know deck the aggregators and general transaction flows you know more broadly within like the ethereum ecosystem right you know even flop launched a wallet last week I think or

like it got off of like you know the App Store jail or whatever like there's this question of like okay well like at what point is uniswap gonna make an argument that like Hey we're like 80 or whatever of like the SWAT Holly and like ethereum does uniswot believe they're like you know extracting sufficient value or like getting their like fair price out of like you life in ethereum is like a shared network right you know today they haven't seemed to like want to kind of like go elsewhere or like I don't know how seriously they've pursued saying oh we want like a uniform chain or whatever I think it was uh someone from nascent um I'm not going to remember the name but someone for like Nathan had like a post on like the inevitability of like udj or whatever right so like we can see

this kind of up and down the valleys back and fundamentally this comes down to like you know an economic evaluation of like you know I am buying a service that service cost me money could I go Implement that service and vertically integrate myself and get you know not pay whatever the marginal kind of like um cost for like the buying that service is but like it's a very non-trivial amount of effort you know we analogize this to like cloud services right if you were like um you know say like Facebook and you were around before like Cloud came up and you're like ah we have all these data centers everywhere we kind of know how to do that we already paid for the expertise to have that we're not going to pay the marginal cost to view the cloud service but you show up like after

clouds you say okay I'm gonna go you know learn all the expertise and like staff the people to know how to run that myself or I could just say yeah okay I'm paying Amazon margin like you know what that's fine I will do that because it is let work for me to bootstrap that's really what we see of like directionally things going yeah I could do this myself like you could do everything yourself but there is a reason people do not yeah no that that makes a ton of sense and I love both of those analogies I want to kind of Zoom back out for a second and talk more about the benefits of what you the atomic composability you get using a shared sequencer and specifically how that differs from two ropes they use the same d a layer as opposed to two roles

that also use the same shared sequencer uh maybe event I'll toss this one to you first if you could kind of just really go down the uh the rabbit wall of atomicity and why that's important for roll-ups yeah no it's interesting that you asked me to compare Dia like sharing the same da versus sharing the same sequencer so if you share the same data availability layer then you're simply relying on the same security assumption and for your data to remain available I guess that has benefits in terms of um you ensure that as long as the data for you know one roll-up is available then it's also available for the other right if one is not available the other is not available availability is

certainly something that you do want to have for it's a necessary component of interoperability but it certainly doesn't give you interoperability on its own um by the way sequencing and sharing a sequencer is necessary for interoperability it doesn't get you there on its own um but it gets you a lot closer and there's a lot more that you get from sharing the same sequencer than just sharing the same d8 layer so what is sharing the same sequence we give you well it gives you the guarantee that if the transaction on one roll up is going to be can be included uh you know if and only if the transaction on another roll-up is included so they get sequenced together

meaning that they will get executed together well that doesn't tell you is what's the result of executing these two different transactions right for that you need to have you know let's say we're trying to do a simple bridge where um want to wrap eth from one roll up to another so the youth gets locked on one row on on one roll up and then minted on the other you need to have a smart contract on each side that implements that right just having sharing a sequencer doesn't immediately give you bridging you need to actually build the bridge and you need to have a cross chain met across roll up in this case message which signifies to one roll-up that the East was locked on the other roll up um of course it's possible if you have a

bridge transaction which is locking and then minting for the lock to fail because it was front run by some other transaction so you would want to implement some weighting mechanism too um the benefit of having a shared sequencer is that uh well a number of different things first of all you um you you mitigate a lot of things that can go wrong if things get sequenced at very different times in the different Roll-Ups even things that could be taken advantage of if if you imagine across uh across roll up swap where somebody's trying to uh if someone agrees on a price uh with another uh counterparty on a different roll-up and they're trying

to exchange an asset then if one person gets their side of the transaction sequence first maybe it's going to wait for the other transaction but now the person can wait for a long time to decide whether they want to still take this price or not whereas if they're sharing the same sequencer they don't have that option optionality right right they could like do something premeditatively to prevent swap from going through to begin with but they can't decide whether or not to take the price if you know uh for for a long time which can have a huge advantage to them uh given how markets work um the other thing that this enables is if you are relying on Builders uh to do much more complicated things around intents uh of users or a complex

um some complex form of interoperability not just making two operations atomically dependent but saying I only want to I only want to go this to go through if so and so conditions are satisfied um maybe it's dependent on on the price and you just saw maybe it's dependent on the time of day there can be many different complicated maybe maybe there's a user who just simply wants to get the best price available at the market at the time across many different dexes right this is difficult to do without some kind of block Builder that's that that has the ability to to guarantee some kind of result um and the block Builder can put up um a um a a a collateral or can set up there can be some kind of incentive

mechanism if that so that the the block Builder gets slashed if it doesn't honor the user's intent right and that becomes much much easier if it only has to to deal with one sequencer which can implement this then if it's talking to 10 different Roll-Ups with each of their own sequencer maybe it will get slashed on one because one of the other Roll-Ups messed up and that wouldn't be so great so um shared sequencing greatly simplifies the interaction that is necessary with block Builders to uh to satisfy more complex user preferences than intense you know I agree I agree with everything I'm actually glad you sent that to Ben I think he we've talked about this previously and I think Ben has like a better kind of like like handle on like

wording this more simply but the key distinction I think between the share of sequencer and what I think people would think about like normal like attack composability right is that like it is guaranteeing the like inclusion of two transactions and like if and only if state but it is not guaranteeing an end State itself so you can't get like that super strong guarantee but only do this if like your victim's not for like an Arbitrage or like a swap case right you can't say like only execute visit I make five dollars you can't get that gear if you can say like only submit these two transactions if they're together as like a bundle and then to bend this point right on like the Builder Marketplace like this is how we're seeing I don't

know what like 90 95 percent of like transactions submitted to ethereum now somewhere in that realm right um I think like it's really really important to think about like how we're structuring that system and the idea that this share shared sequencer creates this like this single marketplace right you know there's a lot of these worries of like oh aren't you like giving up the nav and I'm like yeah but to some degree right like at some steady state you're gonna have these kind of sophisticated actors doing this building the stimulation the block production right for all of the Rolex and presumably if there is value on three different Roll-Ups it is roughly the same parties that are like interacting with all these different Roll-Ups right and there is some benefit to Jeff thing we're going to coordinate them at like this Market that is intentionally kind of designed

to handle this this is one of the big reasons why S3 is coming out of left yet and going back to the point I'm like this kind of unit you don't need a sequencer at all you could just submit it to like the base layer like again go and collect the specialization like Celestia is not you know as a layer prioritize uh or like prioritizing like handling in like an nav Market at the sequencing layer you're like all right one of the first questions you have to kind of think about is like how are we handling the Mev Builders and being designed kind of like up front to handle that system can you know work quite a bit better than saying like oh this is like an annoying emergent behavior that our system is not designed I think just to like summarize in one sentence from

the perspective of a block Builder a shared sequencer can give it the guarantee that a block it's submitting satisfying some user intents and preferences will either succeed or not this is not the guarantee that is given to the block Builder when it has to work with multiple Roll-Ups using different sequencers it's possible for one leg of the block to succeed in one leg to fail and and that even if the block Builder did nothing wrong yet move the thing from like probabilistic negotiation with multiple parties to like they're the single party that like the block Builder is negotiating with if it comes to a deal it's like that deal was like complete rather than saying I have multiple legs for this I have to negotiate and if

there is it because this fundamental security guarantee right like whether you could get a commitment and agreement on these two things right at the same kind or whether you have two parties you can never quite guarantee if you've committed to one and the other one will like bail at the last moment so you mentioned an interesting point there Josh and and near well your responses that brought up the point of Med and if you're a a role with a vibrant economic vibrant Market of economic activity happening on top of it let's say particularly in D5 where a lot of MVD occurs what what is your what is your incentive to say Hey you know we're gonna forego this this ordering of transactions to extract WB or participate in MVP at all pass that on to the centralized

sequencer excuse me the the shared sequencer decentralized sequencer layer um we're gonna pass it on to you and allow you to go through this like does that seem like a realistic trade-off uh for an L2 say Hey you know we do get this simplified process or this decentralization as a service however we do forego uh all this MVP and you know if we look at fast forward out a couple years and think that uh you know Waltz will have this Market that's billions of dollars of MVP extract it uh from ethereum so you know if we do look out into that end state of Roll-Ups and say there is this this Market you know they're they're basically foregoing the ability to uh recirculate this this billions of dollars worth of Med throughout their own ecosystem is that like a a fair assumption to say like you

know the trade-offs are there like this is something that would make sense enough no actually so the the the the Roll-Ups are not uh they're they're not giving up something just the point of using a shared sequencer is not that the sequencer is providing them with decentralization that they wouldn't have otherwise had to build on their own they're using the shared sequencer for the benefit of that and additionally the benefit of sharing security with all the other Roll-Ups and additionally the interoperability with other Roll-Ups which as a part of it provides across domain Mev and one way to think about mevs like there are good forms of MVP bad forms of mv2 but

the the ability for users to now take advantage of cross chain Arbitrage for example means that now all the Roll-Ups are more attractive for all users right it increases the because you're expanding the overall you know realm of of possible transactions that can occur you are increasing the overall Mev right from both both positive and things that you might want to mitigate using other mechanisms and and so now again it just comes back to an allocation question overall the system is generating more value than before the uh the the total value is greater than the sum of the individual Parts how do we allocate

value across the Roll-Ups that are participating in ecosystem so that it makes sense for them to use the shared sequencer rather than running their own centralized sequencers or decentralized sequencers for that part and that too can can be handled if with a proper economic mechanism it becomes harder because now it's not simple you know uh gas pricing we'll just look at where let's look at estimating the demand for for gas in each roll up and come up with some curves and figure out how much each roll up should be allocated no it becomes more complicated when you take into account Mev but it theoretically is still just a solvable yeah I agree with all of that I think

there's another interesting thing towards like you know the the phasedness or the state of like Roll-Ups right like like right now what role-ups are not capturing any V like like so they say right like we have Central I think one certainly the assumption is it like they are not like reordering transactions on that if we go to know the details of like the visibility into like the decision making of that and whether someone has like a you know here I try to submit you know transaction for like the sequencer and the ordering preferences whatever um there's also like an ideological perspective here right like not going like two in the weeds here right but like you know arbitrum very obviously has like an opinionated take on like nav and what they believe to be a more correct or Fair distribution right I I fall into the the Tyrone Camp

of like Fair ordering is like fair is not a real thing like that you just you just picked a different like preference thing and you've decided that like Whoever likes fancy faster is going to be like a more preferred party um I prefer like auction mechanisms right um but like so there's something like they're not capturing it right now right do they think that like a off-chain you know PDF Marketplace which at least is like what I am considering as like the designs paid for how share sequencer were handled like MPV talking to like both the skip team the flashbots team um about these things like there is like an ideological actually arbitrine would probably not use a shared sequencer that had like an ideologically distinct perspective on how like nav and transaction ordering generally should be done um but like I generally think they're on

like the losing side of like an intellectual debate um and then to the point of like you know you Ben Fisher like where um like is there going to be more value on the system like yes right like maybe it's not that huge right but like there should be like more right there is some quantity of like cross chain atomicity and that is like a defined better thing um and again I think like like Benny's point is like really valuable like they're not like that's not what they're like buying right like they're buying all these other features and like part of that if they have to have this like economic discussion about like what are the trade-offs of that but like it's not like they're saying hey we have this inbound Revenue stream and if we go use this system we're losing that inbound no

one's collecting that Revenue stream right now and then that Revenue screen will still be collected by again to my earlier like presumably the same kind of off-chain actors that are doing it on you know ethereum on arbitrum on optimism right you have just made a centralized Market such that they can participate in this market rather than them saying okay I have to go have like my arbitrun bot and my optimism bot and my ethereum bot they can say I have you know my Asteria bot and my espresso bot and that covers like a broader space of things uh and that is beneficial for like you know certainly this like long tail of her life to say like hey I want to have like block production gives good composability guarantees you're like all right great you don't have to like stand up your chain go get your decentralized

that and then maybe try to incentivize the existence of some kind of beneficial you know sequencer market player uh like Builder Marketplace off train you can say this one provides all of that as like this kind of intentionally designed space as a service that was a a really great overview um I just wanted to add one thing there though which is that I think the sequencer shared sequencer while it does need to be aware of the mechanism for allocating whatever revenue or value gets generated in the Builder Market um it can be agnostic to how we approach Mev how different Roll-Ups decide to approach Mev whether it's um using privacy preserving Solutions

like this uh like a threshold decryption solutions to prevent um uh to prevent uh content aware ordering um whether it's some Fair sequencing protocol or whether it's something else whether it's an auction mechanism the sequencer can be agnostic to that and proposer Builder separation in some ways unifies all of these things Pros or build a separation says there's some Builder it's submitting an input as a block and the sequencer is just aware of how much value it's allocating and then how that's getting allocated among all the parties but whether this Builder is flashbots running an auction or it's a multi-party computation among a whole bunch of different nodes running a fair sequencing protocol or some threshold

decryption protocol is really you know completely pluggable and there can be many different versions of this and rollups can choose to use something for building that other Roll-Ups don't use for building and that will impact their interoperability with other Roll-Ups but it won't impact their ability to use the shares so it sounds to me like the main benefit here is ultimately going to be growing the pie and increasing interoperability and you know being Atomic with other Roll-Ups so do you guys not view this also as Like a Winner Takes most Market I'm curious how you guys are thinking about this friendly competition in a way between two you know projects uh I guess do you think the the room is is large enough for multiple shared sequencing designs

or do you think it'll turn into uh I guess unifying on one standard I have a very positive some Outlook and I think that's something that's also characteristic of this industry in general and uh I think we're very new on all of these you know technological developments so I I you know that's just generally my Outlook um but predicting the the you know the the long-term economics of how in what is the study state after a long time it's you know hard to know um I imagine there's certainly room for review but yes fundamentally um shared sequencing just like

blockchains are you know there aren't it isn't supposed to be interestingly many of them uh the more you have than the more you fragment the interoperability uh in some ways what's happened with Roll-Ups on top of blockchains right yeah I agree with that right like it's obviously like like there are network effects right it's like a network like anything else there is like Network effects right um I take the B1 of like there's a thing it's like a category of one right like if you made a thing and you're trying to like we're trying to make a thing and like no one else is also trying to make the thing like maybe be a little suspicious that you're like you're making a thing no one else thinks is like valuable and you have maybe gotten into your own head a little bit much so I think it's valuable to have this from

like you know a less kind of like I don't know like like business Centric View and more of like the altruistic view right like we it's generally better if there are multiple options than like a market because you know you don't want there to be like just one that kind of runs away from it with it and said okay now we have like 95 Orchid share and the burden of like even entering that market is like near infinite and so now you can kind of start to do this kind of value extractive stuff right like from login like the perspective yeah the investors would probably like you to own like 100 of the market share of whatever you're doing from like an end user perspective right like that's not desirable you want to have options um and just from like a general like Market positioning thing like you know

if if like you know after and like espresso or like one and two and like a valuable you know category however we flip that that seems fine right like Coke and capacity both exist and they both seem to be doing quite well and I think the main goal for both Josh and I at this stage is to convince the world that shared sequencing is something that's important that people need to use um and it's great to have Josh here who's a partner in doing that even though Josh is building Austria and I'm building espresso we're taking slightly different approaches um the uh we'll we'll both do much better because we have a unified effort in convincing the world to use shared sequencing to begin with yeah the idea of a rising tide lifting all boats as you mentioned Ben really is

something that this industry continually Rings home uh and it is kind of like encouraging to continue working this issue as well but I want to ask you specifically then about you and the team at espresso have written uh some pretty interesting posts that kind of talk about what are the alignments of economic activity need to be between the L1 and the L2 just given that there are you know there is significant fee capture going on at the L2 and this is kind of I've seen you guys talk about how uh the idea of like eigen layer or re-staking protocol yeah really the appropriate way to kind of like bring the D.A and like have the validator set of the L1 also kind of participating in the shared sequence I love teacher thoughts yes yes thanks for bringing that up because we haven't really

touched on that yet um and so this has to do with the idea that the the if the if the L2 is generating a lot of value that is not captured by the L1 it destabilizes security at the base layer of the L1 um if if if there's a a block that's paying a dollar to the validators of ethereum but it accounts for millions of dollars worth of economic activity on top then it's very easy to now bribe the validator in some off you know uh off-chain Market mechanism to uh to Fork the block or to do something that's not consistent with the validators on this role in the ethereum protocol so you do want to give the L1 exposure and by decentralizing a sequencer you already

do that the worst thing is to have a centralized sequencer that is a permission set of one that the validators of ethereum cannot participate in even if they wanted to by decentralizing a sequencer now you at least give validators the ability to opt into participating in that protocol and thus do get exposure additional exposure to these transactions by participating um Marie staking is a way of subsidizing that so it says oh well you already have you know uh 32 East State for ethereum how about you just reuse that 32 East you don't have to have any new capital Investments to participate in the protocol that's running the Sci-Fi sequencer you can just uh subject that

32 weeks that was locked to slashing conditions if you violate the rules of the sequencing protocol and so it effectively subsidizes all the Assyrian validators to now participate in with no additional capital investment this centralized protocol um you wouldn't want it to I I think it's we've written about this and I think we're still you know going through the designs with eigenlayer and other um researchers out there I think it's risky to build a protocol that's entirely dependent on E3 staking uh especially considering that at the beginning of time the amount of value being generated by uh by by by sequencing for for the role ecosystem

may be low compared to sequencing for ease so you just may not get that much participation at the start you may get one node from from ethereum to restake and now it's completely controlling the entire sequencer uh so you it's beneficial to also have some form of uh Native staking uh that is separate from ethereum staking but the exact design of dual staking as we call it and how you you can have both native staking and E3 staking interact and how that changes over time is still something that we are working through the details of it is a very interesting research problem that we invite everyone to think about in fact Josh I'm curious as you know you guys aren't uh planning at least to my

knowledge a shared restaking design so I guess why did you make that design Choice as opposed to Espresso and the restaking module yeah so we've thought about it not nearly as in-depth as the espresso you know guys um but like we considered it there's like some like structural reasons just like you know we are based around like tender men for like at least like you know you currently were like evaluating whether we're kind of committing to like Alternatives but like they feel like they're structural like how do I just grab a chain by having like a restate token on like presumably for us it would be like the left jump and like what is the design space for that you know stride has some stuff like that but it gives it kind of like technically complicated and like okay you have to get this top in you have to get it over

here like how does that work from like a Genesis event stuff like that um more generally like I agree with like everything bad and sad right like they're like you know trade-offs in this design about getting like like attention uh especially I think for us for like Celestia where it's like a thread like there's um like Celestia is not like established right it's not even like it right um if we were to build on top of it like there becomes this weird kind of impact of like what if we are like you know so the invert right of like we are the dominant way that like transactions are passed through to the left what are the kind of mechanisms that happen there um and what are the implications we never like got committed too much thought to it right again because of

these kind of like technical reasons um but I also think you know and and I think like Ben and like the aguilarity more broadly of like you know aware of this right I have the like back of my mind just like general anxiety over like restaking as a mechanism uh generally because like crypto adds like a whole and kind of like you know the the DJI nature of like people within crypto and I'm not saying like Ben or like the espresso team is it but like something is going to explode because of restaking within the next like three years like someone's gonna do a thing that restates it gets some usage that like blows up that like you know destroys some relevant quantity of money and I just believe that that is just like the way crypto works like we will you know acknowledge that like once it happens so

there is a little bit of anxiety over like refaking and kind of like Rehabilitation like you're stacking risk on top of like a given a lot of stuff but we just have to find that equilibrium of kind of like um like economic I guess like with the opportunity cost right now of like ah you lock this up right like it's inefficient for like an opportunity cost you can you know re-hypothecate it and get more return on it but you've now added more risk to the Vista you know we're gonna do the kind of standard pendulum right we will fly into too far when we lose a bunch of money on it and then we will find some equilibrium in the middle have you all thought about what types of applications you expect to I guess launch with a decentralized sequencer as a roll-up off the rip like

for me in my head it just makes a lot of sense for like if you take uh Ronin for example the the side chain uh for a uh actually Infinity like it would make a lot of sense considering there's really not that much Mev that they're trying to capture uh they've got their own side chain it could be more secure as a roll-up and then they wouldn't have to worry about being a sequencer so it makes a lot of sense for gaming in my opinion so have you thought at all about that or talked to any projects interested in experimenting with the tech we have like a perspective here because like we are building like our own like you know evm to integrate with the shared sequencer kind of the origin of like after the project went all the way back to like how do we get an ebi on

top of Celestia we're continuing that effort you know we discovered this kind of concept of like a shared sequencer um kind of very similar timeline to like like when you know Ben and his team League started working on this right uh uh as I get an outcome of like we want to build an event oh we need the sequencing component okay the design for the sequencing component is actually relatively generic okay we could use this at like multiple places uh but we're still building our EDM there's a couple reasons for that like when we think like there needs to be like an EDM within the Celestia ecosystem there is not currently you know a team you know other than us kind of like heavily committed uh to that effort um and that also we want to like dog food the shared sequencer from like a

design perspective right of like how do you know the interface with which or roll up you know uh integrate with the sheriff's they put very good you're like all right well like go you know use it and see if it is like good or whether it is painful thrilling which applications um I don't think about like really like application but one of the things I think is interesting is talking to various teams like you know they're called error guys like we're all the service stuff you know Argus Labs guide the bearer chain guys you know they're all doing kind of like evm-ish stuff but they see kind of like beautiful oh they're like their customers were Argus and like universe and themselves they want to use the ebm and then add additional functionality to it within like ebm it's like an L1 world right like getting a pre-compile kind of

accepted is like like a Monumental task right because some degree where people will do it's just like impossible or it's not really worth your effort if that's gonna be like a differentiating feature but in these Roll-Ups we can see these datives again don't afford like you're gaining Argus is like a gaming specific you know chain mechanism framework whatever um but like four people who they I want the EDM at the base or I want this chain of the base and I want to be able to modify it have a different date machine we say that to be a very good fit um because right now if you want to go launch a new chain it'd look very very expensive to do the decentralization effort um a lot of the kind of like Roblox stuff you know vop stack is great it's kind of bad at the box but right now

like it is just like EDM right but we do think there's like a lot of room for experimentation by reducing the kind of barrier of Entry to think hey I want to go launch my own chain and I want to do something different and you can still get this kind of censorship resistance block production um without doing it yourself and you need to pay the market rate I I agree with with everything that Josh said but I'll also add that um I I think it's not just app specific Roll-Ups um the the the generic roll up uh product space is very very competitive there are many different projects out there and they all need users uh you don't have any Meb

without users right and I think that forming alliances sharing a sequencer increasing interoperability is going to become a uh extraordinarily you know advantageous move for some of these Roll-Ups and I do think that we're going to see Roll-Ups uh partnering and running on shared sequencers in the near term um we don't have anyone officially doing that yet with the espresso sequencer it's still in development but I I am personally very optimistic that some of the major generic Roll-Ups that um you know are looking to capture value even from muv will find it beneficial to

share a sequencer and figure out an economic mechanism for sharing that value another question that just pumped in my head uh Ben was you know is espresso looking to be VM agnostic media roll-up sets it on top of it I could kind of use any of the VMS today or create their own is that's not something to be theoretically possible yes uh so in our testnet currently we as a proof of concept by running polygon Hermes um but we are looking to support uh the op stack we're looking to support ZK saying we're designing the espressos sequencer to be completely diamagnostic and to support either ZK or optimistic Roll-Ups yeah no that's that's exciting

as well and you know we spoke to prevent that Preston Evans a couple weeks ago building up Sovereign ropes as well so hearing the integration with ZK as well is is also exciting especially given the launch of the ZK VM and the popularization of that but yeah no that's that's also I do want to point out one caveat though uh for optimistic rollups which is that when too optimistic if if two Roll-Ups want to have a bridge then um it becomes much I mean ZK is very very helpful for bridging right because it gives you it can give you instant um confirmation of cross roll up messages whereas if you are entirely relying on fraud proofs for the mechanism of verifying the state of a

different roll-up then uh you're fundamentally going to incur a very long latency um so there is I think we're going to start to see ZK being integrated into optimistic rollups through the bridging between different Roll-Ups and ultimately a full transition to zko only ZK VMS in the long run that's my personal belief um I don't have any problem with optimistic Roll-Ups or VMS today but I do think that they'll as a cryptographer I think the long run is good to have uh ZK rops rather than optimistic roles and we we share that for you you know we've done a bit of work with like optimistic rollups it's like part of like I used to like work on like the roll kit team at

Celestia and you know that we've developed you know some like prototype level like like fraud proof for like the cosmos fdk um and like a general thing but like you still inherit kind of like the complications and like delays of an optimistic roll-up Bridge which is just kind of like like inherently it's just like a poor either experience right like users right now are not thinking about like oh I want to bridge from here to here and I'll check on this like next week right like that is not the mental model of like a user base and so ZK role for all that there is also you know to the point of like like heterogeneity of kind of like many many Roll-Ups right like it is hard enough to like verify like an optimistic proof of like an evm to an evm it is like going to be

increasingly difficult to say like oh yeah I have a Cost Plus SDK you know VM and I want to verify you know a roll-up proof or an optimistic proof of that for upper that on like an ebm like that is like an additionally kind of complicated path we can see these zero knowledge proofs acting as essentially like an interoperability layer we actually see this in like the consensus stuff right with the polymer guys um you know like like VK IDC work right where they could say okay you can go verify um you know uh essentially like aggregated signature on evm from like I from like a tender and IBC chain whereas like it was not computationally feedable to like verify that curve on the evm same with the the sync the labs people right um who have done this between uh

ethereum and gnosis with their prototype but they're live on like seven chains or something like that now right again this kind of how you can verify this light client using music as like an interoperative so that that definitely looks like super interesting well we're already running on an hour and I could actually chat here guys zero for an additional hour but I don't want to take up your entire day so we'll have to do this again sometime in the future but thank you guys so much for coming on it's been a fantastic conversation uh Ben I can kick it over to you to share with people where they can learn more about espresso and yourself and then Josh all pass it over to you after that yeah absolutely but we have a website so you can look at espresso sys.com or espressosys.com we have links there to

all our blog posts we have a number of blog posts out there on hack MD we just released another one today on uh Crossroad interoperability through shared sequencing you can follow us on Twitter Express at espressosis you can follow me on Twitter at benefit so yes I would be excited if people were to check those out yeah for off we're at asteria.org a-s-t-r-i-a.org our blog iflog.astria.org or Twitter is Astria org all one word um you can follow me on Twitter it's jayskybo and j s k y v o w e n we're happy to talk to people we are like like

hiring so like you know reach out and us maybe don't harass us too much but like we are open you know we have like a lever page link on on our um website so by all means you know we're we're looking for you know engineers and whatnot um and yeah I would love you know more people to reach out awesome we'll catch you guys later