App-Specific Sequencing (ASS) & LVR Technical Debate with Yuki Yuminaga & Alex Nezbolin

Transcript

he's basically asking you to show Cella this is kind of The Best of Both Worlds if it's possible we'll take over Finance on day I think it's impossible to answer and basically we'll just have to wait and see welcome back to the rollup for another technical debate this time with Alex nesbin and Yuki yonga about the solution to onchain me and lvr we learned quite a bit in this one frankly we are me noobs at least I am and we talked about the concept of application specific sequencing also known as ass this is a really fun highly Technical and insightful conversation about the state of me the core idea is we need to figure out a way for

applications that generate demand for blog space to be rewarded with me we'll figure it out one day at a time One technical video at a time hope you guys enjoy this one and let's Jump Right In after a quick word from some of our awesome Partners welcome to the ignition Mantle's newest campaign to liberate the growth power of Bitcoin fbtc ignition has officially begun with a total token supply of over 2,000 fbtc with a tvl of over 150 million right now you can go on the fbtc website to see all of the featured campaigns saking opportunities and liquidity pools where you can utilize fbtc on chain today for some of

the best yields available on l2s the ignition Sparkle campaign can can be found at fbtc [Music] ongoing carte is your home for app specific rollups at the Forefront of developing new tools and platforms that enhance the capabilities of blockchain Technology carty enables applications that are flexible and built to scale build on a full Linux environment and use traditional software libraries but in blockchain development go to cy. that's

carte. to learn more about building your own app chain today what's going on guys welcome back to the rollup today we have a conversation that goes deep into the trenches of how blockchains should be built from the builders perspective um in terms of Revenue customizations and sovereignty uh um I think those are like the main three three components but I I know that we're going to touch on more we're going to be talking about kind of app specific sequencing lvr me and more of a deeply technical conversation today I hope to learn quite a few things I know I will so without any further Ado pass it over to our lovely guests for some quick intros and then we'll jump right in so

welcome in everybody um and yeah Yuki why don't we start with you yes um thank you for uh having me on uh this time and I'm Yuki and I'm founding research and uh growth at soral Labs um and I guess just a quick intro myself of um what what I like to do and what I'm interested in the crypto space in general um so yeah I'm personally you know quite fascinated by a lot of this uh meth problem that uh we are seeing today which is also the reason why um I you know I'm working uh with the S labs to build out uh those Solutions uh that that we see in the space such as ivr which we'll discuss more later on and um

yeah um quite quite bullish that we can hopefully come up with some great Solutions uh to to deal with this uh problem that is pling the uh the the liquidity space in general uh right now um I guess uh one thing that I'm personally just generally interested and hopefully seeing from the the the broader crypto space is that um uh hopefully onchain liquidity will uh will take over binance one day uh that's uh that's one one hope that I have it's kind of like the the North Star uh of what I'm hoping to see from defi in general yeah that's the uh infinex thesis when I talked to Kane in Brussels

we were like we got to get all the centralized exchange people over to the dexes Alex welcome in man it's your first time as well so would love an intro yes thank you for having me I'm Alex nlin I used to be an accounting professor at uh different Business Schools in the US and the London School of economics and then I got interested in uh different blockchain issues and specifically um one of the big ones is uh lvr and MF in in general and uh at that point I quit Academia and switched to crypto full-time so I'm now

working on a protocol um that is aiming to minimize some of these issues so are in particular um what I think attracted me to the space is that uh there are a lot of interesting questions that uh the community is facing right now uh some are definitely new but also there are some questions that people grappled with in traditional Finance for a long period of time and so I think it's the field is interesting in the sense that you can apply some existing knowledge to new questions and uh you can develop some new questions some new knowledge for uh

those questions and it can even give you a new angle on how things are done in traditional Finance so you might as well discover something new about the stuff that you think you already know phenomenal well let's dive in uh I I think uh first let's aluminate um this this General meev space as it as it exists today and and the subset of that uh in in lvr and and sort of you know unpack what lvr is to start with uh before we get into ass and application specific sequencing and potential other uh Solutions and alternatives to address uh lvr and me so Yuki why don't you kind of start you know talk us through your

your sort of perception of meev these days as it relates to lvr um and then we'll do the same with Alex yeah for sure um so actually uh to kind of like before uh starting off on uh sort of giving the overview on this I would also like to point to our uh website uh uh with the MEF dashboard that is enabled by our open source MEF tracing platform called Brony uh because on that you will basically uh see a lot of granular details on how mevs are being treated um where it's generated and where it's going and and and based on a lot of this data where I'm going to be talking about

um you know those overview from the way I see currently is that um in the market there is a large large amount of uh Atomic arbs that is being happening that is basically purely onchain uh which uh may include some arbs between different pools across different amm um that basically holds a significant amount of values that are being extracted from um the liquidity providers so far but at the same time there's also a large amount of values that amounts to like uh around like 200ish million dollars of uh 6X arbs that we have uh obser oberved by brones

that are also um existing on chain since the merge and um this all together total up to am around a little over 700 700 million dollars since the merge as well so if you kind of put that in context that's basically about like you know about like 300 million a year since the merge was like you know September 2022 it's been two years that's how much money has been extracted um by the by the by the sort like the Meb participants uh which also includes uh validators uh in in this pipeline because they are the one that are receiving uh Lion Share of those value as as a form of a a bid within PBS

auction as well and then um I think uh Robbie you mentioned about the lvr uh side of things as well um okay so so to kind of like also elaborate a little bit on the lbr side so lbr uh or or I guess uh as I would like to call it in a more generalized form uh 66 up charge that uh we are talking about here so um I would like to consider lvr uh and and those six6 arbit charage that is kind of resulting in this form of lvr as a as a sort of like a that a cost that is incurred incurred by LPS by by being passive mive and participating in amm

and and essentially eating up a lot of those like toxic flows that are trying to pick on their stale prices uh due to the block time of ethereum being a little bit too uh too long as in like it's a it's a 12 12 second block and the price updates uh much faster on binance than than ethereum so that creates a lot of opportunity for uh AIT Chargers especially Market making Market making firms those trading shops to be able to capture those value off the hands of LP and then put it into their own Pockets or maybe bit to to the to the block proposers uh meaning Batts so that's a that's lvr that's the sexxs that have

been captured uh currently by those participants off the hands of LPS yeah yeah so for the community lvr and I think this is the agreed upon term loss versus rebalancing um and and kind of when you go and you put your two assets into a LP I mean you know a lot of the people here have been around since kind of like the defi summer days and the early days of Unis swap preconcentrated liquidity now kind of moving into these different Dex models basically the the the the the muv that we're getting that that that these arbitrages are getting is those those LPS the swaps

that are happening there and that are settling when you're lping into these dexes the the price on chain is slower to update than it is on centralized exchanges and so that Gap allows highly sophisticated actors to basically take profit from what would be LP uh either like trading fees or from rebalancing and then they can basically extract that into their own profits by doing a trade on centralized exchange is that kind of like how it's happening or maybe you could illuminate a bit more as to like exactly what these actors are doing uh that is causing the problem yeah so your description was kind of like you know U

roughly in line but I I I do want to also emphasize though that uh you know uh lvrs and 66 OBS in general is one form of me that constitutes uh some chunks of the entire uh meth that is being extracted there is is however there's still like a pretty large sum of like Atomic arbs that are that are out there that maybe take about like 50% is of the entire like me space as well so just wanted to like keep that in mind um and and and in regards to like how exactly you know those folks are uh actually extracting and yeah as we said Andy like uh they will be essentially executing trades on both legs on on the

onchain side on the offchain side mean like the the Cent exchange like um both they will be trading and um they they were going to be picking on the the sort like the more favorable price on off chain and then uh essentially uh incurring not so favorable price to onchain entities like LP so LP would have to essentially sort like you know buy or sell uh to those uh uh trading shops at a price that they wouldn't do otherwise um and this kind of get into the discussion about markout that uh you know Alex have uh talked about quite a lot previously but um yeah so essentially LPS are constantly uh

being treated um as a sort like a you know uh uh incurring those toxic flow as well yeah perfect Alex could you kind of fill in any gaps here to this this lvr conversation um and sort of give us your perspective on on it yes sure so I think on the lever side or lvr side the thing that I'd like to mention is that um basically Yuki describe the idea very well right since block time is so long LPS end up quoting stale prices 12 second stale prices and so basically

then arbitragers come come in and they end up trading against that liity and if you think about what LPS are doing essentially with their money or what their contract is doing on behalf of the LPS is that it ends up uh buying to reach selling to cheap all the time and then there is a if you mark to the current market price at the time of the trade you end up making little losses every now and then but these little losses they accumulate over time so if you look at markout plots you'll see that arbitragers make money

slowly but steadily usually there can be some rapid increases also when there is high volatility but it's sort of like an up only slowly growing plot for arbitres now for lpce the total income statement will look differently depending on how you calculate it um and uh LPS can still make money even with Arbitrage flow alone and this is because uh the overall position of lpce also has some Delta exposure what we call Delta exposure right you expose to two assets and the exposure changes over

time depending on the algorithm and the algorithm is structured such that on certain price paths you are granted to make money specifically when the price returns to where you started as MLP you're still making money but then on other price paths you are losing money when the price deviates and doesn't come back uh but the nature of the game is such that arbitragers are always making money so what ends up happening is that you lose money in expectation because of lvr so lvr is sort of a running cost of doing business and you can still meure recover that cost again on certain price paths but some you want um but that's a cost and

expectation but then again there is also a broader perspective on things so if there is this cost and the game appears to be a negative expected value game for LPS why are they playing it well because there is this second big component to their income statement which is the uninformed flow sometimes people want to trade against the pool not because they arbitragers but just because they want to trade for one reason or another um and uh those trades can make money for LPS even if you mark them to the current market price so sometimes people are happy to pay a little bit to actually

make the transfer and to make the Swap and so in an Ideal World the am amount of money that you make from uninformed flow would be sufficient to offset the losses to arbitragers Arbiters anyway providing a useful service to you they're telling you where the price is but they charge you for that because what they're doing is that they notice that you are charging the wrong price and then they pick pick off the liquidity that is quoted at the wrong prices but then after that you know where the price should be but then there are other people who come in and just want to trade and you can

comfortably quote around the real price and make money on those trades so the real question is always whether the money that you make on uninformed flow upsets the losses to arbitr and that's how we turn this uh this expected value game from a negative to a positive for LPS um now with that being said there's still a draw there's still cost and I kind of reverting back to Yuki's North Star of of overcoming binance you know this is basically what what binance is doing they're drawing and there's arbitragers and they're drawing uh and they're leaking the value from the onchain liquidity games back to the the

offchain liquidity games on binance by Arbitrage so you know I I think we do and this was initially going to be a debate but honestly I think we have a common North Star I think we all have a common goal to sort of overcome binance make onchain liquidity and blockchain the liquidity hub for the entire Global Commerce um and and ultimately make make all of this a lot more verifiable and transparent so so solving lever lvr is is sort of a common problem that we all have now there's a few proposed solutions for this and I I think it's important to get into each of those one that I haven't heard discussed even when we were kind of leading up to this is like you know binance has maybe millisecond

block times or the way that the prices change you know you can Mark to Market every few milliseconds you know ethereum's every 12 seconds why don't we just lower the slot times on ethereum right so that the price updates quicker maybe that's one proposed solution Yuki maybe you can get into some of the other proposed Solutions and and feel free to put forth what you think is the best solution that's available at the moment and then we can kind of you know critique it and and sort of throw out you know what what some Alternatives might be he's basically asking you to show Cella sure um and and I guess just to also kind of like um uh put some perspective on the lower block time discussion because I believe that was something

that was quite extensively discussed and probably still continue discuss within the efite as well around the block time problem and I mean my personal Take On the Block time is that um I do believe lower block time in general is a ux uh improvements for the users um just because it's no longer like a 12 seconds but now maybe it's a uh 4 seconds or three seconds or whatever that may be so so it it does feels to me that it's a ux Improvement um and and also uh as as it is the case with Arbitron uh that uh uh there will be a sort of like magnitude less

uh economic impact to the lp uh due to lever giv a lower uh block time and um and that is I think is also pretty uh shown well on some other l2s uh as well uh but in terms of the sort of like the problems and and trying to have the solutions for for this um I would say like block time is decrease in Block time is is is like still like very much up in the air like I think there's still a lot of contentions there so I'm not sure if that would realistically be something that we can uh rely on uh what we are trying to pursue uh right now is uh is via this uh approach called

application specific sequencing which I often like to call it uh ass and um um and just for context you know I I before I I hopped on to really you know into this name of as I took a vote publicly on Twitter and make sure that everyone's actually you know happy with this name so that I can actually be confident about that okay I can say this publicly um anyway so so yeah that that was there um so so yeah so we're trying to do this ass uh ass thing as but but what what is ass specifically um so uh some of the sort of like work that has previously inspired and reinforced

our belief for ass is actually done by some researchers who worked on like the verifiable sequencing rule for uh decentralized exchanges uh there was a paper um uh that was also uh written by tun on the um the theory towards uh me on uncertainty where they have extensively discussed and demonstrated that the um essentially like with uh with anything around like Arbitrage sandwich chains front running back running um that oftentimes yields a a very uh complex uh pay of functions for the users and also for the validators

meaning those people who are including those block behind the scene um it's often times uh more favorable to use something like application with sequencing such that you can uh basically tailor and optimize uh one pair of functions over the others so if I let's say if my objective is to say that okay I W to I want to optimize the the outcome for the swappers like or or or like people who are trading on the protocol then I would essentially be uh be be enforcing rules within this as uh setups such that the

users will be let's say getting the best optimal price possible within the the flows of transaction that need to be sequenced all together and in a similar fashion I can also have uh objective function where we can uh we we want to optimize the economic payoff of the LPS right which are also participants in in this case and uh we can have the uh sequencing rules in such a fashion that LPS now will get the most uh value out of this whole uh sequencing that is done uh in in our ass uh setup so in such fashion uh we are essentially trying to

um help LPS help swappers get the best economic payoff possible by having our own appp specific sequencing rules and app specific sequencing in general so that's essentially what we're trying to do uh with uh with our protocol called anr yeah and so so like to to put this into a perspective of applications on the L1 are almost like you sit down at a restaurant and there's a set menu and these are the these are the items that you're having and this is the rules that you're following based on this menu That's the night the application specific sequencing model with custom

izations for different functions it's almost like you walk into a buffet and you can kind of pick and choose what you want to optimize for that in that evening in that application in that chain and basically you can create the outcome that you want for your use case um and you can kind of eat whatever food you want is that a is that a is that an appropriate analogy even though kind of way out of the spectrum of blockchain yeah um and that that would be a yes analogy that would make sense yeah cool so Alex when it comes to acific sequencing are you equally convinced that this is something that you believe can solve the lever and kind of uh this like sex STX Arbitrage and this kind of

differentiation between the block times of ethereum and centralized exchanges or even l2s and centralized exchanges or is there something else that you uh or is there any kind of uh other opinion that you may have yes sure and I'd like to start uh with a quick comment on block times because uh we seem to be discussing two things at once but I think it it actually gives me an opportunity to introduce the my framework of thinking about this issue very nicely so in general how do we compete with binance can price Discovery ever happen on chain um can we get there by reducing block times and uh my answer

to this question in general is that well you can keep reducing block times but binance has a centralized server so they will always have an advantage there it's clear that leading in price Discovery is going to be a pretty long shot for decentralized systems because whatever you are doing in your nied decentralized system I can also do on a server and I can get a better result in terms of latency I can probably trans process more transactions and so on and so forth so there are huge savings by doing this in a centralized way and so I don't think that uh just doing that reducing block

times and I should also mention another component is reducing the fixed cost of a transaction so think about Base gas those two things are known to reduce lever but by doing that can we compete with binance well the answer to this is probably no uh centralized exchanges will still lead in price Discovery does this mean that the whole exercise is hopeless then well the answer to this is no because remember that there are two sides to the income statement of LPS you have the lever you have the losses to arbitragers but then you have massive uninformed flows on chain they're really big and some pair

actually manage to break even and some pairs actually manage to make money by relying on those uninformed flow this is the flow that is not on binance it has to be on chain so if you think about some transaction uh where it has to be Atomic and a part of it is a swap um transactions like that happen all the time there can be liquidations you can be trading through multiple Pairs and so on and so forth so it's a it's it's an area that is not super well studied so I don't think that we understand it as much as we understand Arbitrage flow now but there are a lot of transactions that require composability and they are ready to pay to liquidity providers to provide

liquidity to them and this is something that they cannot do on binance if you have an atomic transaction you cannot do a part of it on binance so even though you may not be able to lead in price Discovery you can still make money because there are big uninformed flows on chain which brings me to the point about app specific sequencing as Yuki described it very well the idea is that if you are responsible for sequencing your own transactions then you can have for example an of chain auction where multiple parties will bid to be the first transaction touching

your pool in a particular block you'll pick the highest bid and you will return this bid to L piece for example so ideas like that have been circulating for a while what what if you are to implement something like this what you would have to do is that you would have to run this auction and then you would have to ensure that a certain transaction is the first one on chain touching your pool this in particular means that if somebody else wants to touch the pool ahead of that transaction that it will probably have to be reverted you will not let that transaction go through so here you are imposing a sequence of transaction of transactions a specific sequence for your

app and can this reduce lever well certainly it can right as we just discussed if you are running an auction on the site and you are reimbursing the preceeds of that auction to LPS then you can certainly minimize losses to arbitragers but what's the downside of having a specific sequencing but the downside is that you have to think about what would happen to the uninformed flow what if there is a liquidation that wants to do a swap on your pool how would that work so they also need to be they also need to compete for priority in the block and uh your transaction and if they want if they want to touch your

pool they have to let somebody go ahead of them but if they let somebody go ahead of them to touch your pool then somebody else will jump ahead of them in the block and do the liquidation so the impact of solutions like app specific sequencing on uninformed flow would have to be considered very carefully if you look at the current balance of if you look at the current income statement of lpce in the E usdc pair for example you'll see that I'm talking about Unis swap V3 five basis points which is the highest volume one you'll see that LPS are losing money on average

but they're making a very big amount of money on uninformed flow they're just losing more to arbitragers and then there is this difference that is relatively small to the amount of money that they make that is relatively small to the amount that arres make but it's a pretty big difference still pretty sizable difference now if you reduce this number by uh 50% and the number the other number stays the same this is good because now they're making a profit if you reduce both numbers by 15% they're still making losses if you reduce one number by 50% but the second number disappears all together then you are in trouble so the question is not just how

to reduce lvr and I know that lvr was uh doing great on crypto Twitter everyone loves lvr everyone loves l we keep talking about it we keep talking about lever Solutions and so on and so forth but there is the other side of the story in traditional Finance if you uh ask somebody who did this for a while what are market makers paid for well they are paid for providing immediacy to the market when somebody wants to trade their liquidity is there they're providing immediacy but what is their cost well their cost is that if they quote their liquidity at the wrong price they can be

picked off this is exactly the same tradeoff in decentralized finance in decentralized exchanges and uh you uh make money on the uninformed flow when you provide liquidity for it you lose money to other treasers but if you are too strict on the ordering side then there is a significant chance that you will lose out on some of that form flow and I think that this is the risk that needs to be considered berx Edge is now live on say Network say users can now trade 50 plus spot and per markets with the unified

order book liquidity of vertex Edge across arbitrum blast mantle and SE the next 6 weeks 850,000 say tokens will be up for grabs as trading rewards every every single week exclusively on say join vertex on say Network now on vertex Edge that's vertex protocol.com actually um I I just wanted to kind of like um clarify uh some some of the comments here um so I I see your points that um essentially some loss of composability

within uh as stabs may lead to some loss of uninformed flow as well right like as you described in the examples but um my view on this is that sure there might be some specific cases specific types of a un influent flow that are lost but um I wouldn't argue that it would be so much to the extent that for example the protocol would basically like completely like lose a lion share of the uninformed flow somehow like that that wouldn't make a lot of sense to me

uh Alex yeah well it's hard for me to tell so we'll we'll have to wait and see I mean there is there are certainly other sources of I'm form flow that you may not be able to lose so right you know I I love the composability part of defi I think that if def is to compete if we are to compete with binance then composability is the big huge weapon is the big huge resource that def has because again otherwise everything is fast on binance uh with um EP specific sequencing we will have to wait for your solution to come out uh the I have seen um one particular solution uh where some

of should I call them your competitors your potential competitors uh they were running a centralized auction right and again you can write you can run things much faster if they are centralized so they were running some sort of centralized auction to be the first transaction in um in in the block that can touch the pool the pool was gated and uh I looked at the performance of that particular pool and it was uh in that particular example it was pretty bad so it was essentially straight down um and you know one of the issues that when we discussed this uh on Twitter one of the issues

that uh came up was that nobody really wants to touch that pool except for arbitragers now you can never reduce lvr to zero right arbitragers will not trade against their your pool if they making losses we have to understand that so whatever happens arbitragers are making money otherwise they're not there and they do have some very real risks when they trade against anything because it's sex deck subit trash one leg is on the deck the other leg is on the sex so if you manage to get this leg done right and then you don't hedge in time then you can end up losing money so they do have some inventory risk and they charge for that and then they have to make some profit on top of

that so losing money to arbitragers is inevitable otherwise they will not be trading with your pool the question is how can you make it attractive to the uninformed flow and uh I think again composability in my view um is a huge source of uninformed flow and I would love in general for the community to kind of reinforce that there are a lot of solutions right now they go against possibility I would love to bring Focus back on composability because that's the way how I think we should compete with centralized Finance everything else can be done faster on a centralized

server there is no way you can compete you can you come up with the most beautiful elegant algorithm but if I can do it faster on a centralized server I can come up with a better algorithm except for one thing when you are composable with everyone else who is in chain so so I I guess I would like to uh make some comments on this and I think you know uh I believe the I think we are thinking about the same uh uh protocol when we when when we're talking about you know uh the the the the protocol that did that don't don't have that much uninformed flow and you know uh I I think I believe that we're probably on

the same page on that one um so obviously that one is like very protocol specific and I I think it's a bit hard to like generalize that uh that as well because uh there are some differences in designs around like okay if you're running a dynamic auction uh or if you're running an auction in expectation uh meaning it's whether it's a expost or exanti auction uh would also have uh a different uh ux for the Searchers and offers who are competing in this uh uh auctions because uh now they would have to be um sort of like valuing how much they can bid in expectation of future uh future OB Char

um and sexx uh Revenue versus if they were to be able to dynamically bid those values and then essentially um assess how much they need to bid based on what's the market uh price actions as well so there's also uh quite a bit of ux differences in terms of the designs around how to uh construct this auction mechanism um also I think um um another thing I would I would mention about the composability part is because like I I I'm 100% agree that the composability is a very important uh aspect of uh uh Defi and uh I do believe that uh OB obviously I don't have a

number on my hand but I do believe that um some lack of composability may result in let's say uh some reduced uh uninformed flow uh now since now people may not be able to create a bundle very easily and they be able to execute certain banch of transactions in in in atomicity however I do want to um kind of point out that that sure composability might be a problem but it's not a unsolvable problem meaning that there are certain Solutions out there that potentially can be leveraged um for better composability for ass

so um just to kind of give a context like uh in terms of what we can do to to to tackle this issue um one approach that I'm personally quite interested in seeing is uh specific to the the the transaction this big transaction bundle that we are crafting with with uh application specific sequencing if we can achieve some levels of a builder commitments uh by having let's say uh the top let's say top three Builders uh be partaking in giving commitments on certain transaction ordering or

combinations with our bundle then we will be able to essentially uh provide uh a level of composability that is uh uh maybe perhaps it's economically guaranteed or perhaps it's uh some sort of commitment uh based on those uh Builders agreeing to say that okay if you pay me this premium on this I will be able to provide you uh Atomic composability between your bundles and some other ex uh extra transaction that are uh outside of uh sella uh uh sella protocols so that is one possible way VIA Builder

commitments that we can provide s some levels of atomic composability and another approach that uh I think some people might be also discussing about uh in in some corner of the uh uh the the room here is the um the share sequencer so I have also uh come across few folks who are actually uh interested in exploring how the Shar sequencer may also be able to tap into this uh uh ass uh Paradigm and then be able to provide uh full Atomic composability between uh different apps

that are also running uh with apps specific sequencing uh in a similar fashion so yeah in general there are like those two two areas of Explorations uh specific to composability issues that I'm quite interested in uh seeing how it would uh develop in the in the future and I think both of them are still uh quite early but both have uh different teams that are exploring and working on solutions to to kind of build this out in production as well so I'm personally quite uh excited to see how this space would evolve and potentially help uh projects that are as um based and and make them more composable across non ass and ass STS uh uh as

well awesome well uh I I think what we're looking for here is is uh ass composability um this is this is kind of The Best of Both Worlds if it's possible uh then we're able to maintain the uninformed flow because Alex you kind of pause it that uh unin uninformed flow really values composability um and that's a main value proposition to the uninformed flow uh we can maintain that if we have composability um but if we have ass we can lower lvr that way we're we're able to bring down lvr we can main we can maximize the uninformed flow um but right now in a in the current state of ass we we don't

have composability um we we have some fragmentation unless ass is done in in some certain ways so um I think this is this is sort of where where it's left uh in this current state but I do have one maybe one or two more follow-ups which is uh we recently had a podcast with Justin Drake he was very fav favorable when it came to composability and and this is like a very key value proposition of ethereum um but also very frankly bullish on on Cella and the way that it it rebates LPS um and incubates that from leaking it uh leaking me back to validators and so one criticism was that uh

currently uh ass and Cella is kind of dependent on uh the next e it's dependent on Unis swap V4 uh hooks which is dependent on uh the next ethereum hard fork and so I was curious maybe we could sort of just unpack like what is kind of like coming up in in the next you know on the road map of ethereum maybe in the next hard Fork that gives way to this new this new mechanism which empowers uh Unis swap V4 uh Hooks and and downstream Empower Cella to rebate LPS uh what is it that that is in this next ethereum hard Fork that is key to this mechanism yeah so I believe that um um right now our uh like yes so we have few

uh dependencies there one is obviously the the V4 and V4 itself also have some dependencies um as well on the next update but I think on like specifically to our side I believe our dependency on the eth upgrade isn't as significant I would say like um to the extent that I for example like if you know uh there were some reason that somehow you know they had to um fix up some of the road map uh as long as the V4 can uh ship uh like uh without any issue I think we should be we should be okay um so I believe our dependency

on that side should be uh not too big of a concern um however one thing I do want to uh sort of mention that is that um um I think in in general like uh we are very conscious of like how we would like to you know uh position and also uh build our s like the liquidity uh ecosystem on top of Unis swap uh because I think unisoft are taking a very heavy lifting by building this whole so like a new amm stack in itself has a lot of gas

optimization that are at the same time enabled um by the next upgrades um and some of which are enabled by next next upgrades on ethereum um that are benefit official for every single user who are trading uh based on the hook of the V4 protocol and and and to that extent like we're like very uh excited about you know the fundamental difference fundamental changes that are it's brought by brought forth by uniso before uh that would uh hopefully also translates to better uh user welfare uh trade execution for uh users using Proto protocol uh like

ours yeah I think I think um uh Alex pointed out just quickly that I think that it was actually some of the previous upgrade um that enabled some of the Technologies and so kind of looking forward to seeing that uh come to light and for from my understanding of abst sequencing as well uh there's kind of there's kind of I wouldn't say an onoff switch but there's a way in which that you can segment which parts of the application you want to be uh uh sequenced in your app or sequenced by the global whether it be a sequencer on L2 or the ethereum validators um Alex I'm curious if that concept changes your perception at all or kind of has

influenced your thinking and then kind of on the downstream of that if we see a lot of kind of app specific sequence apps that have that that property or have that property look I think it's an interesting design space for sure so I very much appreciate the work that all of the teams are doing uh with app specific sequencing with shared sequencing and all of that it just that eventually when you have two apps both of which want to both of which want to sequence their own transactions you will inevitably run into situation where one wants them in one order and the other one wants them in the opposite order and then there is the Eternal question what do we do next and who ultimately decides

to um to uh pick the order right so we know that validators are responsible for picking the order so to enforce their own ordering the only that apps can do essentially they can say well I'm not playing that game with you so if you don't sequence according to my rules then you are not getting my liquidity you are not getting access to my protocal and all of that which again uh might work but it does have a um that does impose some additional restrictions on composability maybe those restrictions are not as important maybe composability is not important in itself I would not love to see that

because I think that if we are in a space where composability becomes not important then really there is not much sense in doing things in decentralized way anymore because now everyone can do their own thing whatever somebody comes up with you can come up with the most beautiful algorithm mechanism or whatever I can always run it on the centralized server and make it more efficient and just beat you on speed latency throughput all of the parameters right whatever can be done that way can also be done in a centralized fashion and so I think that this is the beauty of compossibility in all of this is that uh this is something

that uh you cannot really uh replicate you have to be where everything else is happening and you have to be able to provide liquidity to transactions even though they're not sequenced by you but again it's very interesting design space so some people thinking how to relax it maybe we should be a little bit more mindful about some of these limitations so I certainly expect to see and would love to see more innovation in this area also one thing I would like to say is that there's a there's a a marked difference between uh like the two approach that I've discussed about share sequencing and Builder commitments um

like as Alex said like Builder commitment sorry Ser sequencing is a very complex uh problem like you know there are a few notable projects that are working on this space that are trying to figure out how this uh very complex combinatorial auctions even works when you have uh multiple entities with a different uh um objective functions that are optimizing for different things and doing that commentatory auction is you know IEX I I I don't know if you have a solution for that but it's a hella complicated thing I don't even want to think about how to you know design that but um but then you go back to the other side which is the build commments um I

personally believe and I you know please feel free to disagree with me as well but I I do think that um Builder commitment is already happening a lot today and the reason why I say that is because I believe bundle in general like a transaction bundle within a block in general is a form of a uh Builder commitment or you can argue that it's it's a former relay and Builder commitments meaning that those transactions when it comes and lands to the hand of the relays with the payload they're not being uh Unbound unbundled in in some uh malicious ways and that's sort of like a um implicit commitments that people are receiving when they

submit these kind of transaction bundles so in some way I I believe that uh this kind of Builder commitment is already happening uh in obviously a somewhat trust trusted F fashion right because we still rely on relates to facility a lot of this but there are also um there are also uh like more expressive ways in which we can enable such uh commitments that kind of gets into how I was thinking about okay what if we can enable uh Builders to commit to uh in uh commit to having transactions and executed in combination

with our bundles with some other like external transactions such that they can achieve uh composability between transactions outside of our protocol and transactions inside of our inside of our transaction bundle and then you would a able to receive Atomic guarantees uh Atomic composability guarantees uh via such kind of like a buau commitments yeah yeah I I have not seen the specifics of the build a commitment thing so I I I would have to see it to think about it um I think it becomes a pretty difficult problem though for the Builder and also exactly with commit to but again it depends on details so I'm not saying that look

maybe you don't need full composability maybe you know if you can cover 80% of it good enough right so sort of depends on uh how um how aggressive you want to be on that side certainly some things can be achieved with Builder commitments and again they're probably there already in practice it's just that I'm I'm not too familiar with the specifics of the design yeah for sure and I believe uh right now um uh Prime guys are primarily working on a lot of this uh as well and I think um that one is definitely interesting to uh keep in check on my side to to see how it would

be interesting for us to be uh working with um one one thing that I do have a question for Alex though is that because you've been talking a lot about like the composability aspect of those like ass dabs um do you have some sort of like insights into like how much uninformed flows are generally coming from those uh uh transactions that are attempting to be composable welcome aboard on our journey into the singularity this is the endgame road map for fra and fraxel the new L2 out of the fra ecosystem that is totally based on fractal scaling fractal a modular L2 blockchain built by fracks

with modular and fractal scaling in mind fractal scaling is based on the mandal BR set of fractal images and it explains how l2s become l3s and the entire app chain life cycle goes through its Evolution the FRX ecosystem is made of some of the top builders in the space and has stemmed from its OG days of the frax stable coin it is is now in the LST space the LRT space and this fraxel L2 joins the existing Suite of fra products including fra Swap and fra lend you can see fra fairy the bridge to go from any

chain arbitrum optimism or main net into the frax main net Bridge your tokens and bring your FRA assets into the fractal ecosystem join today to get the best out of fraxel you are still early join today take part in our Expedition into the singularity we have looked at some uh data we haven't circulated it yet but uh my impression is that uh let me just give you my impression and I know that you have the uh the new dashboard so you might want to double check those numbers uh my impression is that generally it's a lot in particular the large transactions um what what transaction

large large uninformed transactions and you actually can generate you know a lot of income is coming from those transactions but it's yeah it's one of those questions that again I think it's impossible to answer and basically we just have to wait and see and uh try and see what happens um I just feel like if again if we lose composability then uh do we have a shot at competing with more centralized Solutions that to me is a really big and open question and again there are chains already that are faster right so you you you almost like I always have

this feeling that you're almost becoming like an L2 essentially you you're not exactly an L2 but you're already sequencing your own transactions right so there's a question if you know there is AUM with lots of liquidity and they used to have first comfor surf sequencing so that minimize lever quite a bit and liquidity is deep there and so if you want to compete with them um then uh a transaction that doesn't need to be composable if you just need to Source e somewhere and uh you just want to be balanced in terms of your general portfolio you don't need to be balanced in the specific transaction you know you can you can you can buy it on a different chain uh on an L2 and uh

achieve a similar result and those are already very deep and equid uh markets uh pretty efficient in terms of minimizing lvr just again on account of having low block times and low transaction cost so there is a lot of uh competition in that area but um again I I understand that there is a difference between app specific sequencing and a rollup specific sequencing is still a little bit closer so maybe it's a spectrum right we're a little bit closer to the main net we're getting a little bit more of the uninformed flow uh than those arbitrum guys but we may be getting a little bit less than vanilla uncope pool so maybe there is something in between that is actually a

profitable and interesting Niche right and and I would imagine that uh the reason why those large orders are oftentimes um uh are part of some bundles is because that um some routing or some solers who are executing those large orders via uh some RFQ so not RFQ uh via like silver based protocols like you know C unx or uh onein Fusion whatever uh might be accessing multiple liquidity pools uh at the same time when they might be executed oh sure they they access multiple liquidity tools pools but it's not just that you can be a part

of a bundle it's also that you can be competing for priority yourself right sure sure sure yeah right so you can have a single transaction that needs to do something you you know you can be doing a liquidation and then you're going through some swap yeah yeah yeah you want to have priority because you're doing the liquidation and so you cannot go through some like other channel and lose priority because of that um but you need to so basically you need to get priority for a different reason so that happens a lot as I understand terms of what volum that's also pretty significant there are things that you cannot actually we say that everything

is on chain so you can look and see but there are actually interesting use cases and pieces of the data that would be great to have but you can actually not collect those that EAS I mean you can collect all the swap information and all that but you cannot necessarily figure out what the intent of the user doing it that is and so it's sometimes ult to have the full picture um and again it's then becomes a question of your prior do you think that that's important to do anything that's not important and I think that people should just try all those different things and basically see what sticks this is one of theages of being

in defi is that we can try everything and see what works yeah I I think I generally agree uh with with the take that yeah we should just kind of try and see how how it works because at the end of the day like as you as Alex you mentioned that uh um like getting exact picture on how the UNL flow and it works or like unfl Flor like goes in terms of whether it's part of the bundle or whether it's a whe whether there's actively seeking for composability or not or how much percentage of that of that is part of the bundles is in my opinion uh quite a uh quite a heavy lifting uh if you want to really try to

figure it out um but one one one thing that I do think is is very true is that um um we we also want to have composability we also want to be able to address this composability uh issues and I believe that uh there are solutions in the space as as mentioned that can be potentially addressing that even to like maybe not to the 100% extent but to the extent that is sufficient to be able to provide let's say uh if someone is trying to liquidate something on a and that they want to immediately sell off some positions um they can still be able to part partake in uh and and and Route those orders via let's say uh our sell's

uh mechanism by ass and such that they can um uh still be able to get the uh transaction execution guarantees uh in in somewhat Atomic uh Fashions um and and and just a matter of like okay what what are the pathways they're taking are they going to be using some sort of Builder commitments are they going to be using some sort of like sh sequencer uh or what like that part is still uh TBD um but I I think that uh it's it it in my opinion it's not a unsolvable problem per se yeah and then I just one one quick comment uh on on a part that uh Alex alluded to about the the L2 uh not L2 uh because yeah I I think initially

when the idea of like ass has been circulated um there's being a lot of uh confusion uh on Twitter uh that people are saying like oh is this is an L2 and I think one biggest differentiation here or the difference between those two uh Paradigm is that one is uh clearly uh executing on those transactions and that uh uh there are performing some sort of like you know fraud proof or uh some some you know ZK for verifications of the state transitions while the ass side is clearly not doing any of that and is not bearing any of those uh trust and

liveness overhead that is associated with executing uh uh State transitions on an off train machine so that is like of one stock difference that I wanted to just point out for people who might be uh confused about the those two different um uh setup yeah yeah I think that's like uh appic sequencing took me off of the app chains ledge pretty hard um I was pretty s on that and now it's kind of like oh wait this is this is almost like meeting in the middle and I think just to kind of wrap the conversation they when I first got into ethereum it was like 2017 and then I think in 2020 I think vital put a post describing ethereum as the onchain Dark

Forest and and I think that all of this conversation today has been like reminiscent of this idea of of kind of Shining Light Into the onchange Dark Forest to make us closer and bring us closer to kind of taking down uh if you will the centralized uh arbitres but really just this kind of the all this liquidity and user base on on binance and kind of bringing a fair system on chain Alex I think made a ton of really good points with regards to the centralized nature of of systems being much more ripe for Price Discovery and much more ripe for low latency but I think that if we as an industry can kind of strive to create the most

decentralized systems with some of these with kind of sacrificing a little bit for latency then you know we can we can kind of find a a happy middle but I also feel that like there's going to be builders in the space who will kind of go on both sides of each Spectrum um you know kind of the left and the right of speed versus kind of full decentralization but I it's been inspiring to kind of hear from your perspectives how you see new solutions that don't have to make a choice between these two trade-offs and kind of build with both in mind to the best of the current infrastructure stack that exists today which I think is like it's going to be super empowering for uh Builders and developers of the coming years who

get into crypto as well as just making the onchan landscape more fair so broadly I just want to say this is an extremely extremely knowledgeable conversation I I'm glad that you guys came and super super fun to kind of understand how you guys are thinking about the onchain Dynamics from somebody who's like in the infrastructure side but also has like lost money from impermanent loss far too many times and like has elped on every deck since like the beginning of dexes so it's been great thank you guys thank you very much thank you Andy thank you Robbie thank you Andy thank you Robbie thanks for having me absolutely thank you guys