Fireside Chat: ETHconomics Meets MEVconomics - Davide Crapis, Barnabé Monnot, Justin Drake

foreign [Applause] [Music] but we are going to be moving over this transitions while in Korea next talk we'll be bringing on Davide Barnaby and Justin to be talking about economics meets web economics and we'll be starting with a presentation so and we are kind of um gonna have like a fireside chat discussion on delay of the land more broadly like we start with like um a brief pitch introduction for me than just in the Barnaby and then probably gonna have like a logo q a uh so I can go to next slide and I'm going to start um at the high level the uh defining um uh Med flows kind of from first principle so if we go to the next slide so essentially um when I want to start is that these energy flows are actually a property not of the blockchain itself but of the economy that we build on top of blockchain so like the system of value transfer uh which includes also like the economic and financial gains that people play and uh on top and so they kind of affect uh the ethereum protocol but also the Injustice infrastructure or hopefully trustless infrastructure on the team protocol and then in like any other protocols that user may use and so essentially um when we think about how can we make ethereum protocol Mev resilient which is an idea that uh I put fourth a few days ago and we actually need to think about how the whole system can be and we be resilient right and this properties uh which I call blockchain SCC properties so security equity and cost efficiency like some of them are generated directly from the blockchain but some of them relate to the broader uh system right and so essentially like um what are these properties like the first one is economic security which in the context of Med uh we know that it changes the incentives from validators to behave honestly so we want to build uh incentives that are robust to Mev with like some of the solutions we are developing is essentially to make these incentives uh robust to this uh process that arose um and that now it's a constant because people are playing these defy games continuously uh the second property is uh Equity so here we think about Equity both um in um access so essentially like uh people should not be censored but also equity and distribution uh people should um uh have a fair game and um have a third distribution uh of value like uh comparable to the externalities they introducing the system so we want to minimize the losses for honest user and then the third one is cost efficiency which affects a fourth property which is the utility uh of using the blockchain and and here essentially we want unethical user a user that are just there to like play the game or extracting red to have many more rents so this is kind of uh the high level goal we are trying to achieve and and I'll argue there is like four ingredients to achieve this uh so the first song is we need to reduce Mev with clever design so essentially and both at the game level so the economic apps that we build should be cleverly designed to account for Med and also at the metagame level at the transaction inclusion game also that one should be cleverly designed to account for Mev and that's what most people in this group are working on and and then the second step is extraction so like there is a potential let me be in the system uh this uh can cause instabilities to incentives so we need an efficient technology for extraction today this is Searchers and Builder and then the third one is uh capture so once the Envy is extracted we need to make sure that uh this is captured and this needs to be captured by trustless protocol it doesn't need to be entirely the ethereum protocol it can be order flow auctions or other protocol uh that are in the broader infrastructure but it does need to be trustless because um otherwise uh there is going to be capture and like um uh this mechanism will not be credible and then finally any VR location is very important because essentially like uh once we capture this Mev we essentially need to make sure that um this flows back uh to the users uh or essentially maybe we can talk about this a little bit more what should be the goal of allocation so if you go to next slide so here is a cartoon view of uh the block supply chain and which kind of includes if you started it for a little bit all the elements that we are seeing today uh including the bad elements exclusive order flow vertical integration between Searcher and Builder uh horizontal competition at different layer maybe Stronger versus not stronger there could be segmentation um and then finally math take can differ at different layers so if you look at this view realize that although we would want a clean supply chain we absolutely have a supply network where like there is several connections across layers uh so we kind of need to be aware of this um and essentially if we go to next slide uh What uh we started doing is like looking closer into all the important metrics of the supply network we launched a rig open problem um which is uh essentially a way to advance research around like more complex properties of this supply network so essentially like what we are going to try to do is create an open source framework to both Define and measure these most more complex metrics that are related to inclusion delays they create for different uh segments of the market validator devious deviations from honest behavior and our goal is to bring together multiple parties in the ecosystem we already have people from flashbots block native aegifi units for clubs that are looking with us at uh what's happening concretely into the market in order to then use this data to like design better protocols and systems and this is not for me I believe uh adjusting is up next okay great thanks David um so today I want to talk about nav presidents which I guess I just named today which is what is the ordering in which um you know Mev to which parties the NBA should flow to like who should have presidents uh on the energy over over parties and basically um in my mind um we have uh we have two types of presidents of this we have the the President's List that we want and the the Precedence list that we have and those those are actually different today so in terms of what we want I think is we want to give users a precedence over Paul other types of of participants over the gmev so if they create and maybe they should be the ones that capture it through this concept of rebates you know called Med share and things like that and then my claim is that the the next entity in the the Precedence list should be uh holders the token holders and the way that they receive these these flows is through a burn so we already have uh part one of the burn which is eip1559 you know I think of the base fees as actually parts of Mev right it's it's uh it's not fancy or sophisticated Mev but it is uh extractable value and it turns out to be extracted by by if holders and there's this other idea called Mev bird which is kind of part two of the brand uh where all the energy not just one part of it goes to the holders so already from these these two items in the Precedence list and the the way that I I see where we want to be moving towards is users capturing all the immedi that they they initiate and then all the rest or almost all the rest going to to to to the holders um and now the reality is that there is a whole you know energy pipeline or chain or network whatever you want to call it of of um intermediaries and you know these NC series will will capture a little bit of MZ but um they will capture a very small minority and you know in a competitive market these these uh should the the inefficiencies should go should go to zero um and the reason why I ordered um these uh these entities in in the way here is basically the closer you are to the user um the the higher the Precedence you should have so if you're an interface for example you're very very close to the user you you provide value by you know bringing the user onto ethereum and so you should have precedence over the Searcher the Searcher is doing you know complicated stuff in terms of of building the bundles and that's that's maybe more valuable than than the Builder which is just bundle merging and then the relay is doing something even less valuable and then the proposal is just doing almost nothing like the the only thing that they're doing is that they're receiving a few bids and then they're picking the highest bid and they're signing it um they're doing almost almost no work now it turns out in terms of what we have today for the users I put the the skull emoji here that the users are being uh are being you know not receiving the flows and so the the the the they're not being prioritized in in this precedence list at all um and in position number one what we have is the proposer so the proposal who you know should be in my opinion the very bottom of the Precedence list is right now at the very top of the Precedence list and that that is something that uh that I think uh needs to be fixed and um one of the the the the the cool things of of of of this kind of model is that we have a clean definition of of a toxic energy and it's basically this idea that um it's Mev that doesn't respect the Precedence list um or at least the the ideal precedence list so in this case any Mev that should be going to the users but instead is going to the proposer is toxic immunity so things like things like front running and and and and sandwiches um and and you could argue that maybe uh proposers you know uh receiving receiving any of you that doesn't originate from the users is also toxic um and I'll I'll talk about that later because it leads to all sorts of um economic distortions that that manipulate the incentive that we've we we've tried to to bake into the protocol as designers um and then in terms of what we have again we have these these more commoditized entities that will receive a small portion of the of the MAV uh in in a competitive market foreign I want to talk about maybe what is the most controversial thing of this this whole diagram which is you know why should holders have precedence over the the the the the proposes and this goes back to this idea of um these economic distortions that happens when you give the Mev to the proposer and so um the way that we give energy to the to the holders is through Meb burn I won't explain how it works that's for another talk but I'll explain some of the the advantages of it so on the left uh we'll have distortions that happen if you give it the DMV to the proposal so the very first one is this idea of reals short-term rework so if you have a very valuable block 100 e for a thousand if ten thousand lead but how many thousand if who knows you know in a single in a single block then really you're you're putting a lot of pressure on the chain to do these short-term reorgs and we haven't seen that right you know yet but it it it might happen uh you know with as the market becomes more more sophisticated and more mature so that's the first thing that is not good when you give the Mev to the proposer um the second thing which is not great when you give Mev to the proposer is that you're basically introducing a lottery all the proposers are Lottery participants they have you have lottery tickets and as in you know most lotteries uh almost everyone is a loser right almost everyone doesn't win the lottery and then there's like the one person the one proposer who wins the jackpot at the expense of everyone else and in order to correct for this fact that almost everyone is a proposal is a loser the proposers need to pool and that creates you know centralization vectors it goes against solo validating and things like that so that's the second economic Distortion that we would avoid if we were giving dmev to the holders as opposed to the proposals um the the third economic Distortion is is what I call Rock pools so it's pooling is rugging the pools um and here the the observation is that even if you have a pooling system it it it's still not incentive aligned and the reason is that um the the the operator of a of a validator that's part of a pool has an incentive to to basically not give the the Mev to the pool if the Mev is large enough specifically is larger than their collateral or their reputation so there's a massive block that comes in 10 000 each block and you know they only operate one validator with 32 if you know they're happy to to say okay the pool you can keep the 32 if as my as my collateral but I'm gonna run away I'm gonna rock pool uh with the 10 000 if and so um if you're even if you like the concept of pools you know like like rocket pool or Lido um any of you burn actually makes these pools uh better because it removes the the rug pooling and then the the the the the the fourth economic Distortion that I want to highlight is this idea of drought so there's there's two use cases for each right there's the the staking which is economic security and then there's this ever very important use case which is economic bandwidth so for example we're going to need decentralized stable coins that are backed by pristine collateral by if and if we're in a in a bull market let's say where you know Mev has just gone crazy then what's going to happen is that the the the the incentive to become a Staker are going to grow and all the if is going to be sucked into staking and then that's going to dry up the the if for the economic bandwidth and it's basically going to to fight the ethereum economy um so in a way the the chain is kind of too successful and it's it's overpaying for security and that leads to a less economic um bandwidth for for the rest of the ecosystem and then another thing that that's not on this slide but I I just kind of remembered is that if you're giving the the movie to the proposers you're kind of incentivizing people to use um derivatives as as collateral and my my personal belief is that if derivatives are are risky to use as collateral because they can lose all their value from one day to another through through slashing and so they're not high grade collateral and ideal ideally they should be avoided if you're building things like decentralized estate work ones and in terms of the the good things if you give the Mev to the holder you have basically more scarcity you know you you increase the shelling point that if is is money for the internet and that ethereum is a settlement layer for the insect of value um but ironically you you're also increasing the the rewards uh for for stakers even though you're removing the Mev portion and uh the reason is that the the rewards are going to tend to the cost of value in either case but uh for most people you know because of this uh this variance around the the rewards most people are going to be earning in less than the cost of value so actually most people are going to be losing money staking a few people are going to be making money um uh stating so that makes everyone earn kind of the same the the same rewards and also from a USD denominated perspective um you know presumably if there's more scarcity the price of ethos is worth more and so you actually have more USD denominated Rewards even though the evening rewards are the same you also have more economic efficiency in the sense that you don't need as much issuance to pay for Economic Security and the the reason is that the amount of issuance grows uh with uh the the the total amount of stakers and so we can we can reduce the amount of stakers and and therefore have less assurance and still have enough Economic Security to for ethereum and then the final thing is that um this kind of related to the scarcity um you know even though the total amount of Eve staked uh you know will go down it's possible that counter-intuitively you'll have more Economic Security and the reason is that again the price of each might be worth more and so from a USD denominated standpoint you might have more Economic Security so all of this to to kind of to justify this this new precedence list I think we should be uh targeting for when we uh redirect Mev to various participants and now on to Barnaby I believe thank you interesting hi everyone I want to talk a bit about protocol credibility and principal agent problems um great so can you move on to the next slide please in October I released a post on embeddling PBS with something I called Pepsi protocol and first proposal commitment it allows the proposal to basically commit to any kind of executable contract and the commitment is fully defended by the protocol in the sense that blocked validity depends on the satisfaction of The Proposal commitment in particular you can write PBS in in Pepsi where you can also do much more than that so it's much more General and Pepsi came from a feeling and I was somewhat unsatisfied let's say I thought we don't really have a theory of protocol upgrades and so we might as well give ourselves as much optionality as we can and Pepsi is really maximum optionality in terms of what the proposal can commit to that's not to say that we shouldn't do protocol upgrades I think we have a good intuition for them we've done them in the past and they worked pretty successfully and there's something to be said for being iterative Community Driven this kind of diffuse pragmatic approach to protagonal development any complexifying economy so processes get more sophisticated you have the division of labor that we call modularization in in blockchain slang if you can move on to the next digesting and so this division of labor it fractures basically trust domains it more and more things get delegated you have delegation and if you have delegation you have principal agent problems these paps and if you have principal agent problems you have potentially misaligned incentives and so we can always choose to protocolize systemic paps Proto principle agent problems but then we have to Define what the system is and in this conversation I think we often let to ask what is the role of a protocol in that larger system of protocol and infrastructure and I don't think we can have this conversation properly without having a bit more theoretical background on on basically protocol credibility and so with discussion on credibility is really taking off at the moment along with commitments to me the groundbreaking work is virgins ethereum is game changing technology literally but recently Sheen from flashbots has revived his conversation he has a much more holistic theory of credibility and commitments and having in parallel very much in the Zeitgeist you're starting to see much more results on credibility in auctions and credibility mechanism design where you don't have let's say a central controller and so what this conversation tells me is that there is this hierarchy of commitments and the reason why we want to put things in protocol is because we see that as being more credible and so I want to just go a bit into why we might think that in and basically it's because the protocol gives us a shelling fence the validator set says anything we put behind the fence anything we put in the protocol is going to be defended with the whole might of a validator set but the protocol can only defend what it sees so most of the upgrades that we do they're really done with the aim of extending the vision of the of the protocol and I call it seeing like a protocol we want to make legible certain outcomes and certain pieces of infrastructure so that the protocol can control and defend them and so the aim is to get a credible signal for instance if we see a safety fault in the FFG Gadget if we have the basically Visa role credible signals of things that are happening in the protocol next slide please and so this protocol credibility to me is the sum of two things is this first protocol introspection and also protocol agency so the more introspection we have the more control and the more we can have actionable responses to events that are happening so sometimes we don't want to go through the diffuse process of community government if there's a fault that we can't control for instance people are driving basically to zero it's not something that the protocol can see as a fault and we need to act as a community to to defend against that but there are faults that we can deal with pretty much automatically like slashing people who are finalizing and doing safety Folds but the trade-off here is that when we build in this protocol credibility when we extend the boundary of a protocol it means that we also need to become opinionated about outcomes that we want and so there is a higher risk that we log in kind of local Optima or mechanisms that are sub-optimal but more importantly when we do that we also start overloading the validator set with the duty to defend more and more things and we can do so but then we must be convinced that we as a community are going to defend these mechanisms basically with the protocol's life for whatever Force under its zone of control for instance we commit to forking out validators who subvert the censorship resistance Gadget that we build in the med burning the finality gadgets all of these gadgets and to me this feels like it gets harder as time goes on because there are other domains like whatever like eigen layer that may directly shift the incentives of the validators such that honest behavior and rational Behavior start deferring more and more especially when the protocol has locked in sub-optimal outcome and by civil climate I mean non-wheel cell maximizing because when this is the case the welfare Gap is basically your subsidy for bribing validators into into doing the the wrong thing and so back to Pepsi to mediate it was nice because it creates this idea of programmable protocol credibility but it may be adults with the goals of zeroing in on a minimal set of gadgets that we are willing to defend with basically the protocol's life and that question is is still open in my mind so yeah I think we're getting closer to this theory of protocol upgrade uh I think commitments credibility both have to do with that and I'm excited to to see what's next I think we'll be overtime but we should jump to the questions awesome great presentations guys uh so we will have a brief kind of question section here before we move on to the next one um the first one I'll go with will be kind of basically getting uh David and Barnaby your reactions a bit on kind of the Mev allocation and distribution type of convo for the most part um I know Davide like you've written previously as far as about allocation the goal is basically distribute the captured Mev to participants in a way that is most beneficial to the protocol um so that last part is really kind of what owns in on the question of how do we Define what is that objective function of what is beneficial to the protocol that we're trying to optimize here um whether that be mem smoothing burning um in particular because specific does address some of those points um that Justin had spoken about as far as the real resistance and some of that kind of stuff um so like how you guys kind of think about what you're optimizing for here yeah a good question so I think the answer is nuanced and it ties to like uh parts of the presentation of Justin and Barbara as well so maybe I'll start and then they can continue so like in the ideal world like what does most beneficial for the protocol means I mean the protocol is just like organizing all the participants right users and like service providers validators in this case right so like on the user side like the ideal allocation is essentially like a fair tax so essentially like them any video contribute to the system and the Nev you get allocated like if you subtract those they should be proportional to the externality that you bring to the system on the other side you also need to use like part of this Mev to make sure that like the validatory incentives are not distorted behaving honestly so I think this ties to like uh done uh kind of Precedence hierarchy that Justin was talking about and yeah so I'm not sure he wants to comment next also because it doesn't the allocation doesn't need to be monetary could be purely fiscal but it could also be monetary in the Ace of burn so what do you mean by fiscal in the sense it's a problem you could or you could organize direct transfers right for example in the guys in the guy in the type of um allocation that an OFA system would do there is a direct transfer back to the user so essentially the user is getting that Meb directly locked through and in facial or deflation in this case yeah right I mean I I you know there's a separation of the consensus area and the execution layer right and like I guess what you're saying is that the execution layer is the fiscal layer and I mean I don't think users would would get issuance right that's kind of the wrong layout because Assurance is at the consensus clear right okay yeah I mean Sonia made a good point around smoothing you know potentially solving some of some of these issues um smoothing doesn't solve the the economic bandwidth drought basically just uh the staking sucking um yeah and it is also unclear in terms of the the rioces even the Rio like one of the ideas is that we have some sort of a small committee that's in charge of the the smoothing but this small committee is maybe not representative of the whole validator set or maybe it is I don't know um but um there's there's some Nuance there a little bit um yeah that point if I can commence I think sometimes we also get the monetary policy that we can defend and if we can't defend things like nav done or things like Maps moving because they rely at the end of the day on this on this majority of committee to to defend these things and yeah it might be something that we we can't get but um I wanted to also mention some research that is done by andersonal group on the dynamic yield curve and notably the idea of capping the size of the validator set and potentially even auctioning of the slots of validator said that's I would say another way of internalizing that Mev allocation in the protocol and in particular if you get rewards from like again layer as a validator you're going to be willing to pay more at this option and so you can even internalize more than the execution layer Mev via this mechanism or you could do both I think that would be the ideal right I mean one of the things that you mentioned is that energy burn was moving requires an honest majority I I do want to highlight that yeah I keep on 559 also requires Anonymous majority um and just for people to to understand this basically if you have a if you have a dishonest majority um they control the full Choice Rule and then what they can do is that they can Fork any any block that increases the base feed and so basically the basically is like decreasing only and it goes to zero and then you've just essentially deactivated the ip1559 and actually there's a lot of parallels between the lg1559 and mug brand it's just essentially the same mechanisms the same Dynamics it's just on the different side of the coin foreign so moving on to one of the other questions that we had uh this one will be for you Justin it relates to several of the earlier presentations I've referenced it um but your recent post on base Roll-Ups where basic idea is letting the layer one kind of do the sequencing for these Roll-Ups eventually um and tldr a very simple Mev implication here is if you implement it in that simple form um basically the Meb that would normally accrue to roll up sequencers would start going to the L1 cells so the question here is kind of why is that such a desirable or undesirable feature of this and then in particular is there kind of a tension here between now all the value goes to ethereum and maybe this is kind of the dominant solution to them and compared to roll it should have their own incentives to continue to innovate and continue to grow over time is that kind of in any way a rejection of like the role of censorship roadmap where we want this private Market of people to incentivized to go make these things right I mean on the on the topic of incentivization I'm not at all worried about the rolled up teams right there's like 10 Billion Dollar Plus kind of market caps I think they're doing very very well and you know for good reason because um you know they're they're doing something very very complicated that requires you know very big teams it requires taking a lot of risk um and and also the the fact that we have these tokens is kind of fuel to kind of almost disrupt a little bit the layer one network effects and create new network effects at Layer Two so um I'm I'm I'm all for like this uh you know this capitalism as they were at the rollout player uh but I think what will happen is that you know what is capitalism in the short and medium term might become you know commoditized eventually it will be just like running water and electricity um and you know there will be less and less of a need for for incentivization um I mean even the roll-up teams themselves right they want to be credibly neutral they want to be maximally decentralized and so they go for example from centralized sequencer to decentralized sequencer they go from governance you know token governance for upgrades and then you know they might remove the governance and become like uni swap you know fully trustless um and I think you know the move from being layer 2 sequence to being layer one sequence is part of this natural progression of of progressive decentralization and then once you've reached the the base roll up you know the the ultimate maybe goal uh is to become an enshrined roll-up um and um you know I'm I'm talking to the to the school Founders and then you know that that's one of the things that they're thinking about like how could scroll becoming and try and roll up so there's kind of this this spectrum of of Roll-Ups um and by the way my definition of a roll up is you know when there is sufficient data Unchained that you can recompute this the the state and so you know some of the rollups that people don't talk about is centralized rollups so for example optimism today is a centralized Bowl app it's a rollout because you can recompute the states but it's it's centralized and I think another type of roll-up that people uh um you know not looking into but I'm I'm relatively optimistic about this sgx Rolex um where the security engine is is sgx but I think that what we what we're doing is we're progressively fleshing out this this whole design space of roll ups with and with rediscovering you know enshrine based um in terms of the advantages I think you know they really do lean highly towards credible neutrality right so it's like maximum security maximum liveness maximum decentralization maximum Simplicity and so all of these things are you know tending towards the the mature side of the roll up so I what I expect will happen is that um you know the dominant roll ups in the short and medium time will be the non-base ones and that eventually you know maybe a based one and uh actually one of the ideas that we that we've been thinking within the ultrasound team is like should we launch an ultrasound roll up and if we were to do such a thing it it necessarily must be you know a base roll up because that's kind of like the the natural Shannon point um now one of the things that I think will happen with with rollaps is that the is kind of winner take most because of the network effects um and so you know Roll-Ups need to be very strategic right on the on the one hand they need to have a you know they need to move extremely fast have the first move Advantage have a lot of of token fuel to do the boosting but at the same time they need to have this credible story but eventually you know they'll become you know maximally incredibly neutral and I think uh and you're adding becoming a base roll up as one of the line items in the roadmap um could be could be a very powerful strategic move uh for for them and you know one of the things that was uh you know highlighted about baseball Labs is that it's a little unclear how you get some of the services like pre-confirmations um but I'm actually you know relatively optimistic that uh you know with things like eigen layer uh that we we can actually have a fast bit confirmation so we can you know have our cake and eat it to have like all the advantages of non-based roll apps and all the advantages of baseball Labs um so yeah the future is advice for Rolex for sure foreign a lot of different options um I think we have we'll squeeze it in one more since we started this a few minutes late um this one kind of gets to the question of what is really the protocol job to kind of provide um to everyone who's kind of using this ecosystem um so it's a question that I had actually asked Phil earlier as well um so curious to hear your guys take um it's regarding and trying PBS and is this something that we actually need to do um is this something that actually needs to be enshrined in the protocol at any point um it's clearly not a near-term action item that's happening imminently um but there is that real risk of you and trying something a very explicit structure and crypto has a strong tendency of finding new existential risks every couple of years and you might find a new one of those a couple of years later and realize you wanted more flexibility so what are the downsides if we were not to do that and kind of how do you think about that um starting with barnabay on this one sure um yeah a vote field sensor was quite interesting to say that uh basically our job is to do research on providing nuclear options almost like deterrence from for people to to not do bad things and for the market to stay clear um I actually I'm pretty hopeful about let's say the development of PBS Justin can probably talk more about this but lately there's been like a lot of innovation like optimistic willing to want this roadmap to PBS so to me the risk was not to say that okay PBS is a bad idea clearly I think you would want this clean interface of the protocol to communicate with what's outside of it I think PBS is is probably part of the right one and but I did want to see more data more experimentation more research and it feels like we're at this point where okay every day there's like a new paper not necessarily about DBS but that can be applied to PBS or new data sets new new findings uh and then yeah but that makes me silly confident about figuring out that nuclear option like the shape of it before we actually need to to use it yeah I mean my take is that um you know with and trying TPS is going to take a long time you know relative to how fast the blockchain space you know moves it will take let's say two three years at the minimum and in in two three years the Mev space is just gonna dramatically change like you every week to know that the only space dramatically changes we're just speed running the whole thing and I'm hopeful that we'll have some sort of reasonable equilibrium and let's say one one and a half years or two years and then that will be close to the end game of you know the non-insrined PBS and then we'll be able to make an informed decision as to whether or not we really want and trying videos but intuitively as a roll-up operator uh I'd say sorry as a Relay operator I'd say we we really do want and try and PPS like being a really Relay operator is not it's not good like um on the one hand you know this is kind of a lot of work uh and you know there's a lot of pitfalls but one of the things that I've been thinking about is okay you know how can the role of the relay be abused like what happens if someone hacks our relay and I believe that if someone hacks a relay they can steal millions of dollars so basically it's it's it's a high kind of trust environment you know you could connect with the builders you could do it with the validators there's all sorts of attacks many of which are not documented and uh you know really you just want to remove this this distrusted entity uh as quickly as possible um but yeah it is possible that there's some very very clever technique that allows us to remove the trusted entity without entering PBS and I think flashbots for example is working on using sgx you know to remove the relay and I think you know as Randall be mentioned this there's this whole roadmap that we have now that's you know been written by Mike noodle basically documenting um you know a progression for optimistic relaying which incrementally brings us towards I'm trying PBS and I I yeah I encourage you to read it because it it it kind of shows you almost how natural the the enshrin PBS uh should be in uh in in a few years awesome this was a lot of fun I don't want to keep uh Patrick waiting any longer um so thank you again all of you guys this one was really really great