MEV 2.0 - Order Flow Auctions & Privacy | Quintus Kilbourn, Barnabé Monnot

Transcript

all right everyone welcome back to another episode of bell curve before we jump in quick disclaimer the views expressed by my co-hosts today are their personal views and they do not represent the views of any organization with which the co-hosts are associated with uh nothing in the episode is construed or relied upon as Financial technical tax legal or other advice you know the deal now let's jump into the episode [Music] hey everyone before we get into it today just want to give a quick shout out to this season's sponsor Rook close to a billion dollars worth of Mev has been taken out of users pockets and that's just on ethereum and that number is only getting larger unfortunately Rook thinks that it's time for a change and they've

built a solution which is going to automatically redirect that Mev back to where it belongs into your the user's pocket so you're going to be hearing all about them later in the show I'm a huge fan of this team and what they're building so stay tuned to find out more all right this is going to be a great episode also uh welcome back by the way from your from your well deserved break looking a little tan there looking nice um we've got uh today we're going to be joined by uh quintus Kilbourne and Barnaby manol ah hope I got that that potency added right um yeah um it's the last time I ever attempt a French accent could you uh could you give listeners a little bit of meditation of what they can expect on on this episode hell yeah so first of all thanks it's good to be back it's been

three weeks in Italy it's been great um feeling much more red X than than before so I'm very much looking forward to this episode which would be the the final one of the season um except for our kind of on recap episode that then we were provided at the end and today we have two amazing guests um we have panabee uh who was working at Rick so Rick is the robust incentives group at the ATM foundation and you can think of it as kind of the like very kind of Hands-On research arm that's focused on economic problems the economic development so they are thinking a lot about um PBS are they thinking about kind of the effect of Roll-Ups um into Mev supply chain generated in the

future also order flow auction so you can think of them almost as they are looking at all of the things that are happening and how they affect ethereum the economics of ethereum and how this at all ECM should react to these things and so we are working a lot with bana Bay um as part of dashboards and the rest of the rig team and so he I know for a fact he knows like a ton about this and has a very strong opinions um back perfect so uh yeah um the other person is quintus clearborn so quintus is one of our senior researchers at flashbots and so he's uh someone who's like one of the leading experts on ofas he wrote very similar pieces on with A's on exclusive order flow on information

sharing and the role of privacy um and he's someone who I learn a lot from personally just so when I have a question a good quintess and just ask him to explain something to me and so I also know these to get along great and so um I think it's going to be very cool to have them here in the same episode and I think we will talk about um kind of how the Meb supply chain is going through evolve so this would be a much more ethereum focused episode so coming out of the coming out of the cosmos going back into the ether and um uh in I mean we we already touched on so many themes right in in this in this uh season we tested about on all of A's we touched on

privacy division of labor uh we touched on on proposed a bit of Separation with that Cutler and we touched on Suave with Robert and John and this would be the episode where we put it all back together so basically we like we have the floor in front of us and be like take out like the whole Mev engine of ethereum we put it on the floor we dissect every piece individually and then we talk with these two how they relate to each other the interfaces between users and Searchers searches and Builders build us and validate us and um yeah and then ultimately we hope we kind of put it back together and put it back where it belongs that's all the intro we need let's Dive Right In all right everyone welcome back to another episode

of bell curve today Haas and I are joined by Barnabas and quintus Kilbourne guys welcome to the show yeah thanks for letting us make this is going to be a really fun one because this is uh this is going to basically touch on a whole bunch of stuff that we have discussed so far this season and sort of the North Star for this episode is going to be the changing of the Mev supply chain uh but each of you guys are focused on very cool uh components of that and maybe quintus we can start with you I'd love to talk a little bit in depth about something that we've touched on this season which is order flow auctions and I know you are super super steeped in that and we haven't really taught we sort of uh touched on water flow options this thing that's happening at the

beginning of the Mev supply chain but we haven't gone in depth at all so maybe if we could just start from a level you just give an overview of what these order flow options are and how they're kind of acting as this new center of grab mate on that web supply chain so I guess when people hit auctions and mud they sometimes think of the bundle auction where you know the absolutions or traders that get submitting uh bits to effectively to the validator buyer block Builders um to have their you know trades included the arbitrary included or whatever um in higher up in the block when it's in some preferential ordering uh when people would order the auctions it's a very different kind of auction um and really the the sort of like what

defines it all the flow auction is that um usually it's some sort it's a user trying to either get better execution or to get some sort of rebate for for doing some transaction um and the way they get this is through leveraging competition so very basic example is um someone's doing a trade and the trade like um on unisaw or something it leaves an Arbitrage on the table um or maybe they're just they're trying to find a counterparty for the trade take the other side um and through some mechanisms or whatever uh you use competition and usually it's you know training clubs or search engines um bidding and then um depending on how the auction is set up and the best that quite a quote ends up being the one that uh executes the

the order and I describe it kind of vaguely because there's many different ways of doing this but uh yeah that's it that's a general idea great and um so what problem would you say these ofas are trying to solve in general a couple of ways of of uh explaining it uh but in general I think blockchain and Finance in general in traditional Finance as well is very complicated and it's more complicated than the average user can can deal with import this is because even you have all of these dexes all these sources of liquidity some of these those sources of liquidity you don't even know about because they're active right it's some someone with like tokens in a wallet maybe um uh but there's also just some sort of like very like low latency information or various technical information it just

isn't going to be accessible to like 99.9 of people like a state of event you know the like price on binance the couple of milliseconds and these type of things uh and so most most users um blockchains aren't really in a position to um you know make the most effective trade or do the like you know um um the the smartest thing on blockchain but sometimes it's also the case that um you know thinking very abstractive or people doing blockchains it might be the case that someone is just they're doing something that happens to be valuable to someone else yeah it's an oracle update and it's changing the state of the blockchain and and um uh that's represents value to some liquidate liquid or something like that

um so the problem with sort of thing I think is it's in part um making these trades more efficient capturing the like the money left on the table if you made it by trade um but maybe also um capturing with the Surplus you create isn't like they're both doing the Oracle update it doesn't need to be an article up that you can think of us right so um if we imagine some like a user trading on Ubisoft V2 then there's obviously things that that Searchers can do in order to to make that execution more competitive for example by by competing on adding a background to that transaction but um do you think that there would also be um kind of other approaches that we can

go about this um where the user maybe doesn't specify as much um how that transaction should be executed like how do you think that's going to evolve yeah and that that exists right so um so if it uses like if a user is a trading against Ubisoft V2 they've sort of specified the full computational par that said okay listen this is my transaction I want to trade against these pools and this way um and there's like some variance in the outcome based on stoppage um but it's not there's nothing uh you know there's not much black freedom of movement there but um you know we could have users sign in like 4337 kind of a tent so um like car stops another example of

this and then um you know the auction kind of finds them uh the best source of liquidity you know whatever that may be in the child swap example users are being um sort of matched against each other and then some other whatever other source of liquidity that saw was in the auction fine um but you could also just have like um an RFQ system uh like 0x or one inch in which case the user says Hey like I want to do this kind of trade um and then a bunch of different routes are offered to the user different quotes um and this will be like to go Lex such a little dragon films or whatever um finding good routes against passive liquidity so it doesn't have to be

electron kind of thing but yeah different auctions so talking about different options with different trade-offs um so what design families for ofas um exist in your mind and how do they trade off against each other I was I've been asked to like give give a families a lot of Affairs before and I find it very hard to do because I think very small differences uh make a make you know important um I mean have important consequences um but like very generally I think you can um one of the interesting ways of segmenting things is saying okay well um you have these kinds of auctions that allocate execution to sort of like an

exclusive exclusive party so um for example um like I think the sponsors of this podcast uh Rook is actually um Hardware because you you know I'm probably oversimplifying but yeah over one order um and um well if you have many orders but then you have some books so the whoever is hold on to the order at least that's how um it's worked up until recently and then they've changed it but um we hold on to one order um and for some planned period you accept bids uh and then you know the highest the highest bid is allocated the exclusive writer execute this um disorder um

and that's one way of doing things um and what this requires is requires whoever's bidding at the time in the ocean to like be able to Value the um the execution of this order um relatively well and because sometimes there's a delay between you know when they're bidding and when they're executing um and this might be the price moves or that like the cost of executing the block changes and and that might you know um but it introduces some risk for the for the um better than the sale in this auction then you have this other like family that spam exists right now on the two sort of canonical or the two examples I know of um or like share flashbots and uh with blocker and they

do this sort of they do this um they work with block Builders and they go through this process where they connect uh user transactions with backgrounds they say the select the student transactions opening up Surplus somehow usually it's because they um they're leaving some Arbitrage on the table and say okay we'll just connect a bunch of backgrounds and then at the when the block is being built we'll agree with the block board then we say hey um you know if you're going to include this a bundle from this user uh in some sense you have to be including the bundle that pays the user the most um and the benefit of this is that you are able to incorporate more information at uh block building talk um but it is also like a different kind of auction

um in the rough case you'll it's easier to match against like active liquidity whereas in the um nerve block on your shirt can you see all that you're ready like fully fully specified um um you know computational part like I was explaining with the V2 result there's many other things you can do as well but I think that's like one interesting uh division let's come along recently so to summarize you're saying that basically the the kind of the biggest um differentiator between the auctions is whether you have a format where there's two options happen happening sequentially um or one option right and in the two option case first the option of the right to execute the transaction to a

single Searcher or a trading firm or whatever but then they still have to see that the transaction is actually getting included in a block and they have to go through the block Builder option and so on and then the second case you have one block Builder doing like the actual block that are doing uh the entire auction and kind of the OFA is folded into the block bidding process right it's not exactly that that clear cut there's definitely things in the middle like what if you have a block Builder a bidding uh in the sort of the two auctions case uh versus like a Searcher um but yes that's like that's roughly um that's a the way I would describe um separate them for sure

interesting yeah so you mentioned an interesting concept the concept of block building time um and I think we want to kind of put a pin in that and we get to that a bit later uh when we talk about um some of the other Concepts in the Mev supply chain so um thanks for that overview um do you think that there are any things Beyond Trading um that that that kind of you could apply and always A2 any types of transactions that you could send to an OFA that that would also benefit from it that that kind of go beyond what people classically associate with of as which is kind of just like you know subjects yeah well of course of course everything kind of looks like a trade at some point

everything's at all uh that's a meme but um but yeah like a very basic example of an oracle update right or um sometimes there will be like a I mean this is not this is a bit of a strange example but I think it's still still useful it's all um do an nft uh this thing where I can uh sometimes you get the price wrong and if if you did have a disability I know if they you'd be able to capture um some of the value of your accidental bad list it um I don't think that's like the Hopeless of the primary use case I think we can maybe figure out the bad listings to some about but um some other way but um but yeah I think there's a like anything that represents value to

someone else right um and your blockchain during complete or at least the ethereum and like smart contract chains um or effectively and and so there's a whole lot of different things that can happen um maybe right now like tax trade is the most popular thing but who knows when let's go let's go forward um but yeah I guess everything looks like a trade if it represents that value valued or something I have a question just about uh maybe Barnaby this is where I'm calling you a little bit here as well I know you've talked um you know about kind of determining where the line for the ethereum protocol sort of exists and one question that I have maybe connecting a couple different interviews from this season especially the the last interview that we get with with Cosmos

is the cosmos approach to Meb is very much we have an application here maybe it's a decentralized exchange like penumbra there's Meb that's being generated and that's sort of baked into the design of that Dex so I would love you know when you're sort of hearing this um Mev is kind of simultaneously a problem and an opportunity a problem in the sense that there can be negative outcomes for users but an opportunity in that there is economic surplus that's often created I would love to get your perspective on who do you think the sort of problem and the the economic value should accrue to kind of the app layer or ethereum layer and you know kind of speaking from I guess your perspective but you do work at the ethereum

foundation like does how does ethereum uh you know kind of think about getting involved here yeah so as always it's very difficult to make blanket statements for how ethereum considers something it's the idea of the Baza obvious is the cathedral but I can talk about what my my thoughts are on the on the topic and I did listen to the episode with Ari um about Cosmos who's saying that you know the transaction supply chain that's the whole of ethereum like ethereum doesn't stop at the at the protocol site and what you're seeing with Cosmos for instance is you have a much more activist protocols like that adding this layer of programmability of of decision making with respect to to

how the Mev is managed and how it's kind of captured and then redistributed I would say in ethereum the trade-offs are maybe a little difference or artistic constraints are a little different uh first you don't necessarily have all trusted proposals so you always want to create designs that are credible even in the presence of untrusted proposals and that adds a lot more constraint to what you can do with the with a protocol and then the because you are not in this let's say app chain perspective you also need to consider second order effects on on what can happen when you building something that's maybe more favorable to one category of applications and then

might have Ripple effects on on other applications that that could suffer from it so the I would say the perspective has been to be a bit more conservative in terms of the update and maybe a bit less opinionated about uh the way that the Mev is managed at protocol level but that doesn't mean that choices aren't being made because when you look at PBS for instance there is an opinionated choice that is made when we're talking about ensuring PBS for instance so um staying with this notion of kind of application layer versus protocol layer I'd have a question kind of to both of you I know

quintus you've thought about this a lot more but by the way you have like a very keen like understanding of economics and markets so I'd be very curious to you your thoughts as well so Tio limovets from like unicef's former head of strategy went on a podcast recently and he said he thinks that um generalized of Ace will um not take off in the marketplace because um all the applications were kind of developed their own um individual order flow options or their own individual ways of dealing with Mev and they are not it so for one they can kind of specialize more but also they are not interested in Sharing any of the economics

um with an outside protocol so how do you think about this trade-off what will happen basically will applications all succeed in developing their own Mev Solutions or we will see kind of a generalized with a layer maybe on ethereum take off and kind of what are like the arguments um for and against these things yeah I think so there's a lot to say um I think it um because it kind of depends what you mean by application as well like is uniswap the application or trading application um but um so the first thing to to think about here is that if if it is the case that we separate things and you make a very

specific auction for every specific use case uh you do miss the like capturing to the complementarity between different opportunities yeah um liquidations and and orbs and these kinds of things uh right now decorations and like trading odds uh you know with background information see what you get what I mean um but um um I think also you have to think about this from like a centralization decentralization perspective like uh I think it seems kind of undesirable uh to have like the only effective way of using Universal being like going through some centralized server and that also like trying to fits the point that I don't know if that's what he was suggesting and

so um it might be the case that like multiple um auctions that that uh um work for for a specific um application why like General preference and I I mean many people with dashboards have spoken about this very interesting idea I think is to have some sort of like General um General layer that acts like a sort of like a programmable auction and I can facilitate many different kinds of auctions for different kinds of applications and the benefit of this is that you're not stuck with you know the one uh auction this is like one front end is offering you and that's the only way of all the accessing you know uniswap or whatever because

you're trusting some um some server they set up but but rather they have a world where um we have many different auctions and they're running in some way that you um you trust um whether it's like you know it is in some sense the centralized um but also that allows people to like go pitch different kind of auction formats if you know this one auction format is working some across experiment the problem with uh like running trusted auctions is that it's very hard for someone else to launch other auction right now the um the biggest auctions have spent a lot of time and they're garnering trust for people trying to get water flow adoption um you know the people involved with blocker will learn how soft spent many

years and it comes out of gnosis um so there's a lot of um like reputation that goes into this and so it becomes very hard for someone else to go like hey I actually have a great auction idea um let me you know set up an auction because it takes them like a lot of time to gain trust from people and and only possible so if you can provide this with like a more General auction layer that can be programmed to be um you know facilitate more specific kinds of auctions uh then you um I think you end up doing quite well um in that you allow people to compete and propose different kinds of mechanisms and and how these work up well but still capture the survey specific specificity of different kinds of what you needs because yes it is true that

um you know maybe the the kind of auction which is best for the next trades the looks a bit different from the should work soil for eight of teens or something like that adding to that um would say there are fairly strict uh results on even the computational complexity of running auctions and so if you're trying to do like this generalized matching engine like you know for sure that this is going to be basically impossible to solve so the question like do you can you put that at protocol like can you can you make the ethereum protocol basically organize the suction seems to me like it's it's a it's a tall order that that's going to be um yeah running against some some fairly

essential or fundamental limits so then the question is if you don't do that what what do you do and I think that was kind of a gist of high school's question which is that you know will we see his wise auctions of just like separate flows for for the user transaction to to go through application per application I think at the beginning we will see that because it's just it's just easier to do when applications have kind of the finger on the on the bottleneck for the user transaction flow um but as Queen just mentioned there are complementarities between these these applications and Iberia complementies complementarities uh there is value and if there is value there are people who

are going to to innovate on formats and figure out how to how to capture that that value but then I don't think we'll ever get to a point where this converges to a generalized auction that that solves everything and matches everyone at the same time and so should the protocol then be involved my perspective is that when when it does we are kind of logging in a certain type of allocation mechanism and perhaps we don't want to do that perhaps we want to preserve the optionality of yeah innovation in terms of the auctions or even the market structures might might change over time so so we we want to give ourselves as much room to to be agile and to and to

let the market kind of figure out the best mechanisms um I would say that everything is already one auction at some sets right this is already you know everything needs to go into the same block and so it's all economically interrelated already um and the question really is like how many you know different is there like you know a different route that everyone takes that everyone goes through some like specific auction burst and then it all Aggregates together later um uh or like you know I guess one sort of like more uh dystopia Utopia perspective was on this is like yes you do have this potential world where you have many different auctions you have one where like there's like three or four

um well I don't know some small number of like servers running in like USC's one AWS um um you know executing everyone's trades and then you have yeah different world different kinds of auction outcomes and so so yeah up in the air and it's very much I think there should be like a a cold war like decentralization and pushing forward that so um that's also I think the more like a important one to make it's not very outplaying and they're like worse fascinating guys um I I want to kind of close the job down of s for now and and move to the next one so one thing that we always discuss and I think one of the original reasons why we're here is is kind of the the problem of vertical integration would only

um I think the fear was that um that might increase would integrate with Searchers and then there would be one dominant mining for it um that that kind of can crowd out all of the other ones that would lead to centralization right and men this has been many years ago and now of course um the LED supply chain is much more complex than just mining pools and searches um but the problem of vertical integration remains and arguably it's it's bigger and kind of more elusive than than ever before right so we are seeing vertical integration between wallets and Searchers between Searchers and block Builders

um between block Builders and relays and um and who knows um what else is there that that we're kind of not aware of because these things are very hard to prove so um could you maybe reiterate why vertical integration is such a big problem for the Meb supply chain and how we can address this problem at a high level or kind of how you personally think about addressing this problem at the height of it um sure so the first question is why is vertical integration such a problem and I think it really comes to why we are building the systems that that we are building which is that we want systems that are efficient but we don't want systems where the

efficiency can be of the extra Surplus that the system creates can be extracted by a single party and so when you have vertical integration you basically take two entities that would be in a in a relation of of checks and balances or maybe intention with one another and now you you you're kind of merging them together and so you might be expanding uh the action space for that bundled entity to to extract more value through whatever is is going through its uh its pipeline so why we why we care about that cultivation is really because we care about having systems that are Loosely coupled and and

each part of the system or each component is unable to to extract any kind of rent on what's going um through through its pipe and if you have vertical integration you you also prevent to some extent competitivity because now uh if you have one Searcher on one side and many builders on the other side there's at least a bit of competition between the builders to to offer something to the Searcher but if the search line Builder are the same uh now they they can do strategies that are that are different and they don't need to be quite as as competitive and so yeah Building Systems that that prevent any kind of to be extracted at any point I think is what we're here for and so vertical integration is a threat to that ever I have a follow-up question for you

there probably um I would love to understand how you think about Builders and specifically Builders getting larger and more sophisticated I think you know one of the the parts of the sort of like a design principle that's been extremely successful at least in ethereum so far is kind of this idea of making basically leveling the the playing field for proposers and making sure that they're adequately compensated for securing the ethereum network I think one of the the trade-offs there and I'm sure we'll get more into the Weeds on all their proposed or Builder separation um is that you have kind of these larger and more sophisticated Builders and without going too far into this we had a whole episode on the modular uh Mev in

the modular world but it seems like that's only going to be kind of exacerbated as we moved up the stack to roll ups so I would love to get your thoughts on is Builder centralization something that you spend a lot of time worrying about or what's what's your sort of thoughts there yeah we do worry about Builder centralization insofar as we care about the signal that PBS is giving us and the signal should be equally the same for for all proposals and so if you now have builders that are more powerful and that capture like private order flow in a sense that can um add some noise to that signal or destroy some of the value that would be going forward so if you don't have the

same competitivity on on the Builder side you don't let the value propagate towards the towards validators and eventually towards the the protocol who's who's behind the validators and so Builder centralization is a bit worrying in the sense that now again you have this one actor in the supply chain that's able to to extract and you rent and which maybe is is not compatible with this idealized version of a builder which is more of a simple pass through but uh that forwards the value towards the proposals and the protocol eventually quick break in the show here I want you to imagine something for me imagine swapping two stable coins on chain paying zero

dollars in gas and instead getting a rebate of two thousand dollars this is something that's actually happened on chain to understand how I want to introduce and thank the season sponsor Rook zooming out for a second the current state of affairs at Med is billions of dollars so far have been extracted from users Pockets using Mev Rook is coming in and saying enough is enough blockchain should drive value to their users and the applications they use it is time to leave the hobbyist era behind us if we want to move forward we want to get this right that's why Rook has built a custom blockchain settlement Network and it's one that gives you full control over the entire transaction life cycle today you can connect to an open source rep node The Rook protocol will automatically match bundle and auction

your orders and transactions in seconds with zero gas overhead also any Mev that's discoverable along the way will be returned to you the user created as a collaboration between the industry's top mechanism designers and Mev Engineers Rook was built from the ground up to be scalable safe and programmable you can get your own mempool choose searches and Builders and link your mempool with others to discover even more Mev you can Define how the Meb is shared and delivered as well and what could basically process anything from transactions to meta transactions and more this is the way that blockchains basically should have been from day one so if you're a user listening to this here's what I want you to do I want you to go to your wallets go to your favorite app your node provider and say Hey I want you to be

working with these guys rock I want the Mev that I create to be redistributed back to me if you're a developer and you want to stay ahead of the game the best way to do that is to follow them on Twitter they are at Rook or even better yet slide into their DMS they are lighting responsive they'll get you set up today and if you do slide into those DMS as always please tell them that I sent you quintus I'd love to get your kind of tag you in here and get your thoughts on just vertical integration in general and then yeah any thoughts that you might have about Builders getting larger and more sophisticated yeah I am I think I bought any sort of caught the Crux of it I think political Integrations you know sometimes it makes sense like it's just like you know the

consensus like they are an execution there like oh what a clinic in some sense but uh makes sense two things apply it um but in the like searchable the case maybe it's less desirable like the like such a round data case that's desirable because then you have I'm like centralization vectors and um uh I think one of the interesting things I can say about other vertical integration about uh in like block building searching today is that I think Potter would come what a way it comes from is that there's you know all of these different searches and they don't they aren't able to collaborate well uh in this in the sense that um you know we might be going we might

be like both uh doing some strategy that touches the same like Unity B3 pool or something changing that price but um because we don't trust each other there's no real like making them for us to do it um I can't you know collaborate with some other Soldier and say um you know let me know what just what state you're going to leave the pools in so I can like better calculate I'll show you for that or um things along these lines and so what Builders do is they play this role network of the circuit we have to aggregate all this information well all these people didn't know out of like From the Block together and I'll be in the SEC special position we have all the information I'll put it all together um and the way things are set up at the

moment that's necessary but what that also means is that um these Builders have like a lot of additional information um which is like super useful uh in in searching hft and traditional Finance as well uh and that's why you see a lot of vertical Integrations because I have this um sort of informational Edge it really makes sense for them to take it doesn't make sense but there's a lot of economic incentive for them to take advantage of it background bundles you know and so there are trades given all this kind of information um yeah I think you can sort of see where which thing you can address developer yeah so you said search US can't collaborate well on on the one hand and you also said um they just have all of

this extra information and so these might be the same thing right um but extra information and kind of um also like a latency Advantage so that they connect later in the game um in finance terms they have what we would call it the last look right they can kind of react closer to the actual event that matters um how do we how should we think about like dealing with this um in general like how so it doesn't build us they arguably have started to become kind of the same party because there is a big economic benefit for us so I decided to be a lot better or for block beta to start searching although arguably kind of follow the skill is in the searching and

not in the block setting um and so it's it's kind of you're seeing kind of the integration that way that all of those parties who originally pick searches they have now become big block bidders we haven't really seen it the other way around that there's someone who's like uniquely good at block building and they then branched out into searching so we're not really seeing that I think that that that's a nice indication of where the actual power resides in the supply chain but how should we think about kind of breaking these two roles apart yeah I think you can think of it uh sort of like thinking of the um the fundamental requirement of like okay all these different strategies after like all these different transactions or whatever end up with the final block uh and then there's tension

which you find in in the highly sensitive environments where your information is super valuable uh and so if if I if someone else knows what I'm bidding for an opportunity or what Arbitrage I'm doing or whatever it is um that means that they can act on that it might not even be that they're on a you know explicitly um you know copy trading need is like a much more subtle things they can do just like sort of um um maybe derive what kind of prices I predict in the future based on my trades or uh maybe just a bit a little bit higher when data share bundles other people something like that um but the point is that if I'm a Searcher and I'm sharing my information with with uh with other entities I'm making myself vulnerable and at the same

time um you know we need to share information to some degree uh because somewhere this block needs to be created somehow and of course you can do like this completely private um building thing but but then we lose you know everyone just sends some encrypted Global transactions and then we all randomly order them um maybe we spam to get better ordering I don't know um and then um whatever that random ordering is or like random but like information disordering is uh we go with that and that's I think arguably also undesirable uh because you lose a lot of economic efficiency um and so now the question is you know how can we um

how does one facilitate collaboration between um all the different the users right and people like Anthony's acting on behalf of users um and you know people who just want direct access to box base and searches or whatever but sophisticated users how do you facilitate this kind of collaboration um and it sort of in my mind that's and you know if I think you've heard flash floods people speak about this before um is the second notion of programmable privacy uh in the sense that you're allowed to you're able to you allow users to walk this line between sharing information that helps them achieve better ordering the block better outcomes whatever collaborate with with

other searches and at the same time um preserve their their privacy even from a user's perspective but if they're users submitting to um some Castle uh trade uh what I don't know I'm submitting to that share um I'm currently trusting flash watts is is uh sort of doing the right thing um with my information uh and only you know acting a certain way and the idea of programmable privacy is to say listen I'll only expose the information necessary to flashbots to be able to facilitate this kind of auction but not bad do you see any parallels between programmable privacy and kind of the protocols that you have been working on is there any

anything that that kind of comes to mind that connects not immediately comes to mind but uh yeah I would say it's definitely part of the part of the stack but because again if you want the the Builder to to reveal uh a good signal of what the value of their blog contains and you essentially want the Builder to to be a bundle Merchant so you want like the time that the information leaves the researcher and gets to the Builder that this this information is not really actionable anymore apart from just putting it in order in in the blocks and so so you really need that programmable privacy path here to First ensure that the Builder makes a commitment that they can't touch whatever has been given to

them and and before that that uh yeah the the users have not been armed during the process of extracting the bundle but did you have something in mind when you were asking that question well maybe just the idea of of kind of credible commitments and how they can facilitate collaboration which which I found to be kind of a connecting theme in a lot of your work yes yeah so we see something that I've been thinking a bit more uh also in collaboration with uh Shin who's at flashbots uh first the idea of commitment which is yeah who's who's making commitments and who's kind of recording them so I've been writing a bit about this um the protocol making commitments is

kind of the most powerful commitment that you that you can get around if you believe in the accountability of the of the validator set and so that's that's a strong commitment you can make but you might also want to be able to have uh commitments at lower levels of uh of the stack and so you want Builders to commit to note front running you or to commit to just doing uh honest bundle merging uh you want all the free auctions to commit to rebating you some of that value so this commitment parties is useful beyond the protocol and it doesn't necessarily always have the same internal mechanics of how these commitments are are

recorded and that's where I would say the credible part kind of comes in so I've discussed a bit in a in a recent post how I think about credibility and in particular The credibility of of the ethereum protocol at making these commitments and that enforcing uh the mechanisms that are enshrined uh in the protocol and the protocol makes it critical because essentially we we have the weight of the whole Community which is which is delegating the the duty of uh of enforcing these commitments or enforcing these mechanisms to to a protocol that is executed by by a set of validators that are that are bonded and that are staked and that have a lot of

money on the line to to kind of uh enforce these with this commitments so that's that's one aspect of credibility and then lower down the chain you still have commitments but but the credibility is a little different so you you might have commitments that are written on Roll-Ups for instance in that case uh there's a trust model which maybe is a little different you have commitment from the Builder side which Suave is trying to make so there's many I think yeah what what this work has kind of told me is that I don't necessarily think of mechanism design or credibility as like zero to one thing like either of the mechanism is there or it's not there it's yet to meet more complicated than that and it's more multi-dimensional and credibility is a is a big part of why this is the

case by the way maybe this is a good time to transition into talking about PBS um and you've done a lot of thinking about this um I'm sure privately but also uh publicly as well so I would love to kind of get your thoughts on the the current state of PBS which is sort of facilitated by the the off-chain med boost solution and then I'd love to kind of get your thoughts on you know what what and if you could even just like sort of explain for listeners like what was the decision why was this decision decide to be enshrined in ethereum and what does the roadmap look like towards that um yeah so what's the latest on PBS is so PBS has essentially been live in its in its current form since the merge uh

and and the in the form I'm talking about is you have validators that are sourcing uh whole blogs from a network of Builders um I think that's been very interesting for us at protocol designer because now you you kind of see the shape of the market like how the market has developed uh so you have I don't know it's been interesting for instance to observe that there's been more diversity uh in the relays now than those at the beginning more diversities the builders though it seems that they are also precious uh towards more centralization still so seeing that live uh kind of out of protocol gave us a lot of I would say useful data to to think about the think

about the problem on my side I've done a bit more work to think about it theoretically so the question I've been training PBS is essentially the question of enshrining an option in protocol and so when you ensure injection there's a lot of parameters that you that you kind of need to to decide uh one of them is do you auction of the right to make the whole block or do you optional the right to make a partial block or even partial blocks uh plural how do you think about the enforcement of the Beats how do you think about uh The credibility of the of the auction uh once the value is is kind of elicited by the auction how do you how do you manage it like what do you do

if you give it back to the proposal you give it back to the to the protocol so there's a lot of questions that are interesting and before thinking about these questions I had fought quite a bit about eap1559 they are very interesting parallels between the two mechanisms in terms of in both cases you're trying to elicit from the perspective of a protocol on a signal a value that is credible and that actors essentially validate us called bias and so yeah these questions have been to me quite fertile so I've been thinking about them quite a bit since the since the last year and and I feel yeah we we've reached a much better uh

understanding of what the the auction could look like we understand a bit more some of its trade-offs and so what it looks like in terms of getting it to to ethereum I think is yeah there's some really nice work for instance by Mike Noida and Justin Drake on the on the optimistic relay and that includes an actual roadmap of making the optimistic relay as close as possible to to what would uh and Shrine PBS look like so yeah I think this is this is probably how how that will go I had some doubts with respect to enshrining the auction because again there are trade-offs that you need to make and so always wondering how is the right Traders and once it's in protocol you're

kind of yeah you might you might be you might be stuck with it but I feel I feel a bit more confident now that we we've seen more data now that we have a real great path to putting it there so yeah that's kind of how I see the state of things what do you think are the current problems with map boost well there's one problem that has been revealed uh pretty dramatically by the attacks that happened at the at the beginning of April which is that um well maybe boost didn't deal very well with equivocating proposals and so even though the system is designed with the idea that uh yeah untrusted proposals can access this market and there's this Fair exchange that happens

between the Builder and the proposal in this specific case it kind of it failed and it failed for a reason that I think was just not uh imagined back when uh back when the system was was designed and so yeah that's obviously it's a it's it's a bad thing that it happened but it also gave like uh some useful data in terms of hardening both math boost there's a proposal by Mike and Francisco called the headlock and in fact that proposal carries over to to how we think about and trying PBS so PBS in protocol so yeah that's I would say one problem the other program for me has always been

the whole block auction um selling away the entire right to make the block I think where there are some examples that show you that this might be sub-optimal in terms of the value capture might be more centralizing in fact quintus was one of the first I think to to discuss this in a in a nav boost issue and showing that you have a centralized Builder we could do better even if the decentralized Builder is able to inferior aggregate uh more value in that case the centralized window would would still win so yeah I still see some unknowns on on that on that question but yeah again with the perspective of putting it in protocol and giving ourselves a path to to

experimenting a bit more with it uh before it reaches there and maybe even after it reaches the protocol uh I I think this is uh yeah these are the two main problems I see I've got a question for you born a bit this is this Harkens back to a debate that we we had earlier this season between haasu and Matt Cutler and it was very centered around relays specifically but maybe we could broaden it a little bit to just talk about off-chain infrastructure in general so something like relays that the plan to enshrine PBS is eventually we're not going to need relays right and they're kind of this centralization Vector but on the other hand if ethereum is going to be very careful about the line for what the protocol does probably it's pretty reasonable to assume into

the future we're going to continue to rely on off-chain infrastructure for relatively key functions of ethereum so I would love to sort of get your thoughts and maybe we can start with relays but kind of broaden it out a little bit to infrastructure that we might not even know we need do you think it's how do you kind of weigh the the costs and the kind of risks and benefits of hey we don't want to rely too much on something like a relay so we it shouldn't necessarily be do too profitable versus hey there are off-chain actors that are providing critical infrastructure to ethereum and if we want that to function smoothly we want them to be economically incentivized to keep providing that infrastructure so I'd love to get your

your sort of thoughts on that cost benefit there first it's clear from even vitalik's end game post that we we're planning to rely on on of chain agents to to perform what you could call call or critical functions of of the protocol and and this is really coming from the asymmetry between the cost of constructing the block so finding for instance the validity proof if we were to enshrine a ZK VM at protocol level um finding the most optimal blocks which is what PBS does optimally in terms of the of a value that they that they return so yeah definitely for sure there is uh kind of this emphasis on on off-chain infrastructure to to perform

these things um but then I would say as you mentioned there is a difference between relay days and and Builders which is that relays a kind of a crutch that we have at the moment because we we need uh MAV boost to function and and if we don't have a protocol to to enforce uh the auction or to be the credible option you then we need someone else in the middle to basically broker the agreement between uh the Builder and the proposal and that person is is very relay but but we perform I would say the Nerf chain rule that is temporary in in the perspective that and trying PBS will will happen and so they don't they don't I would say forests have the

same Dynamic so obviously economics as other entities that are there to stay in the in the long run like like builders for instance and so yeah Builders have kind of been this go-to black box so whenever there's something that requires a bit more computational power when we're willing to to assume all the proposals and we say oh you know Builder can can do it and so um we have proposals now for statelessness where you you have to provide some kind of state to the block to execute it and if the proposals are stateless then the builders are going to be the ones who we need to make that provision as I said before you might have validity proofs in in the future and so it does seem like we we're

relying more and more on this Earth chain actors and again this is very opportunities to to this because yeah we we do it's almost like adding extra computational power to the network as a whole even though that computational power is bounded in time so so we want to think carefully about how do we enforce and create the relationships between Visa of chain infrastructure and the and the protocol itself yeah that's my another long-winded answer but where it can be more precise if you if you like are you worried like it's the ATM Foundation worried about um creating two much pressure on Builders and to to do

additional tasks that could ultimately erase the barriers to entry I think it's a worry if the control and competitivity of Builders or rather the control of the builders and the competitivity of Builders doesn't scale as fast as the number of tasks that we that we give him so if we find that you know there's a there's a good enough market of Builders doing doing the blogs that might put us at ease to add more things to the builders so when I say earlier we censor I said oh there's like all these roles that the builders will have to do but obviously the base of upgrade is is not super fast so we will see over time I will I would

say the the cracks in in the in the Builder Market as we as we appeal and we might take that into account uh when we start thinking of you know Builders should do like a million different things uh Etc and again we speak of Builders as like this um abstraction which is more like you have this um nebulous entity that is that is performing things on your behalf but it's very clear that Builders will not be a central server that is just outputting everything that you that you need them and that Builders themselves will decentralize builder of themselves we figure out that there are economics that work better in their favor uh when they are more more decentralized so so yeah I think we can also lean on that

um when we when we think about uh the role of Builders PBS is basically like one component of it is better than us making commitments to block Builders to include a particular block and that allows kind of this this new layer of Commerce to happen between Builders and validators um you already know some time ago you made a proposal that seeks to generalize this idea of validators making commitments to parties in the supply chain it's called Pepsi proposal commitments yes could you briefly go over like what's the kind of line of thinking behind this proposal and and also what kind of new use cases would it enable in the enemies supply chain yeah

so this proposal came at a time when when I was feeling a little concerned about enshrining something too specific in the protocol and that might not give us good optionality if the market structure changes or if we figure out that yeah that specific option format is is not actually suited for but what we what we're trying to do at the same time I was quite interested in eigen layer and this idea that you can have proposals basically signing up for for making commitments and and making these commitments credible by by putting their stake that we currently have in the ethereum protocol at stake for overall

duties that they and they are then signing up to to perform so I was excited by this because it gave me a sense that oh you know there's like we saw the layer of innovation you can do that is still close enough to the protocol but not the protocol itself so yeah gives you the flexibility without giving you the committing to like this uh this specific mechanism um but at the same time I was seeing some risk with eigen layers so the first risk I was seeing is that the protocol might not know that the validators are committed to eigen layer and if they do bad things on eigen layer and eigen layer kind of writes down that a bad thing happened the protocol might not know and so what's at stake for the validator uh we we with respect to the if your protocol might not actually be

the real mistake because we we got slashed of band by by again so I was proposing first okay we we need to at least ensure that this principle agent problem between the protocol and the and the validators is uh is resolved by by aligning everyone's view on on what is what is happening and what proposals are are doing outside of the of the Bell view of of the protocol so that's the first step and then the second step to Pepsi is saying well what if instead of them signing up to to these additional slashing conditions with eigen layer we allowed them to programmatically create uh essentially validity conditions for for their blogs and so saying instead of

waiting to see the validator did something and then they get slashed which I called optimistic validity we we make them write down what their validity conditions are and we evaluate the block validity based on whether these conditions are satisfied on and this is let's see a pessimistic view of of blog validity and that is it's still a theoretical proposal but it's something that we've been spending a bit more Cycles thinking through and trying to to do a proof of concept uh of this system but essentially what it does is it decouples PBS between two things one which is the market structure which is that there's a legal system that ties

proposals to off-chain entities and the legal system requires that the exchange be made Fairly between between the two and it and the second part is the is the allocation mechanism so what how exactly what is the shape of a contract between between the proposal and the and the and the Builder so when you do PBS the way it's discussed today you you do both at the same time you both enforce the market structure and you enforce the specific allocation mechanism which is the whole block option and so Pepsi was the idea of what if we could only enforce the market structure but validators were completely free to enter into any commitment that they that we want to enter into and there may be plenty of reason why that's a bad idea uh because there are

maybe commitments that we we don't prefer the ideators to enter into such as sensorship or reorgan and On Demand but there may be ways that this gives us a bit more freedom to to explore uh what proposals can do amazing okay that that wraps up our segment on on PBS and I think we're entering kind of the last uh the last uh concluding mini chapter here so um because I I do want to talk about putting it all back together now that we kind of we laid out this entire Mev supply chain in front of us we deconstructed it we talked about the interface between users and Searchers and block Builders with ofas we talked about vertical integration

between um Searchers and Builders and why it happens and that we need division of labor and so on but then we also talked about kind of um the interface between block Builders and validators so quintus um kind of in in your words like if you zoom out five years how do you think the energy supply chain is going to look and also maybe as a side note where do you think kind of um Suave will fit into that yeah I feel like it's a it's a mean by now obviously that's what just solves it all um but but um yeah I I think you know if you want to be really realistic about it uh there in a couple of ways that like the things to go from here um there's maybe like some worst case

route where um things centralized but just enough so that like there's no like real reaction from the community and um and so you still have like you know a high degree of centralization but it's like like minimally acceptable and that's like our all of things and maybe that's like a couple of centralized ofas and a couple of like centrifice Builders that's that and that's really if the world we're trying to avoid um and so maybe to like zoom out to a wall that so I was trying to create and that I you know believe will be the world that exists uh in in by this time uh is this world where yeah you still have an

interface between um you know the protocol and this Market uh and maybe electric one is quite maybe makes sense for it to look a bit more expressive than it does right now it's not just a full block auction at some complicated The Proposal commits to or who knows um um but you know something like a PBS in the sense that there's a market for producing the next block uh and then so with swathfits in is really everything before that um and I said that um you know we want to facilitate collaboration between uh the different users who need to end up with a block um and part of that is um users you know who want to

participate in OFA um you're trying to with the auction trying to capture um sorry get that execution these kind of things so you want to provide them the tools to um share their trade share their information without being you know front drive or screen over and but also give them the power to like really Express hey to give me um the best possible trailer display on the block or or whatever um and the same time for Searchers to um sort you know ideally provides them with the ways to protect the information but also to give them enough information as the blocks reconstructed um so that their strategies all um sufficiently valuable

um and maybe maybe you know 12 provides other kinds of information at block building time so that um you know high frequency or low latency kind of races aren't important because you know we have there's a binance price Oracle at uh um block production climate so people aren't racing to get the most I think the most up-to-date that balance prices in their bundles and there's all like a bunch of details with zooming out what does that like ideally look like um it looks like they're sort of like um set of Primitives on the sort of like middle layer that exists between all of the users um of the protocol really like sophisticated searches or yeah uh just being new

um and um you know through this through this layer uh We've allowed collaboration and and um and hopefully production of the most angle block possible um from like a welfare perspective not just from the latest perspective um but still available enough to compete with centralized uh venues um and through this like PBS like auction with the with the value of the century produced box and ethereum that's like a very very high level view with a couple of specific details yeah Barnaby I wonder if you have anything uh anything to add to that hey I don't have commercial add to to review that quintus kind of presented I

think that's also I as I hope the most decentralized possible uh transaction supply chain but maybe I can add a bit more color on on what my thoughts have been which is I've been interested lately about the path dependency of how things sort of happen uh and and in particular how protocol development is is somewhat oriented by what's happening in its in its environment and if I was to try and draw very rough distinction between Bitcoin and ethereum is that I see Bitcoin people thinking about Bitcoin as like this you're discovering the protocol like you're discovering it as a as a platonic shape the way you you discover like a new physical law or kind of mathematics and and I would say if

the room is is somewhat still has this retention of finding natural laws of its protocol like where the enemy should go how to how to do things but there is still much more responsive to to what's happening on its on its layers and so if for instance uh there was much too much centralization or ultimate censorship uh to our taste I think we would look more Accent on things like inclusion list to defeat censorship or uh pre-mempoon consensus to ensure that transactions have to be seen by by a large set of actors before they are included and not necessary transaction

needed clear but you could you could see commitment to transaction and and then ensure that okay blocks have to be made with transactions that everybody has seen so there are ways to kind of step in and and maybe backstop some excesses of the of the market but these ways they yeah they're not like platonic shapes of protocol they are more like a real political okay maybe actually we we see the market like heading off to the wrong direction and and of the market is not maturing uh fast enough to to defeat the the censorship or centralization pressures and then you kind of have to maybe do something uh about it and so yeah I'm very turned on on what the right approach to that but but it's

something that I've been thinking about a bit more and I think it's relevant here fantastic thank you guys so much um for coming on the show today um I think you both provided really great perspectives and um gave our viewers um a great overview over how the Meb Supply is saying like what are the individual components to look at and how those are going to evolve in the future hopefully but if not also what can be done um to to save the day so thank you it was a pleasure all right what a great uh great final final interview for the season that was that was a ton of fun and both quintus and barnabyte just such deep thinkers

um I'd be curious you know what uh what kind of takeaways do you have from the episode oh well first of all I agree I think it was a fantastic episode I I love hanging out with quintess and Barnaby both of them I think they're both like they have such a you know calming attitude and they are so they do they are so um good at explaining things right um they always can take a step back and you know put things exactly where they belong in context and and find the right words which I I find really impressive so very often when I have a question I actually reach out to one of them and so it was great to have them all together um yeah and um and talk about the Meb supply chain so I thought that we

covered pretty much everything that I I wanted to cover um so I I was very satisfied yeah mean as well I wanted to get a little bit deeper in the weeds with you on auctions because I I thought that was a super interesting part of the the conversation but I wanted to to sort of get your thoughts on that a little bit more and go in depth a little bit and I know you you mentioned to me earlier in the season I think with our in our episode but to ruined that you are a little bit of a nerd about auction Theory and and kind of like going down that specific Rabbit Hole so I would love uh I'd love you to kind of unpack you kind of listed your own sort of segmentation about different types of options that exist can you sort of just

rehash that and then talk a little bit more about how you think about the different maybe just expand or clarify or just go a little bit deeper on the design the different design families that Quinta spoke a little bit about at the top of the episode right yeah so between the three of us I I think I'm like of the two guests and myself I probably the least of an expert about blockchains and quintus I think uh I I think um this was one of the things that he studied and so yeah I think he's really smart about that and most of what I know I already know from talking to people election and and quintus um Internationals so um well it's a good question so um I think quintus uh he he gave a good

overview when you said that um that there are kind of a few ways to to slice the pie when it comes to to different order flow options um he what he drew out in particular was whether there's kind of one option where that all happens kind of inside the block Builder and you rely on kind of the computational power of the block Builder but also on uh what he called kind of block building time right this the block beta bits the block at the last possible moment and they have all of the information about all of the other bits and so they can really optimize for kind of the optimal the most optimal block right and um and that I think that's one of the main ideas behind so meth share

um the the RFA that we built but also um behind any blocker so that's kind of another OFA that's out there right now and that's in contrast to some of to some other ofas that that that um for example are of Q based right or the one um that quintus mentioned uh which is Rook the the current sponsor of the show um which kind of have this two-step process first they option off um the right to the particular transaction and then that transaction still has to enter the Builder auction in order to get it included in a block so I think that was very Illuminating um and um on the other hand I think one big trade-off is going to be um so specialized versus generalized I

think we touched on that briefly um when I kind of cited to liebovitz um so that's going to be a big question like how specialized are the different order flow options going to be so you can have another function it's very kind of specialized for example on trading um or or back running Oracle updates or any number of things but then quintus was hinting at something which is kind of the complementing um a kind of complementing use cases that you can unlock from having kind of one generalized order flow auction right so there might be different uh different submissions to different types of order

for options that already seem related that from happening inside the same auction unlock some kind of combinatorial benefits whether it's batching or whether it's kind of matching things against each other in the sense of like cows or Coincidence of ones right right um and um so one example actually so I used to be kind of very engaged in the makeup community and um so make us liquidations so when when meca was liquidating a CDP that trade was actually routed to one inch and could definitely match with people trading in the other direction right and so um that's a kind of a small a small detail that even something that that's it seems kind of obscure like is like there's a CDP getting liquidated by this

quintus was saying everything's a trade at the end of the day right and for every trade you can have a counterparty um and um kind of the the more information you aggregate in one auction um the higher the welfare that you can generate um from the auction and so that that's really important and then the third thing that I kind of would draw out is is that of kind of um the the auction credibility yeah yeah so we already had um to run on so I think he's kind of the foremost expert in crypto on kind of credible auctions he had this really I think it has already had two really great papers on this topic

um and so the idea of option credibility means basically is the auctioneer um going to be honest are they going to stick exactly to kind of the words of the option that I laid out both in terms of are they respecting kind of the privacy of bits are they inserting any of their own bids are they a censoring any bits or many bidders are they leaking any information maybe outside the system and um so really there are all kinds of things that auctioneers can do um that kind of undermine the Integrity of the auction we had already discussed one example with kind of Google inserting bits into their own ad options to drive out the price and now

um they they kind of facing a big antitrust lawsuit um well deserved for that um but all of this stuff happens on a much much smaller scale right so um for example in the block building Market there's complaints from Searchers that particular block Builders are censoring their bundles or that they are looking at their bids uh and kind of just like one-upping them um or that they are stealing the Mev right so this is all kind of things that that can happen and a lack of a lack of trust in the Mev supply chain really leads to more vertical integration between it makes collaboration because

it makes collaboration between different actors harder right because if I don't trust you then I'm less likely to send you my order flow for example I'm less likely to send you my trade um and so um what this means in particular for OFA is this I think that um kind of you you can gain some more like efficiency in some ways by by running more centralized infrastructure but you can get more social scalability um and more credible neutrality by making a more credible option uh one that that kind of has its rules and shine on chain and that doesn't really have any trust assumptions in the option here itself

and I think this for me these are kind of the three criterias um how you can kind of slice the of a pi and then you can fit almost any kind of Auto flocks and any kind of trading scheme into that yeah I there's so much to unpack there and one of the things that I'm sort of kicking myself from from the notes for the pre-call that we didn't get to was this idea of exactly what you were describing and The credibility of the the auctioneer which was I think it was quintus that called it auction extractable value so sort of a fun play on an Med but just obviously there's an enormous benefit if you're the auctioneer right in the case of Google inserting your own bids because your incentive buys for you know a more uh you know you want to make the the ad

Market more right exactly uh but I would love to get your thoughts kind of can you give your own thoughts on teo's argument for you like uh kind of where the app versus the protocol that ends up sort of uh succeeding because I I thought it was a really interesting observation from quintus that their sort of already is one auction at the end of the day right it's yeah one block right so everything already is economically tied and even if you look at sort of the role of uh I think it's an executor on Cal swap but basically it's you know these executors kind of have carte blodged to kind of go off and find you know go off through the land of ethereum and find the best execution and deliver a surplus to Traders on cow swap in

general so I would love you to kind of tease apart like how you see the different incentives playing out because obviously these exchanges want to internalize some of the economics that they're creating but then I also see the benefit to being a completely credible sort of auction system so I would love to just see get your thoughts on how you see that ultimately playing it yeah so first of all I think the concept of auction Auctioneer extractable value is super interesting and then once you start to look really it's everywhere right so look at the kind of cftc lawsuit against binance there's at least five examples where kind of Finance was also like extracting Auctioneer extractable value on behalf of its us whether it's from like selling preferential access kind of to like

um to their servers or whether it's kind of trading on their own exchange and so on it's like this stuff exists actually everywhere and I think that's that's one of the major bookcases for why more and more economic activity is actually going to happen on top of public blockchains because they are simply more trustworthy you can design more credible mechanisms on top of them and um so at least kind of that's for one for us like internally one of our and off stars is like that all auctions should eventually run on ethereum that all options should eventually run on as well um and so we're really kind of focused on providing kind of a very credible option that can can compete with kind of these uh these more centralized these

less credible systems so um to your question about specialized versus generalized auctions um I think this is going to be a very interesting thing to observe obviously kind of we're batting that there is at least a strong need for a generalized option to exist um that doesn't mean that there were not also be specialized options I think so um because the user and kind of the application come before the block Builder and so on in the uh Mev supply chain they have of course the the Monopoly over their own transactions and they can send their order flow wherever they want and I think we see at least a lot of experimentation um and just economically speaking

um it's hard to deny that there is some benefit to specialization but as as quintus was alluding to so you do have kind of a lot of benefit also from having complementing transactions inside the same option and I think there are more transactions and people would generally believe that are in some way complementary especially um especially when we stop kind of hard coding the execution Paths of transactions so once we move to kind of a more intent based transaction format all of a sudden kind of the the like the space of possible coincidences of ones starts to go up by like two three four like orders of magnitude in my opinion um just because like when you just have

intense they're so much easier to match than if you have kind of hard-coded execution pass but I thought that is other argument was even more important which is if you have one option and and ultimately it's all about kind of allocating like this kind of uh you know preferential block space um then you can't just create ultimately like a more valuable block and you can make um you can kind of fill users preferences in a in a better way right and so I think any system any Oxygen system that doesn't happen at the Block building like at what what Grinders was calling block building time I think it will have a natural a very

natural disadvantage of a system um that does and and so I think those are some of the main reasons why we are betting on um on kind of an auction that happens at block bidding time and that that really like is an argument in favor of having one generalized talk so having all of the information in one place and then kind of auctioning it off or like or merging it into a block all at once so I think that's was one of the ideas behind also like the like unifying Workshop right this which which is suave and so um yeah so I think that's that's kind of how we think about that um credibility also of course would be very important so I think it's it's much easier for someone to build

um like a new a new blockchain that kind of only serves the purpose of um the like running an order flow option for you know that's just for like another chain and make that decentralized then it is kind of for every exchange to develop their own uh order flow options um I think the the letter is much harder to decentralize and um I think we'll see that in some of the designs um that are coming out and in kind of the next uh couple months by the way I realized we haven't really defined a coincidence of wants but that's sort of uh like a classic example of this on in the FX world is sort of like a ring trade where there are maybe you know sort of you can imagine three different participants with different currencies

that are maybe it's like you know the Yen the remnb and the dollar and they're all equal kind of once you adjust for conversion rates like the same amount that you're trying to change so you can imagine just like a circular sort of trade right um so that's just that the definition there and I do think the you know we talked about this a little bit with penumbra um but the batch option this is super interesting uh you know maybe we didn't get into itself maybe it's not worth it to delve in here but like that's it super interesting uh new format as well to explore talking about I mean that kelsov is doing like a bunch of these things as well right so they they have the best options but they also would they kind of optimize for these Coincidence of ones in the sense that

they only uh so in unicep the difference is like every option trades against the market maker even if there's odd bits coming in in the same lot that actually trade in opposite directions someone wanting to buy um Eve and someone wanting to sell Eve against usdc right so in theory they could be matched against each other it would be strictly better for the users but that's not possible in unison they all trade against the same Equity pool whereas cloud service this layer on top where um that the two parties are matched against each other and only kind of um whatever can't be said to through these coincidences of one system routed to um in exchange like like Unison right and so I think um bed structures are

very interesting um because they eliminate some not oil but like most forms of front running and also kind of um they um there's this very interesting paper I think we already shot it a lot like once or twice from Eric Buddhist um his argument is that bad Stockton's actually kind of reduce my Mev that's lost to to latency raises between um between trading firms and so um that would be another argument like of course kind of in this in this like exchange between users and market makers the less the less Mev is leaked kind of to third parties the better and so yeah any any protocol that we think is kind

of advantageous here of course um deserves a very good look I wanted to I want you to transition a little bit and just sort of if users or if uh listeners are sort of sick of me hearing hearing me say this then I I apologize but it's just been super fascinating I think to do this more sort of ethereum perspective season right after Cosmos because they're just such different perspectives where when you listen to Barnaby and obviously Barnaby doesn't doesn't speak for all of ethereum but I think he's relatively representative of the philosophical Ben of the platform which is very much focused on credible neutrality and preserving optionality and not wanting to enshrine a decision whereas in Cosmos maybe that the reason

they have a little bit less uh there's a little less value to secure so they kind of have this nice um is this trade-off where they can actually be much more opinionated about the design of Meb and user experience that's just it's super fascinating to to kind of Stack those two different perspectives one against the other but it has just been a very cool instance to see where these two uh sort of design spaces and opinions are converging like that exact phrase that you just said haasu about not wanting to hard code users intents you can go to this this episode obviously we we talked to Barry uh unfortunately you weren't able to be there but I talked to Barry this season but there's a great episode on I pledge

allegiance which is a great podcast between him and Dave OHA which is one of the founders of Osmosis where he talks about exactly that idea you know where you should be a little bit less prescriptive about hard coding intents in um and I I would be curious just to get like your particular I think you've probably spat between Baker and lydo and flashbots a little bit more time on the ethereum side of things but I would love to kind of get your perspective on sort of that that philosophical division between Cosmos and ethereum yeah you know I I'm really bummed that I I had to miss the cosmos episode because you're totally right I I definitely have a band towards ethereum um but I'm very interested in in what they're doing in Cosmos because they can

move much faster and they can try stuff that we frankly cannot try in ethereum and um but they can nonetheless like yield very interesting results and very interesting data points that that we can ingest at and the reason that they can move so much faster of course what like on the one end is that they are secure kind of much less money um but their chainsawers are kind of much more like specific to what's individual applications so in ethereum everything that we change about the protocol layer affects millions of different users right many of which have kind of opposing interests and the goal like the very kind of specific way for ethereum is to be this credibly neutral

based layer on top of which like many different things can be billed and they can all coexist and if you deploy something then you have a very strong guarantee that your your application um is not going to be broken by the next update and the ethereum foundation is trying really hard to honor that promise right and kind of to to Ensign that that idea of credibility and um Cosmos doesn't have that problem right like in Cosmos um they are like so look looking at osmosis for example is a full disclosure I'm investor um at osmosis but so they have like permission smart contracts for example right so if

you want to deploy is what content on osmosis like you have to get it permission by governance for example right if um they they have a rule that that validate Us in osmosis are not allowed to sandwich like a user for example right even though they're totally could right it's just like this understanding that so it's a validator does that then it's going to be a governance vote and then we get socially slashed and like all of this stuff would completely like be completely impossible to enforce in ethereum right there's no governance layer first of all in ethereum where people can vote with their coins at a particular validator should be slashed um so even the idea in ethereum was like so outlandish that it's like not even

possible to entertain and so because you're so much smaller and so much more nimbled and you can do all of these things that are completely possible in ethereum and vice versa and so Islam just has to take a very different approach it has to be much more slow moving and a lot more stuff has to be bit in ways that do not break the existing guarantees of the system and that do not require kind of any kind of Rapid or breaking changes to the protocol and um yeah I think I think this is um this is the main reason why um um again Cosmos they can just move so much quicker and they can change the protocol in ways that um in ethereum we can't and why we have to look more into building

stuff that's compatible with the status quo and kind of layer stuff on top of it something that I hear out of the osmosis community and I I do think uh it would have been great to get especially Dave is very outspoken about his his views on Mev and it's been sort of antithetical to a lot of stuff that we've talked about this season but um it's kind of this idea of what can your validators do for you as opposed to it's like the opposite design principle in ethereum and then the you know having the explicit aim to commoditize validators I think is what part of what makes ethereum's block space so valuable but it is kind of an interesting thought experiment that's playing out in Cosmos about like if you have more sophisticated validators that understand

the application code as well as the consensus layer it's just pretty interesting um overall and maybe maybe like the the Segway that we can sort of close on here is is privacy um because I know you I know you've done some you you've spoken to Henry even though you went on the episode and sort of looked into penumbra a little bit it was it's fascinating what there is it's just interesting to hear his story you know not starting in Mev at all right he was starting from the perspective of privacy and and protocol design and then he sort of wound his way up at penumbra and I would love to just sort of we didn't get to talk as much about privacy in this episode as I would have liked but I know it's been a big

Focus for you so I would love to get your sort of closing thoughts on kind of kind of privacy and we didn't talk very much that much about like threshold encryption or anything like that but yeah I would love to maybe close with your overarching thoughts on privacy here I also thought when you said that on the interview so I was listening back to it the other day I thought it was really interesting that he said that they hadn't actually started with any kind of application in mind that just been working on privacy and I think this is a very common experience for a lot of people in crypto actually so if you look at bonero if you look at Z cash like this original kind of this privacy or G is the argument for privacy was always more like ideological right it was like

oh yeah people deserve privacy um Monero may be a bit less so I think there was like always more where we need privacy in order to kind of transact on Dark Net markets and so on that uh Monero was always a bit like slightly more maybe with kind of a bit of like the government as like an adversary lecture in mind because Z cash was kind of a bit more um with the idea that it's private but can comply if you want to make that like really easy for you and um I think what we're seeing now now that we have D Phi and that we have a much more kind of merging Gene like thread fire and D5 and every year that goes on and kind of more and more experts from

mechanism design and so on coming into crypto um I think we now all of a sudden see like another branch of people who say well we we don't need to be we need privacy but not for like any other logic kind of reasons we needed because without privacy you cannot build like efficient markets like we actually need privacy for efficient markets like the internet didn't take off before there was like encryption um there was like no like you don't have like online communication without encryption you don't have like people playing online with their credit cards but you also kind of have like efficient auctions efficient marketplaces like if someone could see that I want to buy like a particular item on eBay or on Amazon they confront on me like none of

this stuff will work right so uh now I think we have this much more real like tangible need for privacy um that we didn't have in the past when kind of the lack of privacy was not exploited in the same way and also there wasn't as much economic activity and so I find that super interesting um this whole extreme of of thought and I think we will see I mean just like privacy for me is going to be the big Topic in the Meb supply chain for the next couple years um just because we need privacy in order to have to kind of break this current trend of vertical integration you need privacy to enable division of labor between different parties and you need it to build more efficient markets and

so I think um I think kind of the next 10x maybe 100x Improvement that we will see in ethereum will come from encrypting the AV supply chain um kind of from the moment that a user extends to intent to the time when it's being confirmed on chain I think when we can build that I think just ethereum as a base layer would just be like so so so much more useful um as a as a platform kind of to deploy your mechanisms to build new markets on top of but also to use any other systems that already exists and so I would say yes I'm super bullish on the concept of privacy not for the ideological reasons but just for kind of the very tangible

reasons that I laid out so that I'm I'm All I would also put myself in the camp of uh I would love ideologically to see more privacy in general I will say privacy seems to be a one of the rare sort of bipartisan issues in in Washington at least like people don't seem to really value their privacy that much and I do um you know talking to some people in in crypto who've like worked on privacy Solutions I wonder if that's kind of this thing where we might need to wait for an Overton window for people to to care about it um but I I would just sort of put myself in the camp of hoping just yeah if I'm an ideological perspective that privacy becomes something that we build into our systems and people start to value a little bit more but I totally hear you

on the more pragmatic aspect of yeah for options to work there needs to be sure privacy it was this way with um I forgot what it's called the like main encryption protocol online um so I read I read about that like some time ago but uh there was a big debate about whether encryptor should be allowed online right and I think it was like not any of the philosophers or anyone making kind of these ideological arguments that ultimately suede kind of the United States States government on that being okay it was um it was just like the very pragmatic kind of industry coming together and saying like it's not gonna like it's not gonna work you need encryption to do Commerce online and so I think there will have to be a similar push

eventually in crypto like we need like it needs to be okay for things to be encrypted otherwise like Commerce on the blockchain is not going to work and so um I would if I had to guess I think that that time would come uh when I'm not sure but in the meantime we'll just keep building and these sort of closing dots fasu I mean I know we've got our sort of individual wrap up uh just the two of us but like close I know anything you want to bookend or leave listeners with or anything like that I thought that the one thing that we didn't touch on um here now in this like post discussion I I thought it was very interesting um just the position that barnabia takes is part of from the theater foundation and like the kind of how he shows like

us how they think internally which is that um they are they are kind of very focused on like not letting things get too far out of hand right so they look into different things they look into like PBS or data centralization they're also going to start looking Osa um and then they basically think about things that could be deployed but not with the intention of deploying it more so as a kind of nuclear right kind of as a way of of like um kind of disciplining the market and I think you see a lot of parallels here with just like Regulators in in other markets right um and so far I think they've taken like a very light touch approach which I

think is exactly the right um way of thinking about like a system like ethereum um and so yeah I I think that's really interesting this kind of uh these paradets between like the ethereum foundation and like a regulator I think it's you know in this season I think we've wanted to stay relatively in the weeds and a little bit technical but it it's also been yeah I found myself sort of wondering what's the right mental model or framework to you know to assess protocol design and I think it's much less looks like a business and much more it looks like a a government and and Regulators I think John Adler there's a great talk John Adler giving and one of the defcons where he explicitly sort of went into more detail about the analogy of uh

regulation on ethereum so um all right great episode also I think we can leave it there thank you so much Mike [Music]