It’s My MEV and I Want It Now - Tom Schmidt

foreign [Applause] [Music] hi everyone uh my name is Tom I'm a general partner at dragonfly we're a global crypto Venture fund um we've done a number of Investments and and sort of been writing about Mev for a while where early investors in flashbots and um you know I sort of want to walk through I I guess maybe more of a VC's perspective on what's happening in the Mev space I think um there's been a lot of discussion a lot of research put into um how do we design um fair you know order flow auctions um how do we incentivize people to um you know uh extract anything in a way that is like not malicious um and I think the the mechanism design research has been going on has been great I think that the problem of course is is how do you make sure that the market recognizes maybe and the outcome that we want the the mechanism that we want um and versus uh sort of you know having the market and sort of what the market decides be sort of some be sort of sub-optimal and so um when we started talking about is um sort of the different actors that are circling on different parts of that maybe stack and how we might go about sort of preventing any single actor from uh maybe seizing a part that we don't really want them to so um I think so far everything we all the sort of discussion around me deal and researcher around MVD has been sort of related to sort of the world that we live in right now where um you know users sign a transaction it hits the public mempool um Searchers uh uh uh you know CDs transactions they assemble bundles send them to Builder and then it eventually gets validated um and and I think unfortunately like that sort of state of the world is is slowly sort of changing over time um this is really good uh tweet from block native um where they basically try to drill down and figure out um well you know what percentage of total uh uh transactions that we see actually aren't going through a public mempool or going through some other method um that sort of breaks the our mental model of the world um such as a private mempool that would we sort of necessitate a change in how Med and Meb supply chain is is being structured um and so we found a 3.8 percent this is older I think it's actually much higher and I'll be talking that a little bit later um and so you the the problem is like we can design really great options um but if we don't have incentives to and encourage people to participate in the options they kind of end up going to waste um and so this sort of gets back to this topic of like um uh I guess block centralization and and um uh Builder centralization and um SC John has this really good uh flow chart um talking about you know so that inherent centralization and sort of positive flywheel um of uh um private order flow and you know I would argue this is already kind of happening um already it'll get into a little bit um and I think so far again kind of the discussion in the Meb Community has so far has been uh sort of finger wagging oh this is bad oh we shouldn't you know um have private order flow we shouldn't have this sort of centralization of Builders um and instead I'd like to sort of you know think about it more and sense of well how do we encourage an incentivize that people who maybe are privatizing workflow today to um make it public um and and sort of get back to this sort of censorship resistant um decentralized open world that we want to sort of live in um and so uh this is you know very somewhat famously classically um safon's uh uh Mev supply chain um going from user to wallet to search or to build a validator um but I would argue this again is like almost too abstract and this part right here is actually kind of where all the activities is happening right now in the Mev space um and one thing you might notice from the from the supply chain is that basically the further back you go in the supply chain um the more leverage that party has so um you know the the um while it actually has a lot more leverage over the Searchers and the user actually has a lot more leverage over the wallet um and so it's this isn't quite as simple as sort of how we think about um transactions historically um and so to get into this we'll sort of break down you know uh classically how does having the transaction get formed um so you have some sort of UI um they send a signature request to the wallet to the webview provider um the wall then asks the user to sign um you get one of these little dudes um just click OK um the user sends the transaction um the wallet then sends this signed transaction to an RPC endpoint um and then um that enters sort of this public mempool it gets right across all these different nodes eventually you know either through the traditional means of just you know validator finding it including it in its block or on a search or sending it to a builder and then and then getting included in the Box it somehow gets validated and then the DAP also updates to UI um but you know all these sort of different components of this of of the of this life cycle they all want to increasingly maybe for themselves and this flow sort of I sort of sort of started to break down um and so uh you know you have some memes that kind of you describe you know what's going on a little bit it's got to get these out of the way unfortunately hmm and so um you know there's a couple different camps here and you can see why every different cat every different stack um will sort of conceptualize why they actually think they deserve um the the order flow and why they're actually trying to go after it so you know I think right now a lot of people think of order flow they think of RPC providers and there's been a lot of effort a lot of development being put into having a sort of custom RPC provider um uh which we'll get into some examples in a little bit um I think in practice this is like a really deaf really tough entry point for um new companies and new projects um a you know a lot of wallets won't even let you choose your default RPC provider um this was not the case for Manifesto recently um it's it's super clunky users don't really understand what's going on I think there's also even questions around how it's actually modernized and how this actually gets uh paid but you know in concept one sort of way of entering the supply chain is basically um well when the wall actually sends the signed transaction to RPC I'm not gonna it's gonna send to sort of the conventional uh you know uh eth you know um uh mempool I'm actually gonna you know sort of replicate what that interface looks like and then do something kind of funky with the transaction um so one angle here is um you sort of convince people to use your RPC because you sell them on some sort of security angle so heartbeat um which is a dragonfly portfolio company um is going down this route so um you can send them all of your transactions um you can send your RPC to them they intercept and scan every single transaction and depending on the level of security that you choose with them they can either ask you for a two pack for every single transaction so that you know either they'll text you and have you approve the transaction or if it's something that their simulation picks up is potentially you know scammy or malicious um then only then will they actually you know um you know ask you for two fact and so it's sort of this nice little add-on or module into how people use RPC today which is again just sort of uh you you blast it into the mempool and sort of hopefully the best um and I suspect this is kind of the path that like some of these other simulation tools are also kind of going to go down out of basically hey I'm sort of like Norton Antivirus send me all your transactions before they hit the public then pool I'll make sure they look good and then we'll sort of send them on um another way to sort of think about this is um rpcs that go after this sort of ease of use so it's all these usability ux wins that um in aggregate just make it you know uh strictly Superior to just sending it to inferior or something like that so um example would be gas Hawk um they uh uh that basically um try to model out gas price fluctuation and then send your transaction when gas prices are lower and and sort of save you money and sort of pocket some of the spread um it's even something like flash path protect or the one or one inch Rabbit Hole um you know they wait to make sure that you're not paying for sales transactions you can pay in you know arbitrary fees Etc both this and the previous category in theory should be yeah strictly Superior everybody should use them just using a standard yeah you know in Europe for example or PC endpoint um but for a lot of the reasons that I mentioned earlier it's just really really hard to get people to switch over um if they can't even switch over at all um and so this sort of goes back to this other option which is well instead of you know having users switch their RPC over what if we just um sell directly to wallets right because wallets again they're further up the supply chain they have more leverage um and then we'll integrate with them directly so um you know in concept this is great for wallets right It's A New monetization Path because um you can sort of uh benefit from some of the med that your users are generating you can Outsource the med extraction to somebody else so um you know maybe it's a bit simpler for you logistically or from engineering perspective or maybe you know users actually even get uh some of the Meb directly as well and sort of create stickiness and so we've seen some development down that path as well um but ultimately this all is sort of again downwind of of wallets wallets really have a lot of Leverage over which LPC endpoint an order is going to get sent to where transaction is going to get sent to you um while it's obviously they need to monetize um you know not everyone can have so the mirror mask swap style money printer and so they're all looking for ways to sort of um think about alternate sources of monetization um we'll also have a really great um order flow not all order flow is created equal right like order flow from a bot entity or order flow that is toxic order flow from a professional Trader that might have some sort of alpha um you know that's not something where there might be that much Med that you can extract um but if you're uh you know catering to people who are less price sensitive less time sensitive um that is actually a order flow that can generate a lot of Mev and therefore is is you know power per pound more valuable than other types of transactions that you see um so again you know this is a sample for metamask where um very designed to pursue this method of um sort of swallowing up um the RPC providers instead of having their own layer on top of them where they can you know again provide a lot of these ux Security benefits as well but also potentially extracting maybe so before stuff again hits the public mempool um they can sort of do whatever they want with it um yeah they can do it themselves um so it was like a tweet interaction between Bert and uh Mike from from rainbow from a while ago discussing the idea of hey what if rainbow actually you know just extracted uh any of you from user spots itself um or you know um uh uh partner with somebody else um or we're going to be auction it off which is I think how a lot of people conceptualize it the question is like you know what reason do they have to do that um and how do you sort of make those economics so actually work for wallets um you have to actually um cater them and sell them a product that they want because they have so much leverage in the ecosystem um I think the Silver Lining here is um I think uh while they're actually finding that just charging on these sort of swap products is actually a much better monetization monetization strategy than Med extraction um and so it might just be that you know they might sort of leave some some you know Juicy Fruit sitting on on the table instead of trying to you know extract it themselves because there's so many other uh ways for them to actually make money uh I mean again all this sort of discussion so far has been in sort of presupposing the current sort of uh stack right like uh the RPC providers sort of assume that users will be able to switch to RPC endpoint um and wallets are sort of assuming that well I will actually be able to receive transactions from users um and that of course is also changing over time as we're seeing um that like sort of the conventional way people sign and create orders um is actually changing so um this whole sort of bit is is kind of what I want to drill into a little bit um so dapps really you know kick out this whole process of the user click something or tap something on a UI um and then the DAP says hey you know wallet please sign this this transaction um but the transaction itself is going to be a conventional ethereum transaction it can be an intent-based transaction right can be like a open c order or like a zero X order um or this is the case during Sushi where you're basically signing a hash that represents your intention and then that doesn't even enter the wallet right there's no transaction to um relay that just goes back to the DAP and then the DAP can do whatever they want with it so in this scenario um Sushi as a partnership with manifold where um they have this product called Sushi guard where basically uh when you swap on Sushi you sign this intent it goes to Sushi guard they can extract an EV from it um and then that can go back to Sushi so the wallet doesn't actually see the transaction in this flow and therefore the RPC providers don't even see this transaction um I think this is kind of not optimal um because it's really confusing for users you don't really know what you're signing you're signing this text which can obviously as we've seen as sort of the permit style attacks and other types of attacks um after being malicious in some scenarios so people have been sort of conditioned to be on guard when they're signing a transaction um less so when they're sending a message but I'm message can you know given certain conditions also be be kind of malicious um and of course you know this also sort of depends on the nap um so um a lot a lot of uh uh you know daps I mean we're talking about the the protocol layer as well as maybe the UI layer um it really depends on what percentage of your volume or percentage of your every activity is going through your first party interface um or versus third-party your cases so if the majority of your volume is going through an aggregator or a third party or maybe just Bots um You probably don't actually have that much leverage in terms of how do transactions actually get generated whereas if you have you know a great um site or maybe you are slowly vertically integrating and creating your own wallet or creating your own app um you actually have a lot more leverage in terms of uh you know again choosing um uh where orders are being routed and sort of where order flow is being generated so uh this is like a screenshot from um I think you thought who's one of the founders of friction basically looking at the percentage of toxic order flow on on uniswap which you can think of as sort of a rough proxy for the percent of volume that's going through uh or not going through the user why and instead being generated uh programmatically or going to through some sort of third party um so I I think you know zooming out a little bit um all this is to say that um uh the the sort of the the land um underneath uh Ned world is Shifting pretty rapidly um and so you know building a product that assumes that people will be able to find a transaction in a public mempool um and then you know engage in some sort of uh uh searching process for it I think is is not um probably gonna be long for this world so the question is you know how do we make it uh EV positive for these different players to um cooperate and participate in a open auction um that is fair to users as opposed to relying on their uh benevolence to not generate malicious I mean the not not you know not self-extract um you don't really want a world where you know the majority of uh transactions are going through better mass metamaster self-estraction and then they also maybe you know uh become become a builder it becomes all vertically integrated which is um kind of the trend that we're seeing um I think there's also a question of like you know where will that Mev actually end up obviously everybody's trying to go after it um but there's also a question of you know why why doesn't this end up with with users there are efforts to push you the med that is being generated back to users versus going back to an RBC endpoint going back to wallet or going back to adapt um but I think that ultimately relies on competitive pressures across these different players to to do so no one would really do it but it is sort of like credit card points where everybody everybody else is getting some sort of Kickback you're sort of forced to do so as well even though um it's sort of hurts your bottom line um and then uh you know I'm also I have a research piece coming out soon um that sort of has more data on um trying to map out and and quantify the public member or the private member problem um but if you want to collaborate uh feel free to DM or email me but I think in my mind we're sort of in a state similar where we were maybe two years ago before any of the Explorer where we know there's a problem and we have some sort of little Snippets yeah three different papers or through one-off research pieces but we don't really continuously monitor this or track this and you know based on um some of the anecdotes as well some initial data um it seems like it's a growing problem if we want to live in a world where um you know there are all transactions that are just going through a private mempool effectively um so ultimately I think when we sort of look ahead the big takeaway here is the people further up the supply chain um have more power um and so um naturally everybody's trying to you know move further up um but It's Tricky given given some interest or sort of already entrenched and so um you know everybody's acting really rationally uh nobody's trying to be malicious here um and so we can't just really finger wag and say hey you should not be privatizing order flow you should not be self-extracting instead we should make it economically uh reasonable and rational for them to give up their order flow and participate in an open auction um and I think um that's where I'd like to see more more sort of discussion happening versus I'm just sort of assuming that it will happen um I think things like Mev share are a great step um and obviously making more accessible for people to participate in an open auction but I think ultimately question we need to make sure we're answering is um whether or not it makes sense for um these different actors to open up uh on order flow um to to an open auction versus again sort of keeping it private so um that was it kind of quick um happy to answer any questions um if people um have them actually I don't I really don't even have uh I do close the tabs I don't get Echo but happy to answer questions we are running a bit on or to the time of time that's right we're still behind on time uh so we are just going to move forward this uh but thank you uh the viewers who are confused by drewins equations appreciated the break on this one uh and we will turn it over to quintess now who's going to talk about cat I think