Interview with a Searcher 2.0 | Dean Eigenman, Anish Agnihotri

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 huge fan of this team and what they're building so stay tuned to find out more you can be a competent Searcher and generate your own and be competitive as a Searcher then there's a high likelihood that you should also be a builder right simply because who are the best Builders they're people that have their own order flow all right hasu so today we're going to be talking to Dean Eigenmann and Anish agnihotri a couple of Searchers and I think this will be a very fun callback to the original I think the sort of canonical there are very few interviews with Searchers out there but your interview with a Searcher

from 2020 really stands out so um yeah why don't you tell listeners like what are we kind of trying to probe for here in this conversation yeah I I mean I I look very much forward to talking to these two guys I I um to our Niche I need I would consider a friend of mine he's been on the podcast before he used to work for Paradigm at the same time when I was there and um I think he he's so good at breaking things down and explaining them in a concise way and um yeah it's really insane how young he is I think that's something that I mean you you'll be hard-pressed to find a person who has kind of spent more of their time in crypto as a relative percentage of the age which is uh I think it's very interesting

um Dean is he's a super colorful character I think he's very polarizing um he always has great stories um he's working for this so he has his own Mev company project blank I think they are mostly active on uh on this alternative uh chains so not predominantly on ethereum I think they do a lot on um kind of these like alternative evm chains um so I think we get a very broad picture here yeah all right let's get right into it all right everyone welcome back to another episode of bell curve today at hasu and I are joined by Anish agnihotri and Dean Eichmann guys welcome to the show thanks for having us for this episode we're going to be kind of

diving deep into both of your roles as Searchers in the Meb ecosystem but I think before we get into that I think Hassan I would love to know kind of just to kick things off you know can you guys just give our audience a little bit of a background about each of you and how you kind of got into the space I think that'd be a very interesting segue into our conversation sure I can I mean I can start um I've been in this space for a while um starting like 2016 2017 um I met a guy on Reddit called Matthew deforente we were talking about smart contracts at the time and he was running an audit shop called CK labs and was part-time on the ethereum security team and so I started working together with

him and then around the same time Nick Johnson was looking for help on ens because he was the sole developer there uh started helping Nick Johnson and them was doing those things full time for a while uh at one point we shut down TK Labs because auditing is ruling and not really at one point worth it anymore um then I left that kind of took a little bit of a break and then joined status uh working on protocol research for messaging um worked on some of the peer-to-peer message protocols there that are still being used and iterated on like the iterations of uh whisper that the EF had originally proposed um and around that time as well I

co-founded dialectic together with Ryan who was previously a co-founder at polychain um yeah and been doing dialectic ever since that's been like almost three years ago now I think uh kind of hard to tell and uh part of the things we were doing at dialectic we wanted to look into Mev and started a fully fledged team from there to kind of uh yeah do that maybe Dean to to stay with you for a second how would you say um what what we say was the relationship between what you did previously which is kind of security auditing and protocol research um how did that allow you to transition into searching and um what would you say was your first Big Break as a Searcher that made you realize this is something

that I can do for a living that's something that I find really interesting and so on I mean for from when it comes from the protocol research and things like that uh the main thing is for the research I was doing it required a fundamental understanding of the full stack which you kind of need for Mev as well uh which is one of the reasons why I like spending so much time on Mev um and I find it the most compelling thing to do in crypto right now is because it incorporates everything right you need to understand the networking you have to understand the clients how they function um you have to understand how the evm works in detail and things like that and so for example for the evm part uh my experience from auditing

um obviously helps there because we had probably a a better understanding of that Paradigm than um you know just a normal smart contract developer would have I got into crypto I'd say probably 2014 so coming up on 10 years next year which would be exciting as an Enthusiast similar story to Dean was on Reddit making websites for people got paid in Bitcoin because that was the only way as a kid and so stuck around since then say say hello adjour to make people realize how crazy and so long you've been in the space yeah I'm 20 now so I got in when I was 11.

um a couple years ago now um and since I've stuck around it's the most exciting thing to me it's the closest Hands-On I can get with anything in crypto and so that's why I stuck around um in 2016 2017 I unfortunately was a kid so I didn't buy into the ethereum pre-sale but uh I got my parents to buy me a bunch of R9 280X graphics cards and was a minor early on stuck around the ecosystem did a bunch of Open Source work with Bitcoin eventually found myself in senior year of high school joining polychain um another place where Dean's co-founders worked and co-founded I think um there was a part of a prop team for

about a year in early D5 summer where we armed a yield from borrowing across centralized exchanges and put Capital into the new and you know protocols that had only been around for two or three days at that point in time left that in 2021 and joined Paradigm as a researcher where I worked mostly on open source protocols and building tooling and about 10 months ago left that to go independent as a Mev search or solo um that's fascinating uh both you guys I think maybe to start at a super high level I think it'd be great to give the audience sort of a perspective of what is the landscape look like for our Searchers today when I say that I mean I think at least for me going into this

season when I thought of Mev Searchers I thought of kind of um you know a bunch of individuals kind of writing Bots that were looking for Mev and finding a way to profit is that a realistic understanding of the landscape or has it kind of professionalized in our other is it kind of dominated by larger shops that are searching for Mev like what is that whole kind of landscape of Searchers look like well I mean that's kind of a hard one it's hard to be able to actually tell because you don't know who's saying who's speaking the truth and who who's not right the only thing you actually know for certain is what's happening in on chain um from my like feeling and kind of what I can see is I think

if not professionalizing people are starting to group together um I think there's probably a lot less solo Searchers than there were one or two years ago because yeah the competition has just gotten so much more Fierce um but whether or how much of the space looks like that if yeah it's hard to say for certain yeah yeah if I were to Ballpark numbers it'd be like today there's probably an upper limit of 500 Searchers total just based on but of course it's like you know you can't guess from addresses but at least that's what my high level number is and of them yeah say majority are probably in like small teams of like

under 10 people and that's all of crypto or just ethereum I'd probably say all of crypto yeah there's a diet overlap I even think for all of crypto that 500 is a very optimistic number yeah like I'm counting this as anyone that is sent maybe like you know a bundle in the last year or like a couple of yeah last year to a relay practically speaking it might even be like a fraction of that number there might be 500 but then I think the amount making money is like yeah or the amount who can live off of Meb because there's a lot of people who do Mev just for the fun of it who aren't actually making any money I I've seen like a lot of the the various Mev

discords it's just like people messing around like Mike on Ticketmaster yeah that's exactly right I'll tell you guys a little Mev Story how I lost a bunch of money um Dean you mentioned that the searching landscape was getting much more competitive can you describe uh just give the listeners a little bit more of an understanding of what that means and maybe in particular how is searching or the landscape for searching change based on when you first got into it and what it looks like today I mean should we count the stuff we were doing in 2017 with like ether Delta arbs and stuff like that as searching because I mean back then back then it was easy right no one was like doing anything you were in the mempool you didn't really have to you that you

didn't even have the PGA back then right you could just kind of do transactions and that was it um and then you know people started PGA there was more competition and things like that and then at one point flashbots um came and what flashbots did was democratize I guess you know this dark mem or yeah democratize access to the to validators or a democratized collusion of sorts and what that effectively did is it caused a race to zero on the known strategies that other people are doing because now you're competing on a different landscape right and so I think from those strategies the things

that people are doing over flashbacks and stuff like that the competition has become a lot more professional because they need to optimize for like very few things to be able to win um in that paradigm and so you see yeah a smaller concentration of people doing those things um and I guess probably more people are looking for stuff outside of there that they can do without having to pay miners tips and things like that yeah yeah I think like the the Overton window of you know Mev has shifted a lot over the last couple of years where in 2017 2018 like Dean mentioned I remember you know rving on zero X by just

pointing clicking tokens and yeah that was very feasible right because nobody was looking at it right for context you know swap V1 I think came out Devcon 4 of 2018 and we didn't see a sophisticated actors for that pop up for a long time afterwards and so it's sort of just like as we've seen volume rise up people have started to you know more eyes are looking at this and saying hey there's opportunity here and because of that a window has shifted to now this is more mainstream and yeah there's trading firms out there that are now dealing with Mev which you know four years ago I couldn't have told you the uh there was anyone Beyond you know friends and family type of thing you know people I see at conferences who are just arbing this by clicking to put a lot of this into a little bit

of context so uh you probably can't go into a super great detail but um as to as to you too can you describe uh in Broad Strokes like what do you do in Mev um what chains are you on what kinds of opportunities are you focused on and uh are you doing this alone or um as part of the team yeah so I am a solo Searcher so it's just myself and because of that I focus mostly on the long tail so differentiating here between the short tail which is the usual you know Arbitrage liquidations things that large teams are mostly focused on I don't do a lot of that most of my work is on a broad long tail of opportunities which

looks like you know nft Market making nft mints uh probabilistic backgrounding on protocols that don't have Solutions like flashbots spam um keeper Bots a wide range of you know the Obscure things that you need to keep the chain running but aren't necessarily you know as sexy as say Arbitrage or Atomic um you know liquidations and things like that um as for chains I'd say predominantly on ethereum made net um although I basically deploy the code that I have to any evm based execution environment since most of it is largely the same across each and every single

one and that includes some of the new upstarts like you know evm on Cosmos similar to Kanto and things like that where there's opportunity that makes sense how would you how would you say to stick with you for a second before we go back to Dean um as someone was active on on the holiday VM chains at least in theory how has uh how is so far the the the kind of shift to roll ups in terms of like Roll-Ups have not surpassed ethereum in in pure volume numbers but how would you say this has has impacted the nav that's available on these chains and and maybe um related to that also how it is working on these chains different for for you given that they don't have

um they don't have kind of me V origins or if um they have these sequences that that kind of what are transactions in the first come first serve manner sure uh two questions there I think the first one probably with as transaction flow moves to l2s how that's shaped up for the Mev World um I think for me it's not much of a difference excuse me it's mostly just a different execution environment um with most of the same Paradigm I think it's a difference in the node software and the networking uh wherever volume goes generally Mev Searchers will track that very well and go there as well and so we already see there's yeah dominant teams that are across arbitrim right and just the way

that they do it is slightly different so on arbitrum and on optimism these l2s you currently have centralized sequencers and so it's much more of a networking game of how close can we be to the sequencers how close can we be to the order flow Originators where these transactions are originating from how quickly can we you know take a look at what state transitions are happening and then send and fire off a transaction versus on ethereum where you know you have Solutions like flashbots and so it's much more competing over the tip to the validator for at least most short till opportunities um so that I'd say is probably what the initial difference is as for the differences that I notice moving across l2s and other execution environments

I'd say for at least a long tail Searcher like myself it's much more beneficial to build up a domain expertise in networks that aren't ethereum like if I can understand the quirks of a smaller L1 if I can go deep into the protocol code that sometimes is a fork of Geth but has its own modifications that change up the ways that you know networking works or changes the way the mempool works it makes it much more competitive for me to compete on that Chain versus say you know any random Searcher from ethereum that can come and deploy their code on any evm environment and so for me it pays to basically double down on these more obscure L ones that trading firms and other large teams don't focus on to get that Competitive Edge that's a great

answer thank you taking the same question to you Dean um what do you do broadly and yeah I'll ask it a what if so we kind of do the opposite of the niche right now we don't do much long tail um I think long tail is less viable for teams more viable for for independent Searchers um that answers the questions well uh we run it or we are a team um and yeah the thing we do as well as we deploy right now pretty much on any evm compatible chain and are also ensuring to get um expertise to Corner markets outside of evm compatible chains

um especially with things like Roll-Ups coming right now um you know the cosmos ecosystem gaming gaining tractions and things like that I think that looking around in that space is going to be very interesting can you guys um just in terms of when you're sort of looking out there in terms of the world and trying to assess different opportunities can you describe what sort of makes for a fertile environment in terms of extracting Mev so anishi mentioned that where kind of transactions and volumes go right um so like layer Twos for ethereum for instance what what about the applications that are built on top of those l2s um like for instance there's been you know uniswap V3 for instance there's

been that's very fertile ground for extracting Mev and kind of the whole amm model so outside of just where volumes are like what about the apps that are existing in these environments like how does that also play into where opportunity exists for you as a Searcher as a long tail Searcher at least I'm app agnostic and my work basically pushes me to whichever apps have the highest volumes with the least competition so I'd say in the short tell maybe you know you're looking at how can I optimize these amm hot paths how can I get really good at this type of liquidation and that carries well over to other different types of protocols for me that's more of a problem right I want to go places where people have not

fared because that gives me the competitive Advantage right so a lot of my work is an absolutely new apps and new protocols that are doing experimental things that aren't very well defined by like an amm structure since that lets me get very deep into those code bases understand how they work and find out where those inefficiencies are before other people have figured it out um and so at least in the in the life cycle of an opportunity for me it's much more important to be there early and not even to do well but to just be there early and understand everything than it is to say you know understand the intricacies of uniswell B3 very well and then adapt my strategy across every Dapper where that's deployed and can you

um you know maybe without definitely revealing anything that you're working on right now but maybe like an example of uh you know a past opportunity that you're able to exploit I'd love to just kind of get in the nitty-gritty of what that actually means um so like how do you go about what are kind of the steps for determining sort of the logic of different applications and then maybe if you could just give listeners one sort of concrete example of an exploit or an opportunity maybe that's expired just so we get more of a concrete understanding of what that looks like in practice sure yeah I'll give one that's you know been spoken about decently well and has you know repos of people that have worked on it um a lot of my work I'd say greater than 50 is in the data ingestion analysis of

what happens on chain right so it's consuming a bunch of ethereum data other chain data and then writing rules on top of that to identify what a profitable transaction looks like so say a very naive rule could be analyze every single transaction over the last year where a user has inputted less than what they received in output right and a good example of what that catches is say something like keepers for a protocol where a user is spending gas fees to get some protocol token in exchange for running some services for a smart contract or a protocol or an app a concrete example being say being stopped protocol where they have a keeper that you can run it's a Simple Start contract

function you call once an hour that progresses some activity in their smart contract and in exchange of that and they emit use some quantity of bean token which is their dollar denominated stable coin um and so essentially it says something like you know every hour you can get I think now 25 but previously it used to be a hundred dollars for just calling this function and that's an easy opportunity to spot on chain since you're looking for transactions where users spend very little in gas fees and ethereum and get a decent amount of dollar denominated stable coin on the other end and so that's something that pops up across one of The Trackers that I'll write it's even easier because that's something that's on a fixed schedule right it happens once every hour which

means that if I do some you have naive clustering I can easily see okay there's somebody calling something every hour that makes them money and so I will go in and I'll investigate okay what does this look like what is this opportunity how much money is there to be made is it worth it for me to go and write the code um naively you would think oh 100 bucks an hour for free okay this is fantastic right that's you do the math out twenty four hundred dollars a day 30 days yeah that's almost a million dollars a year practically that's usually not how it works um for starters if I'm finding an opportunity that somebody else is already doing there's that aspect of competition where it's likely that it's already been bit up to the point where

you're paying upwards of 90 of the validator and so in the case of being if you're trying to do that today you will maybe make a couple of cents every time you call it simply because it's such an easy strategy that 99 of the of the tip of the extracted value goes as a tip to the validators and so there's not much there but like I said earlier it pays to be early right so if you were one of the initial people to come across this and or if you do it in some unique way right like for a period of time um last year I did it where when nav boost first came out there was upwards of 10 to 15 percent of stake that hadn't onboarded to MVP boost right you had 80 adoption of people using the relays

versus 20 not so if you just default it back to that 20 percent of people that weren't using a relay and just submit transactions in the mempool you actually got better price execution on running these keeper Bots which meant that you were much more profitable compared to somebody who was you know a less sophisticated actor paying 99 tips directly to flashbots or some other relay um so that's I think a concrete example of a type of work where it's much more figure out what's going on on chain where's the opportunity and then find some unique way to extract that value we could touch on so many interesting points that I need some I'd be I'd be interested to to kind of hear your take Dean on just the competitive aspect of searching right so Mike was

asking well what does what makes it for a photogram for searching and I'm sure like you need some opportunity to be there but then the other big thing that you need is like you want low competition right so you can actually as you were saying Quantum Market because like it doesn't matter if opportunities big if the competition is also bigly not gonna catch much but even opportunity small but you can control it entirely then then probably you can make a very nice return so how do you think about the different layers of competition like search about this user Searcher versus search uh Searcher versus Builder if that exists on any of the chains that you're on but also kind of Searcher versus validator how do you think about kind of the these different layouts and how do you try to position yourselves

within them in order to get the best possible outcome yeah I mean uh validator is the absolute end game of competition right we saw that last week um so that's I think that's that's a problem that a lot of energy probably haven't thought about enough given what happened um for those who don't know there was a validator who ended up uh ex or muzing uh Searcher uh last week um the I mean the way we look at it the way we try to analyze competition is not like what is the value of a given strategy it's what is the maximum

extractable value for us um given that we join this competition like are we able or do we think that looking at this competition and doing an analysis of it which includes analyzing the way they send transactions analyzing all of their smart contracts do we think that we're able to easily um you know push them out because that's a strategy as well which we've employed right you don't always have to have the biggest pie in the beginning but you might be able to easily um burn and burn competition out right that's kind of the thing the way I like to go about it because a lot of the strategies that we do are um

pretty common but from what we can tell sometimes you know it's still solo Searchers and things like that and then they don't have 24-hour coverage to um you know optimize their strategies uh and react to everything that's happening on the network whereas we do um and so yeah I like to burn out competition and Corner them as much as possible we don't we don't go into things where we see that the competition is sophisticated as well as employing some of the strategies that we do so as Anish said you know have or a lot of his job is doing data analysis um that's the case for us too we do a lot of data collection and run various

tools to kind of find anomalies on the network uh signed patterns things like that hey guys quick break from 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 I actually would be curious I don't want to go too far down this

rabbit hole but as two Searchers I would love to get both of your perspectives on the exploit that happened last week um because again that kind of two perspectives that also and I laid out on the last episode is you know on the one hand right circus are generally very sophisticated actors and very well compensated for what they do um and these particular Searchers were sandwich Bots which isn't maybe the most uh sympathetic quarter of Meb in the world but on the other hand uh what you know you you don't necessarily want uh Searchers like you guys to be concerned right about some of the infrastructure kind of Upstream of where both of you interact so I would actually love to get both of your perspectives on the exploit that happened

last week All is fair in love and Mev yeah I'm on the same page uh I'd say you know as a sandwicher you explicitly take that risk of using the infrastructure and you should be well aware of the fact that you know there's no certain guarantee that your front slice and backslice will always be in the exact position you want to and this was just a good example or at least the most explicit example we've had of that not being the case and that's all I think nmev generally you have to go in with the mindset that things will break and the yeah the intuition that you have of this the way a system works is not necessarily the

same as both the designers of that system and the users of that system well said guys well said um I you know I actually want to pick a little bit at what you said is kind of the Searcher versus validators being the ultimate level of competition definitely seems like searching is a very sort of PVP uh zero-sum game um and I'd be curious like you know just to clarify something you said it's kind of part of your strategy especially for kind of like short tail Mev just sort of crowd out other would-be competitors by accepting like lower maybe profits than you would on otherwise would on trades to for Less well-capitalized teams to basically crowd them out of Trades and then how do you kind of think of you

know is your sort of mental model that you are kind of in a comp in competition with validators or proposers to extract more of the value of Mev and if so how do you sort of trip away or win at that game over time I mean in the end I definitely am in competition with validators because I don't like paying them um and so I think the obvious end goal there is to run enough validators um where I you know do Miami V from um yeah uh there's definitely been scenarios where we've seen on train that competition is accepting none of the or none of the profit and bribing as much

as possible to like burn out competition um definitely seen that happen um I said All is fair in love and Mev I think people will kind of do anything to get a slice of the pie you know what you say that's uh that's one of the advantages of being a big up firm compared to like a solar charger that you have the balance sheet in order to to take a loss maybe for a couple of weeks to months on a certain opportunity the expectation yeah yeah and also just the adaptiveness right have you experienced any of that yourself on each yeah I mean all the time from the other side oh remix say it it's it's very common thing right

um sort of the game theory of the Grim trigger of you know if I can't make any money then neither will you or anybody else's I'd say very common in Mev across basically all types of met right and I'd say it's probably when I enter a new strategy with a competitor it's probably the first thing that happens right where they're like okay I am willing to play The War of Attrition and to crowd you out of this simply because I'm willing to take zero profit fascinating okay give us some like PVP stories here come on that's that's what the listeners want uh I don't know what I I have a fun one but let's start with you then you've got to

tell us come on just thinking about it I mean the one the one thing we see often now is with Sandwich Box uh sandwich Trappers which are essentially Bots that will purposely um trap sandwiches in yeah very interesting ways uh where the sandwicher thinks they're gonna get a huge amount of profit and then ends up walking away with nothing um yeah so you sandwich the sandwich and things like that we've seen stuff like that which I find relatively fun because that one's like that one's a funny one because people talk about the ethics of sandwiching and I'm like yeah but if you're sandwiching the sandwich or are you a bad guy where Bandits are private all right um

I've not witnessed it on ethereum okay well except this one time now or last week yeah yeah but that that's a different context because you had the validator in play right yes yeah so it can happen on chains that that basically don't have private channels minus yeah I'm on I mean on certain chains you also have private channels but the private channel is only 30 to 40 percent of the wild leader set right Dean just for the listeners could you describe the mechanics of that attack like what does it mean to sandwich the the sandwiches because you know in the this ethereum example I obviously wrote one of the validators was involved there was sort

of a beating right of sandwiches where you gave the impression to to these sandwich spots that there was going to be a really juicy uh you know front-running opportunity and then obviously that that backfired me ended up getting exploited so what are the mechanic is it kind of a similar thing on on other chains where there's a lure so to speak it's a similar thing um it often revolves around yeah you you do a kind of bait and then you do multiple transactions and most of those transactions Bots won't simulate because they're not or they don't look related um and that's where the real juice is in so you bait with a shitty trade um that's as much detail as I'll as I'll say

interesting fair enough um I I would be very serious um you know from it seems like especially for more of this like I guess the heuristic that I'm getting uh from Eugene and Anisha is sort of for a long tail Meb opportunities that's kind of better suited for like single person uh sort of searching outfits and then when you get to the more shortail side of things it looks more like uh almost like a hedge fund kind of like a high frequency trading shop is my my mental model um I would be curious you know because at least on ethereum now you have PBS so you've got Searchers or sorry you've got uh searches and then you've also got Builders right now just be very curious like if you were a sort of a hedge fund or high frequency trading firm and

you're saying hey this mbb stuff actually looks very similar to what I do in Trad five markets and they were trying to decide like do I enter kind of the searching space or the building space like what would be the mental framework for deciding where to play in that value chain or if you're a shop would you just say I'm gonna do the both of these things which might actually be happening right now but I'd be curious to just get an understanding of like where the different competencies lies maybe a niche this is proposed to you because you're you're on ethereum but between Builders and and Searchers all right so don't forget your PVP story you're not up there okay I'm sure I can start with the the PBS I'd say if you're a sophisticated trading firm

if you're a centralized shop that's coming into ethereum it's it pays to do both right if you can be a competent Searcher and generate your o and be competitive as a Searcher then there's a high likelihood that you should also be a builder right simply because who are the best Builders they're people that have their own order flow right it's either you have transaction flow yourself so say something like flashbots protect feeding into the flashbot Builder and or you're a market making or trading firm or an meview shop that generates arbs that generates liquidations that has all these types of transactions that nobody else has that you can then stuff your own blocks with to augment the

profitability of those blocks and make your Builder the winning Builder so if you're a shop in the traditional world you want to do both and that's what we see today right places like Jane Street SCP winter mute all of them run their own internal searching operations and at the same time have builders that they operate um simply because it Pace to do that yeah and maybe just to like give some further intuition on on why that's the case if you search I will send the order flow to a bunch of different Builders then all the the order flow will make the blocks of all of these Builders will available they will compete with each other and bid up the price that they pay to the validator um whereas if you are kind of your own

builder and you don't share your order flow with anyone then um you you now have less competition you can like you end up paying yourself like the search outside of your business place kind of the better side of your business but you end up paying less forward to the melody land so it's almost like you can you can get a rebate on on what you've paid yourself in the first step so um I need um what are some other benefits that you can get by um by being a lot better um that you don't have as a searcher other than the financial one sure so the most direct one is you can do types of strategies that require you to have access to the entire block so with bundles you can only sort of do this

Atomic set of transactions that are all bundled together versus with the full blocks you're at you have the complete Liberty to have you know bottom of the block are top of the block ARB do things in the middle of the block you could sandwich entire bundles that other people are submitting to you and there's existing there's existing builders that do exactly that right where you submit them your bundles as a Searcher they will then run their own algorithms to search around your bundles and then give you a kickback for submitting order flow to them and so it opens up sort of like this entire space of strategy that you can do around building efficient blocks and doing new types of things and I expect that it'll only get more sophisticated as we move

forward where for example one of the things I've been looking into is say if you can know for certainty that you're proposing two slots in a row from a single relay or a single validator set then you can get into some really cool things of having you know cross block MVB where you're building the blocks for both this block and the next one and so yeah there's a there's like an infinite opportunity space so you can do anything do you think that um eventually kind of uh all of the major trading firms would have to be block Builders and vice versa do you think we would see it ever like a reversal of this trend where the the need to be a block data will decrease

I'd say it depends I'd say it's probably in their best financial interest to be block builders although if if you're not particularly dominant um and you know if you're a shop that has very little transaction order flow or very infrequent then it might not pay to have you know dedicated teams that maintain the node software and run your Builders and you could instead you know delegate out to other builders that take very minimal fees yeah like for myself it's it doesn't particularly pay to run my own block Builder right I have some transaction order flow that other people don't and I have some long tail transactions but it's not enough for me to justify the

infrastructure costs of running an entire Builder operator you know operation I have a question for for maybe both of you and you know I've heard um you talk about this a little bit in the past but if you guys describe what sort of the capital requirements are for searching in general and how trades get funded like Dean I I heard in uh the interview that you did on zero knowledge that you've used um flash loans in the past to fund some of your your operations so I just be very curious like is searching again depends on I guess short sale versus long tail the type of thing where you need to come in with a good amount of capital to fund some of these like especially like sandwiches right where you have to move the price of assets and specific pools or is it the kind of thing where you can

use Flash loans or just in time liquidity or just be very curious about what the funding sort of model looks like for searching so from every bot that we track which is like 20 30 Bots across various Networks they're pretty much all using flash loans still um I think the one of the nice things about using flash loans is that you don't have to keep any capital in Native currency meaning you don't have to take on any inventory risk in the end you still want to be smart about that right even if you're operating on a chain and believe in a chain you are operating a business so you want to reduce your inventory risk as much as possible until flashbots

gives you the nice opportunity there um yeah and most of these you know or sorry not flashbots flashlights and most of these venues where you can get flash loans you can get more liquidity than is reasonable to hold anyways on a eoa which you need to run about um yeah so yeah I think most people probably still use flashstones yeah that's it for my Atomic strategies still use Flash loans for the non-atomic ones short duration inventory risk so nothing yeah a couple of blocks maybe we could um talk when you say like non-atomics I think a niche one of the

um at least one of the surprises for me uh going into the season hosu definitely came in with more knowledge than I did but you know my understanding you know going in was that most of the Meb being captured was Atomic and atomic is a very interesting idea when it comes to me because it's like really pure uh riskless sort of Arbitrage but my understanding is that the majority of Mev that's being generated today is actually not necessarily Atomic so could you describe kind of the the change that's undergoing there yeah sure I'd say that's that could be split between the non-atomic strategies that exist in short tail versus the non-atomic that exists in long tail so so in the short tail and at least it's if you can

exclude the atomicity and you can remove the need for Flash loans you can be much more efficient with your execution and that's what we see with things or you know trade on different venues and so a good example of say non-atomic trades or at least shortly non-atomic trades um in short tail are things like C5 defy Arbor stat Arc right where one leg of the execution happens on chain and another leg of the execution happens off chain um and so if you were to look at just simply the on-chain leg you would see that's non-atomic because they take some position on chain right they're either long in acid or they're short in asset and they hedge that out off chain using centralized venues to close that Gap

um this pays for a couple of reasons the first being that you have more venues that you can trade on the second being that you can take some duration risk and you can use your own inventory so say if you're a market maker like wintermute and you already have you know 200 million dollars of ethereum inventory and you have some internal price set it's now much more efficient for you to only have to execute one like of that trade-on chain since you already have inventory off chain um and it also lets you capitalize on things that aren't just you know extractable value from ordering right where from any transaction on chain there's sort of two types of extractable value there's the extractable value you get from ordering that transaction in

some way and the extractable value from the signal of that transaction right um and Stefan and I'll get it Frontier had a really great piece on this which breaks it down but it's a you know just from seeing a transaction what Alpha can I get from that right just from the fact that it's in the mempool if if you make a massive trade a beat right where you say I'm buying 100 million dollars a beep then put that into the men pool I don't even have to as a Searcher be the one that's sandwiching you or front running you simply knowing the fact that there's such a large trade that may execute gives me insight into the fact that the price of each might go up and I can then go make a trade based on that so at least in the short tail that's

where the non-atomic is very dominant in the long tail it comes into more obscure types of strategies where it's you know I might have a good prior on the fact that you know there's an 80 chance that I make a profit on this trade but it might not be that Atomic trade right and a really good example of this is that got commented last year is something like nft mints right where I'm spending e gas fees today for some nft that I may be able to sell for a profit in the future and if I have a very high prior of the fact that I could be profitable that's not non-atomic but it's decent enough for me to participate yeah that makes that makes an enormous amount of sense um I'd be curious because like one of my sort of mental models for for this was

you know when you talk about statarp at least in in Trad five this is what you know house and I have discussed this but what Citadel Securities is sort of famous for right and uh you know they have sort of this practice where they have all this um they have a huge Advantage seeing this sort of dumb order flow from a venue like Robin Hood and while they don't trade directly against the customers on Robinhood they use that information as signal to make more informed trading decisions so the example that we've talked about is you see a bunch of buy orders for gme it's probably relatively safe to assume that there's going to be a bunch of buy orders for AMC or very correlated type asset um so I guess are the same it sounds like similar games are being played on

ethereum uh or in crypto so to speak so far and I guess my my question for you is if if if the if the opportunity set trends that way so that most of the Mev being realized is non-atomic and that has higher Capital requirements and it favors an organization like wintermute that has a bunch of ethereum on their their balance sheet do you see the overall trend of Meb extraction again moving towards some of these larger sort of hedge funds if that's the way the opportunity set moves and Dean open question to you as well yeah I mean generally speaking that is the way I see the trend moving right it's people that have the

inventory and can take the risk look I mean the question I I see the trend moving that way as well the question is always what at what point do we still call it Mev is some of this that ARB stuff actually I mean um I think that's one of the big debates as well right because in my definition of nav I only view the atomic things really as Mev um I've although we do non-atomic things as well I don't I call it me because that's what everyone else calls it but that's not really what I for you as muv maybe I just have a very old definition of the term

but it would be I guess it's it's an interesting question but um I think if we expect that the non-atomic firms are going to crowd out the the atomic Traders because they have a lower cost they can bid up the opportunity more than is it doesn't mean that the Mev goes down I think that would be illogical right then if he doesn't disappear just because it's being captured by non-atomic players yeah the well yeah I guess I mean you could say the Meb goes down that's a different attack now um yeah I don't know I don't know I think that's like a broader discussion on how to define things but in the end I see

the space going that way as well right I think that a lot of this stuff is going to professionalize um and a lot of like strategies from trying to find things are going to roll over into how some of this stuff is done I think a lot of Envy shops right now are unsophisticated um they haven't looked into things like how inventory is handled at hft firms and things like that but then I also think that hft firms will inevitably have a lot of opportunity to do um that they didn't have to do in um in uh the Trad five space just because of the nature of the system that they're interacting with right the way to win an hft is you pay more money to

be closer to the to um to Wall Street essentially so that you get the shortest cable um or that's one of the big ways to win right um that's not really the way that you end up winning in any V or I don't think that's how you end up winning an MVP in the long term because otherwise you're just giving all of your money to various validators or you're using a public private Med pool like flashbots which is of a disadvantage to you as well for what it's worth really quickly I don't think this inherent centralization that we're going towards is a totally bad thing either right for what it's worth a lot of what happens in Mev today is reapplying the

same things that hft in the traditional world has seen 10 15 years ago but just doing it in a more decentralized and publicly auditable way right where today you can't see what order flows Citadel is trading on that comes from Robinhood if Citadel were to do that same type of thing in ethereum today or with you know the various designs that have been proposed to that type of stuff it'd be much more public and auditable right so it's like a better way of doing what tradify has already experimented with I mean just on that uh on that thing Dean you know you're starting to get into kind of a thread of the the formula for winning in hft in Trad Phi is you buy up space in in New Jersey that's very close to exchange right in this kind of the

Chicago New Jersey pipeline so to speak if you had to you know be super reductive for Mev describe okay so that's the way you win in Trad fight hft what's the way that you win in crisella movie well I think it's actually I think it's like almost outdated right that you only need like a need to be very close to yeah that's that's a very outdated it's a very outdated view um of it I do you have the you have that plus the engineering chops that you require right um now because everyone has the shortest cable you inevitably have to do

optimizations elsewhere elsewhere being uh making sure your code is the fastest um and I think in in um Trad fire looked more cable less code optimization whereas in Mev right now it looks less cable more code optimization that's at least how I view it um but I think that probably shifts to an equilibrium on both whereas probably in hft it might already be at that equilibrium I think it probably differs by uh how the sequencer works right so on ethereum layer one where you have Mev auctions uh you definitely you have more code

optimization than size of balance sheet and on on layer 2's where you have a kind of the latency auction now the like quote-unquote the cable length is very important but something else that I was alluding to I think it's it's a the optim the code optimization I think it's the size of the balance sheet I think it's the proximity to exchanges but it's also how fast you can move how fast you can ingest information which is what both of you were talking about in kind of your data ingestion efforts and that's also how fast you can move information across the globe um in a way that doesn't rely on like public internet infrastructure right so I think we're seeing a lot of the hft firms operating their own microwave Tower infrastructure I don't know if any

of them are operating satellites that's probably not faster I'm not sure so um but yeah I wouldn't be surprised to hear that sleep from my perspective I think what works in the traditional world with centralized exchanges is basically a one-to-one mapping to how it is in need right from my perspective the three things that you need to do on right is how optimized your code is how low level are you how good is your networking stack right how quickly can you ingest transactions how quickly can you emit orders how quickly can you cancel orders and how much exclusive order flow do you have from venues like Robinhood and you know things like that and the way that I see it that's identical to what makes you

successful in crypto save for perhaps your crypto Nativity and how deep you can go into protocols right other than that it's you know how close am I to the sequencer how quickly can I ingest transactions right how much of the validator network do I see how close to the MMI right do I have exclusive order flow from some dapps or some exchanges and things like that um so yeah for me it's mostly the same thing right it's what works in the traditional World works really well in mbv which is why there's a lot of traditional firms that are now coming into Mev and are doing decently well right it's their bread and butter inside of you working with exclusive autoflow

so not one that that you I'm not speaking about autoflow that you create rather but that that you purchased and the purchased and then interact with in some way I think you'd be hard-pressed to find any Searcher that would admit that they pay for order of thought okay that's also an interesting answer I mean this is something that we kind of touched in an issue sort of just hit it with your with your three buckets there but I'd be curious um just how you think about the The increased role of latency in in the searching game like we um I think josu's kind of flagged it as like what you don't necessarily want at least from the perspective of ethereum uh you know is

to be an Mev system that's completely optimized on on latency because that just presents for certain risks uh sort of geographical risk and Regulatory risk being being very prevalent there but I would just um I'd be curious to to understand how latency is kind of the role of latency has kind of evolved and maybe going back to kind of your early days of searching like how much of that was a vector of competition for you uh how much you know is it do you kind of spend uh how much time do you spend thinking about latency today and how do you expect that to kind of evolve into the future sure couple of things to unpack there I guess um the first thing is how does late and see play a role today and I'd say Network optimizations are

probably the easiest pickings I can get in Mev right where even before I try to optimize code hot paths um if you're a competent Searcher you write code out of the box that isn't particularly unperformant right it's it's not a horrible code um where you in optimizing that might shave off a couple milliseconds but you're much better off trying to optimize your networking stack where you're playing in the tens of or hundreds of milliseconds to get around the globe and so I'd say in terms of latency for at least myself just optimizing networking and getting better latency has the best Advantage um it's especially important in l2s where it's the entire game for things

like arbitrary and optimism is how quickly can you see a transaction and how quickly can you fire off your transaction and that's true of things like you know like you know Dean's on BSC and I'm sure he can say that you know back running on BSC for example is 100 a game of like networking infrastructure right it's how quickly can I see and respond to these transactions and so I think that's where latency plays in on ethereum um yeah I mean latency continues to play a role um even in ways that we wouldn't expect like a good example is um there's a latency that's involved with simulating blocks that are submitted to relays right so if I submit a block to the flashbots relay today it takes say 200 milliseconds for them to

validate that that block is valid which means that I miss out on the last 200 milliseconds of a 12 second window to submit a block versus say ultrasound relay where they have optimistic relaying um which means that I now have better performing higher paying bundles and blocks simply because I can fit more transactions and don't have to worry about that latency loss due to simulation um so that's latency in like a different way that it affects ethereum today and this is like data we've seen over the last month and as they've enabled optimistic relay I I don't know if you have anything to add Dean but Anisha that's just interesting to hear you said that we had uh Justin Drake on uh in our first episode of the season and he was

actually talking about from his perspective as the operator of the ultrasound relay just how important latency was and some of the small adjustments they made so that they could lower their their latency even further and that win with more blocks so just interesting we can hear you validate the other side of that yeah I mean I'll validate it again the optimizations we focus on 95 of the time is latency networking because yeah usually a wind there is a 10x multiple on what you would get over some shitty go memory or whatever language you're using optimization anise I would maybe push back on something that that you said a little earlier where you said maybe it's not so bad for crypto if we end up with a similar

Market structure as chat file um because I mean I would also argue that I think um like uh wholesale P5 gets sort of an unnecessarily bad reputation uh in terms of the execution that it provides the execution is great but I think what it does it's very centralizing on the market structure in a way that's not a big problem in thread fi but in crypto you kind of have um you know the block Builders also playing this very important role in providing like very fundamental Network guarantees and I think if you have too much of a concentration on that layer then it ends up affecting kind of the whole crypto stack and that's my concern

why why I think we kind of need to try to evolve the market structure in a way that that helpfully leads to a to a less concentrated outcome that makes sense to me although my Counterpoint there is block building today is already centralized to the large extent in terms of competition where you know the same five or six block builders make up 95 percent of blocks on ethereum today and I expect those same entities are also going to be the ones that are you know taking proprietary order flow and then putting it into their own blocks and so from my perspective at least it's like you know the shift doesn't really look much different than where we are today right and it may actually like you said better execution it's probably going to

be more beneficial for users at the end of the day and you know less we'll go to validators and their Kickbacks and things like that and the entities that are involved will largely stay the same from my perspective right so it's like we're only getting better from here on out right we're not really making anything detrimental um no I get what you're saying don't be not like that inherit centralization Factor but it'll become much harder for the average entity to compete and to join in on uh compared to today I agree with that yeah and I mean they always show techniques that we can enable I think that that kind of uh make uh stuff like censorshipada right um stuff like enabling kind of private mempools inputs that may also be private

from the Builder uh but also techniques such as partial block building um where maybe the the station committee builds part of the block and the block Builder builds the other part of the block or whatever I think um yeah this is all stuff that the talic laid out in this endgame post which I would recommend that we want to read so I think yeah it's very interesting debate so maybe we can uh sort of start to to wind down here but I would love to um I'd love to just get your thoughts like actually moving out of the weeds here for a second and sort of getting your your thoughts on you know one very popular debate in Mev is it kind of manifests around the fair ordering versus you know auction maximalization but you know there is still kind of a very consistent Camp of

people who say Mev is this sort of bad negative externality that should be avoided and we should do everything that we can in terms of mechanism design or thinking about it when we're designing consensus or how these systems work to mitigate Mev or just sort of eliminate it and then there's kind of this other camp that says Mev is inevitable whether it's in crypto systems or Google AdWords or wherever um wherever it kind of exists in life and ultimately what we should be doing is extracting that Mev and then making sure that it gets distributed you know back to the users or to the validators that are you know um securing the network or you know Searchers that kind of come up with these algorithms or whatever it is so I'd love to kind of understand where both of you fall on that that spectrum and then like you

know not you know it's kind of a high-minded word to use but sort of the morality and ethics that underlie Meb I think Mev is inevitable it will always exist in systems and you should protocol designers application designers should think of Mev when they are building their applications and you know attempt to minimize it to the extent that's possible but it won't be it will never be possible to completely eliminate it from my perspective um as for the ethics frankly I think in my role as a Searcher I do a service to the ecosystem by providing better execution and extracting what value you know if it's it's kind of like if it's not me then

that value isn't extracted at all right and this leads to more efficient prices this leads to potentially better Kickbacks for users in the future leads to better revenues for validators and so ethically I think it's perfectly fine to extract MPV yeah um Mev is inevitable I have always set that you know there might be we might be able to reduce it to some meaningless amount but inevitably it will still be there um I do think that we should spend time um you know reducing it as far as possible on the protocol level and I think the eth uh core research and stuff like that is definitely leading on that front

um I ethically I feel the same as a niche I I view Mev as an accelerator from an accelerationist standpoint it's like you know somebody's gonna have to extract it and show uh where the flaws lie um and you know looking at it in theory we only get so far you know a lot of the issues probably weren't apparent to us two three years ago until people started actually doing Mev um and we and we uncover all these inefficiencies on this network and so yeah I think not only is it doing a service from um from the points that Anish mentions but also

um you know it's helping to uncover and uh turn ethereum and all other chains into more anti-fragile systems both of you dropped so much um good knowledge in this podcast uh already so thank you for that I do want to end on the crescent if if you if you had to kind of take an intern in the Meg space like there's a new like tang and hungry kid upcoming Searcher wants to break into Meb what how would you recommend they go about it what is the kind of skills that they need to pick up and uh would you recommend kind of they reach out to what's their starting point know how to code and know how ethereum works and then just DM me

um for anyone who's DM me before knows that I'm pretty responsive on Twitter um and I've it wouldn't be the first time I hire someone who sends me a Twitter DM and a GitHub link do you take add-ons yep we also have we have I mean most of the project block team is publicly Anonymous um we have add-ons even at dialectic at the fund so yeah yeah I'd probably say know how to code very well understand ethereum and then honestly there's a lot of opportunities that are still public you know flashbots maintains a GitHub repository of you know here are some opportunities that

need Keepers or Liquidators or things like that and even there there's still money to be made I think the only way you can get into it is by doing and so you know whip out a code under and even if you won't make much money try to be competitive in one of those strategies and that'll teach you a lot and get you down the revenue give me come on give me like an underrated quality in Med what's something that like nobody else appreciates that that you like if you had to hire someone that you would actually value a lot I mean the thing I mainly look for is someone's resilience how many times they can get punched in the stuff I can keep going um and that's more so than any other job I don't think that's a unique skill I think you just like need to have a far

higher pain tolerance working in the Mev than you do in software engineering in general any advice on how to test for that I think you can easily tell that from a person's character from like three days of working with them that's probably the most practical like yeah work with someone for a bit and see how fast they give up I mean the most common Trope is probably that you know mbv is sexy and everything I'm going to do is landing arbs and making money and I'd say that's maybe five percent of it 95 of it is you know taking systems you've already got making sure they don't break maintenance

maintenance maintenance over and over in an environment where inherently everything is set up to break around you and so yeah plus one on the resilience that look for people that don't mind the grunt work of just being in there and just doing it every single day and it might not be the sexiest work of finding arms that's that's once in a once dude I've got a question for you actually I would uh just as someone who runs an Mev shop I mean what's your strategy for maintaining you know Finding and maintaining people in a super specialized field because I would imagine one of the one of the challenges is a bunch of people would say hey like uh like I could either work at project Blanc or I could do what Anisha is doing

and then just go and sort of find my own long tail Mev and I'm assuming that that's a pretty you know attractive opportunity kind of do my own thing how do you sort of keep um my answer to that as always my answer to that is always you could make a million dollars a year and work like 18 hour days or you could make 10 million dollars a year and work 10 hour days and I'll offer the latter but anisi is really happy with the former pretty attractive uh value proposition there Dave that's uh that's pretty good if um I mean yeah do you you pay them well that's how

all right I mean I think I said all of these things are going to turn to teams just because the yeah the overhead of doing several strategies by yourself it's a it's a lot of work like everyone we hired on the team went through a burnout because they were trying to do it themselves all right guys well you've both been uh super generous with your with your time um if folks wanna you know getting either get in contact or find out more about the work you do I will plug you uh Dean here and say the project Blanc website is probably my favorite uh website I've seen in in quite some time you've got a very simple message on there but uh people can people let's see I have to change it while the

lawyers were setting up the incorporation because they didn't like what it said actually you know what could you just say because project Blanc has a very cool sort of History I'm a huge nerd for this kind of stuff like what's the what does project block mean what does that refer to it refers to these two guys in France back in like the 1800s or something and they did Arbitrage because what happened in France is the stock exchange was in Paris and the way they would notify price updates is there will be a coach every day that went from Paris to wherever they lived which was like two or three days away by horse

but Paris or France had this great infrastructure of uh semaphores everywhere across uh the entire country uh to which the government could communicate messages on what these two brothers uh realized is that there's an error correction system and so what they did is they bribed the guy at the semaphore in Paris and the guy where they lived to um do price updates within or put price updates within the error correction system of this semaphore and so they get price updates uh within hours rather within a days and were able to trade based on those price movements

some two hft should and nothing ever happened because there was no law um for this so it was truly the first Mev right like they um were taken to court and everything but the French government was like we have no law for this they this is fair game that reminds me I think there was this uh it might have been the Rothschild there was one of those old you know families that made a ton of money in sort of the old world but they had a similar strategy in the Battle of Waterloo they paid a messenger uh to basically run back to to London uh faster than anyone else so they could trade off that information so yeah Mev existed long before long before crypto um so yeah definitely check out uh Dean's website and then Anish if you

wanted to people on to follow you or find out more about your whatever just what's the best way to do that they can just reach out on Twitter the easiest way to reach awesome all right well this has been a super fascinating conversation guys uh thank you so much for for joining us thanks for having us all right um man what an interesting conversation it's uh I was really looking forward to this one like you know hearing from people who are actually doing kind of live searching um so thankful that both those guys kind of gave us their time oh you're what were your takeaways hatsu my takeaways I I they weren't they weren't sharing quite as as much and as openly as I was hoping to be like quite honest with you but I I

still thought there were some real nuggets in that conversation um I think um especially Anish I think he he dropped uh a lot of interesting um just like high level descriptions of of how the Mev space works and what it takes to be a good Searcher I thought that they both did um I think this like control of the networking stack uh having exclusive order flow but then also just like deep and very like low level understanding um of the domain that you're working on and like ingesting information and so on I thought this was a very very holistic overview that we got from them um about the real kind of operations of

a search or of a of a searching team yeah it was you know what I was trying I like the the buckets that Anish sort of gave us for what makes a good Searcher and I was kind of hoping to get a stack rank for those like if you had to prioritize sort of um you know networking versus code optimization versus kind of your your latency or your view of the network um but it sounded like all all three of those things were frankly very important and some of that just depends on your strategy like I would sort of guess based on that conversation that for short tail maybe your view of the network is slightly more important or access to proprietary order flow and then if you're long tail that's slightly less important and you really want to be

you know your networking stack and code optimization you got to be yeah so Anish didn't say that but I know for a fact that there are long-term long tail Searchers who deliberately do not even use flashpot's protect relay and not any of the other builders either because they actually use kind of they use the public mempool that they kind of like direct to a validator connection so if they run even their own weather data in order to to get a transaction um mind in a way that other Searchers don't observe it after the fact right because it's a lot about kind of stealing uh it's not about like an unbundling like we've seen in the past

like like a couple weeks ago in in math post it's more about such as analyzing after the fact well other structures are doing and then copying the strategies and getting alerted to nice opportunities and so if like all transactions that are being mined by like the main block Builders those are under the microscope right by the entire social community and that's why some Searchers take the deliberately kind of the long way in order to get transactions mined um that that maybe may be more expensive maybe have higher latency but it doesn't matter because they don't have a lot of competition in the particular uh strategies that they are doing and so it's okay so for them it's actually more important to stay under the radar

yeah that's very well said I um it's funny and I even in your interview interview with the Searcher when was that three years ago maybe in 2020 uh he set Mev Senpai said something very similar to that which is for some of my my juicy opportunities I won't submit it through through flashbots uh this was when it was still proof of work so it's kind of interesting what I don't know if you have something to say there I had another I had another takeaway actually which was you know a lot of this season I'd been thinking about you know I kind of went into this episode wanting to ask them quite a bit about like the search reverse Builder Dynamic but it was interesting to hear Dean say ultimately where he viewed his

sort of competition and his end game is in between himself as the Searcher and the validators and that was a sort of a light bulb for me which I guess when I looked at it like should be sort of obvious um and I'd be curious you know what you kind of thought of that and then I guess the the obvious conclusion to that is yeah if I was a Searcher you know I could either send these bundles in these you know interesting Arbitrage or whatever sort of extraction opportunities to someone else and and hope that it gets put on chain or I could just run a validator myself and that kind of seems obvious so you have to put that in a context that that kind of Dean's firm is on a lot of chains that do not have proposal Builder

separation so um whereas you would as a Searcher you can submit your bundles to block builders in ethereum and then you can get access to um to to a lot of the blog space a lot of the values to map boost that same mechanism um does does not necessarily exist right and so um for many opportunities um you kind of do not have this luxury of submitting um directly you don't don't have this channel where you can submit to to validus and so um that's why very often it's it's the best thing that you can do is actually run a lot of the editors themselves and this is I mean search us see this as

something positive but at flashbots we see this as something that's very negative right so this was this vertical integration between searches and where did US was the entire reason why flashbots end up creating uh the original mathcaf right because we did not want that um uh kind of this like vertical integration between the two roles um to the point that some validators would end up making a lot more money than other validators and causing centralization on the validator layer so um I think it's very understandable that um players like and Dean and project blank are kind of pursuing this route of running their own valve datas but this

is something that the uh this layer 2 earlier one ecosystems really need to think very carefully about um if they see that this is something that's going on and I can assure you it is going on on kind of all of these chains um like what do they want to do about it and my response would be you need proposal Builder separation in order to protect the decentralization of your validator validator layer yeah very well said and you know just to give listeners a perspective you know the 12 there's a very big difference right between the operate the environment that Dean is operating where there's 12 validators that are operated by binance versus how many validators are there on ethereum there's some 500

500 000 000 yeah yeah not at the same time but yeah yeah exactly so just super super different environment obviously but um it was just it is it was just interesting to hear his hear his perspective there um you know you know what else um I thought was really interesting how they both said um 80 of searching is basically latency optimization so I think in the previous episodes we were talking a lot about this latency between like a builder sending a block to a Val data or kind of a search for sending like a bundle to to to a better but I think what this episode really Drew out was

looking at the entire almost like life cycle of an MAV opportunity because they both were kind of alluding to this idea that you first you need you basically need this elaborate uh pricing models right you want to build this picture of what the fair price is should be like a what the what the market prices are of different Assets in the world and then B what uh you think the correct prices should be right and then you want some kind of mechanisms to act on that and so there's a lot of um ingesting information as fast as possible and then simulating against it and then Translating that into some some opportunities that you want to put

into action and we only talked about the latter but I think what both our niece and Dean were saying is that actually the former this like ingestion step this is actually the even more important part and there's nothing really that um for example core developers in ethereum or flashpoints Etc can do in order to make that any easier for these selections right and so this will probably always be something where a lot of competitive Advantage can be had and that can end up being kind of a very centralizing Force as well did you agree with what Anish said that basically everything that worked in Trad fire and hft is largely going to work in crypto

yeah I think um I think it's it's like I think it's probably right but I don't think it's desirable right so I I think we can just empirically say that all of the strategies that are proven to work in trade fire they are coming to crypto they are coming in a big way so there are um kind of major trading firms in crypto that are now running their own microwave Tower um basically right so it's just like this stuff is is like coming and it's not just that it's like crypto firms kind of adopting these techniques no it's also from the other end you have the big firms from outside of crypto now pushing into crypto and they bring all

of these expert bring all of this expertise and they bring all of this existing infrastructure right and so I think you kind of see this uh convergence um from both ends I think I think it is on us so we have the opportunity to actually build systems that are maybe more fair more open and kind of uh are better in some ways than the system said we have in sharify uh even in even in uh in chat fire so for example from uh professor professor Eric Buddhist we have seen sort of excellent uh proposals uh about kind of the the New York Stock Exchange adopting frequent batch options and he's kind of made all of these models show how they

can eliminate uh the Mev that is most of the Mev that is kind of leaked to these latency Wars between hft firms and um but they are kind of like political and historical reasons why these things are very very unlikely likely to be ever adopted in trade fine but here we have kind of the chance to build entirely new systems um that do not have the same amount of entrenched players yet um and so I think I think so I definitely think that it's correct and it's a threat that we need to be very Vigilant against yeah what do you think about I liked by the way that word the life cycle of an

Mev opportunity I thought that was a really interesting concept um the you know there's there it seems like in Mev it's almost like centralization whack-a-mole in a way where you know you focus on trying to prevent centralization in one part of the supply or the value chain like let's say the validator level right now but then what I heard Anish say sort of in response to one of your questions as well right now 95 of the blocks are built by these highly centralized Builders right there are like five of those and if you're a builder you have a lot of advantages that you don't necessarily have as a Searcher for instance you can you know do like you can have bundles that sandwich other bundles right you just have the entire view of the block and

there's a much richer design space for the for the sorts of uh extraction opportunities that you have and then even I know we'll talk about this um uh with with with John and uh and Robert but uh you know and you you I heard you discuss this with like PBS kind of at the at the layer two level right or at the roll-up level uh right we're sequencers today it's almost like taking a step back and uh now we don't have PBS on on layer twos and that's where a bunch of the the value change is gonna happen so maybe less of a question to you than just an observation that it's just very tough right and centralization pops up in a whole bunch of different spots so all right well I think we can call it

bear um I think the the next um episode that we'll be doing is actually with uh some of the guys from Solana so we're going to be talking to uh Lucas um and Eugene from Solana and that's a very different uh ecosystem uh Mev works very differently on Solana than it does on ethereum so we'll have a lot of specific questions uh for them about that but uh for now this has been a ton of fun haasu and yeah I look very much forward to the episode on Solana because I think it'll I think especially with Eugene I think we can have a great conversation about the pros and cons of latency um I think you make the he'll make the argument that that you want low latency

systems because they can generally provide a better service but then you still need to balance that with kind of providing a fair uh any of the extraction system which is where I think then G2 Labs is going to come in and we will try to understand exactly how how they can provide a flashbots like offering on on a low latency chain so that's a very interesting prototype I would say or case study where if this works then maybe we can find techniques kind of to to have lower latency and still preserve the this kind of decentralization guarantee absolutely all right signing off [Music]