Episode 216: A Dip into the Mempool & MEV with Project Blanc

Transcript

Welcome to Zero Knowledge. I'm your host, Anna Rose. In this podcast, we will be exploring the latest in zero knowledge research and the decentralized web, as well as new paradigms that promise to change the way we interact and transact online. This week I chat with Dean Eigenmann and Edgar Arutonian from Project Blank, an MEV focused team working on arbitrage and MEV strategies.

We revisit the topic of MEV, what it is, and how participants in the network can capture this. We also dive into the concept of reorgs, accelerationism as a philosophy, and get a look at what goes on in the dark forest. But before we start in, I want to let you know about our upcoming ZK Hack Mini event. Last fall we did a seven-week event with puzzle hacking and workshops all focused around ZK learning. Now we bring it back for a mini version of the event. It's set to last one to two weeks. It starts on March 1st. If you missed the first edition or if you want to participate once again, be sure to sign up the newsletter at zkhack.dev or sign up directly for our first event to get updates on our partners for this edition and our schedule.

Also, if you're looking to jump into ZK professionally, I want to remind you to head over to the ZK jobs board. There you can find job posts from some of the top teams working in ZK. It's a great place to find your next project. I've added the link to all of this in the show notes. Now I want to pass the mic to my producer, Tanya, who will let you know all about this week's sponsor. Today's episode is sponsored by Polygon. Polygon is one of the top Ethereum scaling platforms for developers with over 130 million unique user wallets. They're known for Polygon POS chain, one of the most popular chains built on top of Ethereum, sometimes clocking even more daily active users than Ethereum itself. And the team is actually working on a spectrum of solutions, including EVM-compatible zero-knowledge roll-up

solutions like Polygon Hermez, Polygon Miden, and Polygon Nightfall, which is built in association with Ernst & Young. They also recently announced their acquisition of Mir Protocol, now known as Polygon Zero. These solutions inherit the security of Ethereum and use the power of zero-knowledge. Over here at the podcast, we've been tracking Polygon's journey and have recently had Bobby and Threadbare from Miden and Jordi and David from Hermez on the show to talk about their respective projects. We strongly recommend you check out those episodes to learn more about their vision, project roadmap, and progress so far. Polygon is also hiring and have a few jobs listed on the ZK jobs board.

You can check out that link in the show notes. So, thank you again, Polygon. Now, here's Anna's interview with the guys from Project Blank. So, today I'm chatting with Dean and Edgar. Dean is the co-founder of Project Blank and Dialectic, and Edgar's also the co-founder of Project Blank. So, welcome to the show, Edgar and Dean. Thanks for having us, Anna. Thank you. Okay, I think we need to start. What is Project Blank? Well, it is a team that Edgar and I put together. Edgar, who was previously at Flashbots, um and we decided to start an MEV team together. Um and so now it's a team of, I think, five guys. It's essentially run like a prop shop where

we do arbitrage and um execute other MEV strategies, uh while also doing a ton of research, which will probably be opening up relatively soon. Mm. Let's dig in a little bit to kind of the history of this project and maybe the history of your work in the community. Edgar, this is the first time we're meeting. Dean just said you used to work at Flashbots. Like, tell me a little bit about your entry point into the the space, maybe even leading up to that. Like, what got you interested in MEV stuff? MEV, I think for me started when I tried writing my own arbitrage bots.

Okay. And then there's a whole history of work that I did at at that point. Mostly on my own. Actually, all of it on my own. And then there's some period past. This was in 2020, at the end of 2020. Uh just to get to the focus of Flashbots. And um I got a conversation with Georgios in January 20 20, first time he told me about Flashbots. Mhm. Um and I kind of really wish that I started using it then. Cuz I burned a lot of ETH not using it. Um but uh there's a whole bunch of other stuff that happens in between, and long story short there, around March, I think it was March, I joined uh Flashbots.

Cool. And what were you doing there? Like, you sort of said you wished you had been using it, but like were you developing with it? Were you building it out? Or were you like utilizing it for something? When I worked at Flashbots, I worked on MEV Geth itself. Mhm. And uh research project on SGX. I see. This is actually something I've always had a question about. Like, the Geth client it's it's like the the Flashbots Geth. What is that what it's called? Yeah, MEV Geth. MEV Geth. So, MEV Geth is like it's the equivalent of the Geth client that is developed and maintained by the EF, I guess, but it has some adjustments, something that allows people operating these nodes to actually tap into MEV,

right? It's regular Geth with like 300 lines of code added. Oh, really? Okay, so what is what are those 300 lines actually do though? They basically control how the block is made. And at the moment that the block is made, if the miner is running Geth as if it was a MEV Geth, Mhm. it checks another transaction pool. Like, you have like usually just one transaction pool in the guest client, but here they make like another transaction pool. The private transaction pool, that one can only get populated by trusted IPs. Basically something like that.

Mhm. And how does that actually like the way you describe it, it sounds very very like light touch, but what what is it doing? Like why would anyone use this? What is it actually accomplishing for the miner or yeah, the node operator? It lets the miner make money without having to do anything extra other than use a different binary. Okay. And that's kind of this MEV, which we should potentially define for the on the show. And I know it's actually changed definitions sometimes. Like is it miner extractable value? Is it minimum Well, like what's the what's the acronym?

The acronym was kind of like um it's too nebulous now. I think it's just whoever controls block production. So it originally means miner extractable value. Mhm. It's like if there's some entity that controls the order of execution. That's like the exploit the the bribable entity. So like in the general case, that's the miner. Was there like another word they're using instead of the M in MEV? I've seen people say validator. That's kind of true, too. I mean Mhm.

we work a lot on BSC, so it's validators, not miners. So it's kind of interchangeable. Cool. Um okay, so Edgar, you were join you joined Flashbots, you were working on this geth client. You were I guess part of that team. What made you want to switch and start something new? Like what's what's different maybe about what you're working on now than what you were doing then? Uh working on the clients is really hard. Uh but also there was this kerfuffle in the summer about reorg and basically I wanted to step aside to focus on the reorg path and not bring any bad attention to Flashbots.

Got it. And actually, I kind of want to maybe later on explore a little bit what that was and what that is. But before we do that, let's actually hear a little bit of Dean's story. Dean, we met many, many years ago. Uh I think it was in Zouk, maybe? No, I think we met in Paris at at the first ECC. ECC yes yes. Cuz that was like my first outing. Dean was so young back then. I think I was 17. Yeah.

So, you're 17 years old, you're going to your first conference in Paris in this space. What were you up to at the time? So, at the time, I was still working in like Tradeteq in Switzerland, but I was also on the side doing audits together with Matthieu De Ferrante at ZK Labs. Uh Matt is an amazing security researcher who was um also part-time on the EF security team. And he had started an auditing company, and we'd gotten into a argument on Reddit, and then at one point, he was like, "Do you want to work with me?"

Nice. And so so that's how that's how we had met. Yeah, and then I did that with Matt together for one or two years. At the time, I also joined um ENS cuz Nick Johnson was working on it completely alone. Mhm. Um and so I I joined Nick on that. I helped write a lot of the code. Actually, most of the smart contracts deployed now are the ones I wrote because um I don't know if you remember remember, but there was a bug in the ENS contracts, and they needed to all be migrated, and Nick was getting married while that was happening.

Oh. Um was handling that while he was getting married. Um but then after ZK Labs, I joined Status to work on messaging research cuz uh they were using Whisper, which clearly didn't work, and so we started designing a a new messaging protocol. I think you've spoken probably with Oscar before. Yeah, Oscar and Sanaz. Yeah, they were both on the show actually talking about the project Vac, uh VAC. Yeah, so that was it originally started essentially with Oscar and I. It was a research group underneath Status, which has now grown to a relatively large team. So, I had left or I left just as Sanaz joined, I think. And then I like

tried to do my own Web2 startup. Let's talk about that. Oh god. Come on. I I got sick. Honestly, I got sick of crypto. I feel like being so close to the sauce all the time. Yeah, I I fully rage quit. Um I I think that crypto has like this thing of like burning out everyone who works in it. Definitely. and I was I was pretty faded. I felt like things weren't happening fast enough, and we weren't really uh delivering on our promises, and so I I had left or crypto to try and do a Web2 thing, and then felt the pull pull me back in, and now I'm ironically bullish, which is not something that I've ever been.

That's funny. Let's talk I want to talk a little bit about that experience though. So, you left crypto, which you know, we've over the years I've known people who've rage quit crypto never to return again, and some who took a pause. You were one that took a pause. What did you find kind of on the other side? Like did you feel when you were leaving like I want to get out of this space cuz it's so XYZ, and it's better over there? Or were you just like I want to build something? I thought crypto was salady, and it turns out the traditional tech is equally salady.

Um like I One of the things I noticed is that also you can't carry over credentials very easily. Like I think in crypto if I were to raise money I could fill around probably in the next 3 hours. Mhm. Um not the case in traditional tech. It's very good to do this though. Like to to take the step back and to like that it was like an 8-month period and I think that was probably one of the most like important growth periods outside of the stuff I've done in crypto.

And it made you want to come back. So you like after that experience you actually wanted to come back into the space and now you're excited about it. I had founded Dialectic while still at Status actually. Okay. So Ryan and I did that while still at Status and for the first like 3 to 6 months it was essentially me, Ryan, and two others. Yeah. And it was just us doing like OTC deals and things like that. And then at one point it just blew up and now it's a team of 20-something people.

Um the fund started with sub millions of dollars and now it's like grown way beyond that. Um I remember we had our offsites um in St. Moritz which we do which we've done twice now. We had that this year in December or last year in December, sorry. And Ryan and I like sat down looking at each other and like both like can you believe that it it's grown to this scale? Wow. Especially when we were like looking at the list of the people we're still hiring. I think we're going to hire like another 30 or 40 people in within the next 12 months.

That's wild. But like so Dialectic is a VC, right? Like what is it maybe? Maybe you should define what it is cuz why are you hiring so many people? It's a family office. Mhm. So it's structured as a family office meaning that we have clients which are ultra high net worth high net worth individuals and we help them on their personal lives as well as like their financial lives, right? So we we help them um invest and things like that, diversify their portfolios, but we also help them on personal things like ensuring that they have PAs and someone takes care of booking their flights and mundane stuff like that.

Mhm. Um that's that's what family offices do. Okay, so it's a family office, but why are you are you hiring because you want more PAs for the clients or are you like it sounds like you're hiring more as like an expansion into the technology? Um so, we're hiring just because one of the things which has changed about VC is that you're no longer a capital allocator. Um especially in crypto where there's so much free money, right? Um no one like as I said, I could probably fill around in 3 hours if I was raising for something in crypto at this point.

However, that that's not what most developers and most teams are looking for anymore. They're adding they're looking for uh investors who can actually add value, which is the meme that every investor will tell you that they're a value-add investor. But, in crypto, what that means is having your own research team, having a development team, having marketing, having hiring, and having all those things for your portfolio companies, right? Like um as a VC or as a fund, you now have to provide all these auxiliary services for your portfolio companies to actually be a compelling uh partner.

How does Project Blank work with that? Like, how does it connect to this Is it a different org or is that within this kind of research part of Dialectic? So, when when we were starting Project Blank, there were multiple questions on whether or not we do it as part of Dialectic or do it as its own company. Right now, it's it's it's it's completely own entity. Dialectic owns a part of it, obviously, but the the reason why we did that mainly is because of remuneration. People who work on MEV are making so much money themselves that like you can't convince them to join an organization where most of the equity is already doled out, right?

Okay. We needed to create a new entity so that we could give double-digit equities to the developers working on the project. So, you kind of did describe it at the very beginning of this episode, but like so it's not about client maintenance, it's about using existing tools to like leverage MEV. Like you like you sort of mentioned the strategies and stuff like that. Yeah, Project Blank essentially is a team of four or five people who are executing strategies. Okay. So, you can consider Project Blank as a prop shop.

What does that mean, prop shop, when people say that? It exists to make money. Yeah essentially. Okay. What does it actually stand for? Proprietary trading, I think. Yeah. Yeah, yeah, it's proprietary trading shops. Which Which are Usually you have them in like bigger funds, but it's a smaller teams of people who are doing high-frequency trading and things like that. Why did you choose this This may be more to Dean cuz Edgar, I think the reason you would choose to work on this makes total sense. You were already working on it before, but Dean, why would you choose to do something like this? Like why did you choose to go in this direction? Cuz there's lots of ways to make money. Like what's Why Why MEV?

So, we had been looking at MEV for a while in the team. We just didn't have the expertise. I was the only one who's ever really worked on clients. We had developers as well, but like MEV is a very unique set of things that one needs to know to be able to be successful. And so, it requires a very specific skill set from the people working on it. And so, we were always looking at it and we're thinking about strategies. And then at one point, I DM'd Edgar just as he had left Flashbots and we started a conversation and it kind of grew from there. So, it wasn't actually fully planned. It was like, "Okay, we have this opportunity.

Edgar, are you down?" And we managed to make it work. Interesting. Like I'm trying to picture what the the team structure, like what do you guys actually do day-to-day? Maybe Edgar, this is one for you. Like, are you researching what people are currently doing and kind of trying to emulate that? Are you coming up with new strategies yourselves? What is that even What kind of work is that? It's all of it. Okay. It's really all of it. Every Everybody on the team is kind of um it's like a 24/7 thing because we're always trying to improve our own strategies and also noticing some of competitor strategies and also trying to reverse engineer or understand competitor strategies.

So, it's it's kind of like um I mean, it's so evolving that there's always like its own set of challenges in addition to the fact that sometimes like the ground rules themselves change. Mhm. Um so, that also makes it even more challenging. What is a strategy? What does it mean? Like, can you paint one Can you paint a picture of what that would even mean? Like, you're running a piece of software. How do you run it differently? Well, I mean, like let's say a strategy is sandwiching. Sandwiching is one MEV strategy. I think most people They know that one. Yeah.

Right. Somebody makes a trade and you want to make sure that the trade will go off at the worst possible price for the person who does the trade because you'll be able to skim off some of that kind of diff. Mhm. So, that's one strategy. Liquidations could be another strategy, but there's more like granular levels, too. Like, how do you go about doing that sandwiching? You could be like, let's say a sub strategy could be like sandwich the sandwitcher. Oh wow. Right? That starts to sound more fair than just the sandwich strategy. Yeah.

Yeah, yeah, exactly. Well, there's always a bigger fish in the mempool. Um so, that's kind of what what I mean by strategies. Like, but how sophisticated are they? Like, are they sandwich the sandwitcher? Or are they like Are you altering parameters? Are like I'm just kind of picturing like what do you actually do to even to make these strategies happen? What are you doing? You could kind of try to imagine it like a decision tree or something like that. Like uh a mem- mempool transaction came in that you've never seen before. So, like what do you do with it? So, some parts of it you start to like analyze.

Let's say like some sanity checks. Preliminary stuff. And then you can start deciding what do you want to do with this transaction. Do you want to run this transaction? Do you want to predict what this transaction's impact on prices will be? Mhm. Do you want to know where most likely this transaction will end up in the block at what index? It kind of like goes deeper and deeper on each one of those kind of sub decisions. And are you just writing scripts that like are are taking all of these decisions into account? Like you're kind of like if it's higher than this or if it falls into this threshold then we're going to pay attention to it. If it does that then we're going to do this. If it has Like I'm just kind of Yeah.

Is it something like that? Yeah, exactly. Not scripts though. We use compiled languages, but yeah. Sure. I'm not a coder. I was just saying whatever. But that's exactly it, basically. I guess then going back to that other question like how sophisticated does it get? Like how big are these strat- like these trees of decision-making? Is it usually like you're creating multiple very short decision trees or are you doing like very elaborate Yeah, I'm kind of trying to get a sense for for what this looks like. It can get complicated because sometimes there's also a component of are you trying to front run or are you trying to back run?

In one of the two, let's say time doesn't matter. Mhm. If you're trying to front run someone, time isn't that difficult because you could always do a higher gas price. Mhm. But if you're trying to back run someone, the speed of your execution is is super critical. Right? If you take too long, then you won't land it. Back running is when you want to match the gas price. Got it. Yeah, yeah. How do you do that? How would that work actually? The back running. It's it's a issue of speed. It's like There's like some network overhead and then randomness and then just how fast does your code respond to the new transaction that you've just seen for the first time.

Mhm. Okay, so you're creating these strategies and I think you said this just before Dean, like you have a team of like five or so on doing this right now? Yeah. Are you all individually doing it yourselves or is this something like a like is this a real collaborative effort where you're like sharing insights and picking up new kind of tricks that maybe you've seen from the outside or one of your colleagues? It it's fully collaborative at this point because no one really has the incentive to do it themselves. Okay.

Um especially with the strategies that we're doing, you're always drinking from the firehose, right? One of the reasons why we were able to put together this team is because everyone who's in it has burnt out before. Um I think we have a team of probably the best people in MEV um who were probably making the most money. The problem is you're constantly changing stuff. You're constantly fixing stuff. You're constantly like modifying code to ensure that it's fast enough. And so you're never not working, right? Like our Telegram chat which we have together is 24/7 someone is like updating code and writing in it that they've changed something.

Is it a single repo then? Is it like you're all contributing to the same thing? It's not that you're individually creating your own strategies. It's like this is one joint effort. No, we have a parallel effort because a lot of MEV is creativity. And having that kind of creativity, you need to give some kind of freedom. Mhm. Uh and so we actually have three parallel bots. Basically in in some state of production. That are always being tuned and tweaked. Because sometimes we might want to scratch an itch on some idea.

Right? Adjusting a bot is not like that easy usually. Uh you know, moving things around. It's very hard to know if you move something around and it doesn't cause a mistake later on. Mhm. So, we have multiple approaches. And this way people can maintain like this kind of new blood of creativity always. But also that we can share and funnel pass through ideas that um are maybe aren't good ideas. Because like once you get a MEV idea, you have to build it. Mhm. And that takes time. Right? And I mean, it's a fun building this idea. I'm not building another one.

I don't want to tie down a teammate on that. So, we we have a it's like it's very collegial. It's not hierarchical. It's like at the sweet spot of a low team count. Yeah. That everybody just genuinely loves it. And uh we're just kind of like spitballing ideas with each other. I think small teams are always better. That's my my theory. But uh are you when you're building this thing, like where does the capital come from? Like are you cuz you sort of said this before like each of you could be earning a lot on your own. So, is this capital that's contributed by the participants and creators of the strategies or is this like you have existing capital that you can do things with through this? I'm kind of I'm kind

of trying to picture like cuz to take advantage of MEV, you are also often trading something or you're doing something with funds as well. So yeah, I'm just curious where that comes from. We did a small round when we started the company, um which was led by Dialectic essentially. And that's where the capital comes from. A lot of the strategies we're doing actually don't require that much upfront capital because we use a lot of flash loans. Oh yeah. Let's talk about that. So this is flash loans MEV. But why don't we define for the audience if they haven't heard about it before what a flash loan is.

So a flash loan is a loan you take out for the lifespan of a transaction. Mhm. So I take out this loan and then I pay it back at the end of the transaction. And so what that means is for arbitrage, what we can do is we can borrow millions of dollars and pay back the loan at the end and we just take the delta of the profit that was made minus the cost of the loan. Wow. Had flash loans not been built yet, then you would have had to have a lot more upfront capital to actually see this happen, to see any profit, eh? Yeah.

So MEV plus flash loans equals a business that can be founded with very little capital and can still already start to take advantage of these strategies. Yeah, pretty much the only upfront capital which you would need is the cost of a transaction. What are gas fees on something like this? Is it Are they crazy? Like are you doing a lot? Some of our strategies inevitably run hundreds of thousands of transactions. Um but it's on like a lot of what we're doing is on BSC right now and so it's throwing pennies at the wall.

And that's why that's why we can do so many transactions, right? Because it's like it's essentially going to a casino and just throwing coins into the slot machine and hoping something comes out. That's the analogy Edgar and I often use when it comes to MEV on BSC. Actually, that was another question. So, you've decided like you're doing most of it on BSC, the Binance Smart Chain. Do you do other chains? Like MEV as I've always understood it kind of starts from Ethereum. That's like where a lot of the initial research you know, the fact that it's a Geth, that Flashbots is a Geth. It's not a fork, is it? It's a Geth version? What is it?

It's a It's a fork of Geth, yeah. It is a fork of Geth. Okay, so it's a fork of Geth. Yeah. Yeah, I'm just like are you looking at other chains as well? Are you all Are you just BSC and Ethereum? Uh so, we don't do Ethereum right now. Um we're on BSC and we're looking at pretty much every other chain as well. Ev- anything which is EVM compatible we're looking at. We're also slowly starting to venture out into non-EVM compatible things. Really? Like what? A lot of Osmosis and Cosmos stuff we're starting to look at. Cool.

Yeah. The The reason why we don't do Ethereum itself is because the cost of a failing transaction is much higher, right? Um for some of these like arbitrage transactions and things like that, you could be out thousands of dollars on a failed transaction. Do you When you're doing some of this sort of strategic research, do you ever look at ways Cuz like I mean originally when the MEV conversation started, it was presented as a problem, right? Like this is a danger, this is a bad thing in a way. We should try to prevent more MEV. We should try to add new strategies. Like, you know, we had actually had the Osmosis guys on talking about using threshold cryptography to in order to actually make some of

these attacks impossible. At the same time, there was this attitude of like instead of making it impossible, let's democratize it and allow anyone to easily run this software and make use of it so that it sort of levels the playing field. But in the work you're doing, you're creating strategies in in to kind of access the MEV, to use it, to to get it, to I guess that's what it's called, to like capture it. But, are you also looking at ways to mitigate it? Like, are you interested in how networks could potentially like reduce some of these attacks?

Yeah, so we we work with some teams or we're starting to work with teams to look at how to prevent MEV. When we actually started Block, Edgar or and I had a long philosophical debate on how we view these things, and I think the the way we both look at it at this point is from an accelerationist standpoint. Someone has to do it to point out all the gaping flaws, and if it isn't us doing it, someone else is going to do it. Mhm. Um and I'd rather Edgar and I do it because we, as said, actually are interested in preventing these things or um ensuring that chains are a bit more resistant to some of these MEV exploits. Like, uh we

keep a relationship with the Flashbots guys. Um I'm constantly talking to Phil Daian along with other people as well on especially any any project that claims to be MEV resistant, I'm always happy to look at and actually give feedback. What kind of feedback could you imagine actually giving? Not feedback to the uh chains that claim to not have to be MEV resistant, but rather like as you're doing this work and you're finding kind of like maybe ways that it could be mitigated, how do you actually communicate that? And actually, I guess by by this I also mean like are you open-sourcing all of the strategies eventually? Like, are you publishing, "Hey, we figured out how to do this and this, so you should so maybe the chain should be aware of it."?

Um I think that we will start open-sourcing strategies once they become unprofitable. Mhm. If we open-source them now, it's an alpha leak, and inevitably the bottom line of Block is still to make money. Okay. And we can't maintain that status quo if we start open-sourcing the strategies. Got it. Edgar, I want to touch on the topic that you raised kind of when we talked a little bit about your background. This idea of a reorg. So, from where I was sitting, I saw a little bit of this on Twitter, but I wasn't following it super closely, to be honest.

What I did see was that there was like a a proposal to kind of go past the way that MEV had been considered before, which was like the sandwiching attacks and stuff like that, more into the reorg strategy, I guess. And then there was a lot of pushback. Maybe you can tell me a little bit more about like what it is. What is a reorg? What does that even mean? And what happened? The root of the problem is that in Ethereum, when you get a block, as it is right now in proof of work, once you get the block confirmed, it's not really the block.

That is to say, the update that has been done, all the state updates that have been done in that block, you can't really rely on them. Mhm. But we all kind of do rely on them. Everybody likes to have fast transactions, right? Like um people in the ETH world don't really like the 15-minute or the 20-minute Bitcoin block time. Mhm. It's just not usable for smart contracts. So, to get the faster block time, there some compromises have been made, let's say. Because like computer science is all about tradeoffs.

Mhm. And one of those tradeoffs is that after one block in proof of work Ethereum is is accepted, it's not really reliable state yet. Like the data inside of the head of the blockchain is not reliable yet. You have to wait a few blocks. And so reorg kind of exploits this, or tries to exploit this, by letting people pay to redo the last, let's say, two blocks or the last three blocks.

So, reorgs naturally happen, but the idea was to let people pay for it. Is that something that is happening actually all the time? Reorgs happen all the time naturally in uh Ethereum, in proof of work Ethereum. Do Do reorgs actually happen on other systems? Like, is that something that is just common to all proof of work blockchains? I mean, you would you'd sort of said like Bitcoin doesn't potentially cuz so but it just makes it very, very slow. No, Bitcoin also had reorgs. It's just that with the longer period of time that you wait, the less likely someone can pull that off, rather less likely a reorg will happen.

Mhm. The reorg issue is because it's like it just physically takes time for a block to propagate around the world to a sufficient number of the uh hash production power or depending on the chain, maybe it's validator power, something like that. So, because of that, you have this issue that where sometimes you might have competing versions of reality. And that also happens on BSC, which is a proof of authority network, but they also have reorgs there, too. So, tell me a little bit about how you would create something or some sort of piece of software that would actually like use that. Like, so far what it sounds like is kind of this organic around the world blockchain thing based on timing and like where maybe the source of the current chain is and how

quickly it's moving. Like, that sounds very kind of like I don't know, something that's just going to happen anyways. What could you do to actually manipulate that? So, I I worked on basically the MEV Geth, how MEV Geth manipulated the production of the block, I just extend worked on extending that part. To where there would be, let's say, a third transaction pool which takes proposals to do a reorg. Ah. When the guest client when a maybe guest runs code that makes the block, it basically has like a conditional or something that will say, "Hey, will I make more money from redoing the last three blocks or will I make more money

continuing as I was?" May- maybe just can you explain like how multiple clients would actually respond to a call like this? Like do they all agree to it? Is it something that's proposed and then accepted or is this something that's automatic? So that part required a little bit like more extra coding because I ran into some issues of having to create some kind of coordination between the miners because they would have to build on my fork of the chain. Like if I created a re-org, Yeah. Ethereum is like it decides which chain to pick by being like the greediest chain. It's like the heaviest chain.

Yeah, the longest. So I need other people to keep building on my chain. So if if I can get the next miner to also keep building on my chain, then I need to coordinate that kind of stuff. Like how am I going to get that message across to them? So I started to have to extend like the P2P protocol just slightly, just like little things like that. Mhm. But there was like an issue. I'd have to go like backwards in time to then promise more money than what a competitor could have done already future in time kind of. So it became a little bit of a mind bender.

And is this something that you actually respect out completely or was this just like some space of research that you were starting on? No, I got quite a few commits in and that's when I I ran into massive pushback around the same time that I also had the technical problem of how to communicate to the next miner to keep building on my re-org chain because I wasn't going to start out with like Okay. all the hash power, right? Although I did have some miners agree to run my code, it wasn't going to be enough hash power to do it on demand whenever I wanted. So, I needed to have some kind of way to have other miners join the collusion.

And you didn't get you didn't solve that point when the pushback had had started though. The pushback had started like immediately. It's just that like when I got to that technical problem, the pushback was like at its zenith. Yeah. And so, what kind of happened? You got people started to get really angry that you were going in this direction basically. They were angry that you were incorporating this. It sounded like people felt like it was kind of heresy. Like you should never suggest such a thing. That it was kind of against Ethereum or something. They got all bent out of shape and nice people became mean and people who said that they'd be friends just disappeared.

It was just surreal because to me it's still like some open source project. But I mean, obviously it's not. People have put in so much money into the Ethereum world. And truth be told, if reorg was successful then the smart contract adapt experience that you have would definitely suffer. Like you would have to wait for quite a bit of time. And probably no Uniswap trade would succeed. Um you know, everything would get undone and then redone and then undone and it would just it would be a race to the bottom.

So, in the end did you decide to abandon that work? Did you walk away from it? Or did you evolve it into sort of a different direction? So, Vitalik wrote me a telegram message that changed my mind. That and a few friends from flashbots. But he basically was saying it would be really crappy to poop all over the work that the smart contract adapt devs have been working on for the past two three years because they truly believed in blockchain. Mhm. And to like really like crap all over their effort because like the foundational platform that they use would not be usable. It would be really mean. It would just It

would just be like It would be friendly fire, I think, because I do believe in blockchain. I want it to succeed. It's strange that it came down to like one contributor's decision to go in a direction cuz like so you decided you don't want to do it or you you would potentially like change course, but like could somebody else do it? What if there was somebody who's just malicious towards Ethereum? Could they not develop something like this? That always kind of like struck me weird. Why is it that something I did, I'm just like nobody, causes so much like issue?

Because the blockchain has been around for 5 years, 5, 6 years Ethereum. Somebody could have done it. And while I was researching it or in my prep work for it, I found some material from like years ago, like literally like year or 1 or 2 years after Ethereum came out, talking exactly about this issue. And then I had a few people DM me on Telegram telling me that when they were close to the EF, they mentioned these issues. I don't know. I think when people get threatened, it's better to put your head in the sand and not address it, especially if there's like such a euphoria in some like collective vision.

You know, it's like nobody wants to hear bad news. Yeah. Um Like that's just one aspect. I think the other aspect is like it's not easy to code up. Okay. It's very difficult to code up correctly because if you mess it up, then the miner has lost money. Mhm. And the miners are risk averse. You know, flashbots kind of normalized the miners running forked code. But before that, they were super wary of something like that. Yeah, yeah, yeah. And they still are super wary. Okay. are super wary. And they should be because it's like it's a lot of revenue that could be gone.

Mhm. And how do you conclusively prove the counterfactual that I would have made more money or less money if I didn't run the flashbots geth? It's still kind of like it's not necessarily 100% known. So, the miners are risk-averse. So, even if you could code it up, how would you code it up? As an anon? Okay, so like this anon GitHub profile is going to be run by like the miners with like hundreds of millions of dollars of infra? Yeah. Or what would it be? Like you then have to work for the miners? It's kind of like also a little bit icky. So, it's like there's a lot of factors. And also you probably need to have like some kind of Chinese connect.

That it's kind of fascinating the fact that like the social pressure in a way does keep it's like reputation and social pressure actually in despite it being very very open source and it sounds like anyone could come along and do it. There's this sense that like it would need to be trusted and for it to be trusted there probably would need to be a reputation. To build up that reputation, you'd be in that community and potentially care about what what the community thinks of you. But do you see this kind of as an attack vector in the future? Like what if there is a malicious actor who is anon or pretends to not be anon? Maybe there is a name and they seem like a real thing and they befriend everybody, join a

team, build this thing. Like I'm just trying to think is there is there a way cuz like the the idea doesn't really go away. The idea that you could do this. The idea was there from day one. Yeah. As soon as the protocol that Ethereum proof of work uses was basically like immediately pointed out there are these kinds of like time banded or these kinds of like reorg or these kinds of issues. Mhm. Um and so it's this overlap, this intersection of like the technical skill the reputation the ability to like just get in contact with the miners. The miners are people, not robots. You know, and not all of them have like a email address that you could just send, right? And then like

you have to validate that your thing works and conclusively prove that it make more money. It's not easy to intersect all those things. Yeah. Yeah, the backstop here is truly the miners, right? Um I think that even if someone develops a reorg client, most miners probably wouldn't run it because it is a race to the bottom. But don't you think they might run it like in the short term? Like wouldn't some of them try if it does make them short-term more money or they get some sort of they have a strategy that they could use for certain things? Like it does seem like very like forbidden fruit or something. Like people would want to do it a little.

The problem for a reorg client is you need buy-in from a lot of miners, right? The MEV gets one miner can run it and prove that they're making money because they're just determining how the block that they're currently mining is being set up. With MEV or with a reorg client, you need buy-in from multiple miners because as I said, the miner in front of you needs to build upon your chain. I think basically only Ethermine can get away with running it on their own. Ethermine being the mining pool. The lar- Is that the largest mining pool?

You You need basically like a third. You need to get to like a third of mining pool power to pull off some of these attacks. But there are groups that actually control a third. I think Ethermine is at like 32% or 31. I mean, they hover around there. That's the actual amount you would need is like 30%? For some of like the uncle bandit attacks stuff, yeah. 33. Interesting. Why don't they do it? Uh that's a good question. I think it is a recognition that it would probably hurt the chain. Mhm.

I think there was probably many legal issues that would open up. I mean, when I was pursuing reorg, I got many, many legal threats and legal opinions, right? Especially if, let's say I do the reorg and then someone pays to redo a liquidation for someone else. And it's like a $40 million liquidation. That's like whoever had $40 million has enough money to litigate. Yeah definitely. You know, that's true. So, it's not entirely a slam dunk code as well thing.

Mhm. So, since then though, like you've and I guess this is a question, have you shelved that project then? Have you like put it to bed and now the type of strategies you're looking at and the work you're doing is so is quite different? Uh yeah, I shelved that idea out of respect for people who gave me the alternative opinions. I shelved it. Mhm. Was it hard to do? To let go? Yeah. Um well, it was because I I I left Flashbots over it. And Flashbots is an amazing team with the best people in this space.

Um I can only say good things about each and every one of the people there. Mhm. So, you know, walking away from that team for an idea that didn't pan out is a little bittersweet. That's fair. But you found a new team. I did. And that came a few months after anyway. Got it. Okay, working on other EVM compatible chains kind of makes sense. I feel like a lot of them are also running some sort of geth fork or some sort of like equivalent thing that would be easy to run these strategies on. But when you talk about working on something like Osmosis, these other chains where the the whole system is very different, they have different prevention mechanisms

even for MEV or at least planned prevention mechanisms. How do you approach these? Like would you see yourselves building out this this client infrastructure again or doing the fork yourselves with the add-ons or would you see are you waiting for like another team to build that and then use them for your strategies? Well, so even on BSC we have to build our own forks right now. Okay. Um so not only are we building strategies, we we build a lot of infrastructure. Cuz MEV geth doesn't exist on BSC and so we're in the for to extract MEV on on BSC, we're essentially in the normal mempool, which is why timing becomes so important because we're we're just fighting the normal

mempool. All of a sudden I'm confused about how you can do this on BSC because like if it's a proof of what is it? It's not proof of work, it's proof of authority. So it it's like a limited set of miners. Are you in it? Have you been accepted? So like how how do you actually do anything here? Like you could create a geth client but they'd have to run it. We need to really efficiently send transactions to the validators, which means that we need to try and figure out where validators are and things like that so that we can we can reduce the time span from which we see a transaction to getting our response to a validator. And so it's an interesting game cuz it's micro optimizations, right? We're optimizing for the

nanosecond there. Cuz we just need to get transactions onto the network as fast as possible. And where do you live in that game? And actually this actually brings up a whole question on the strategies in general because I I got the impression like if you're utilizing a strategy, you're acting as a miner or you're acting as a validator. But is that not the case? Like the MEV extraction can be can be happening by an agent that is not a miner at all but lives in between the mempool and the miner. We're searchers. Searchers have strategies. For for miners, the strategy is really just run just run the software. But for us we have to custom code all kinds of software for the custom strategy.

Ah, yeah. And this is the bot. And minor or validators running the software, then you create the bot and you call this a searcher. I actually hadn't heard that term before. Yeah, we run a bot, yeah. And the bot is called you call the bot a searcher? Uh I guess you could say it that, yeah. I mean, searcher is like the entity that's like doing the looking for MEV and trying to claim it. So, yeah, I guess so. Okay, got it. So, on in the BSC side, like you still need the validators to actually be running a particular kind of software so that you can utilize these. No, they run regular BSC. That's why speed is so important because we don't have this privileged private mempool.

I see. I see. So, they don't have the MEV geth, but there is there has always been MEV anyway. So, you're kind of playing like the older game, right? Yeah, we're playing the pre-flashbots MEV game. Yeah, yeah. Are you using actually strategies like that like that used to be used in that case? Not really because the game has become too advanced. I mean, some stuff stays the same like the concept of front running or or rather the concept of like PGAs or back running. Some of those things stay the same, but some of them are like less important. Like PGAs aren't as big on BSC as they used to be on ETH before flashbots.

So, I I think this has been a really cool kind of journey through the I I realized in this conversation that I I understood what what the research work on MEV was roughly, but I didn't actually know as much on how it would be to like like use it, what the games being played were, who the people what they're doing, how they're actually using this thing. So, I think it's been really cool to kind of go through this journey with you both. I have some other questions that are less related to the MEV work and more related just kind of to general kind of actions I've seen going on maybe more in the dialectic direction. So, Dean, I I wanted to ask you a question.

Maybe you could help like update me a little bit on Sushi. The reason I'm going to ask you that is I know that there was like a proposal that you were part of I guess like a few months ago. What's going on with that project and are you related to it? Were you working on it? Was there some connection there? So Ryan and I have been close to the Sushi team since the beginning. When there was that rug pull that initially happened, we had reached out to Maki back then already and offered our support. And so literally 2 hours after after that rug pull happened, we were on a call with Maki to like discuss next strategies and things like that cuz we were we were very bullish on the idea of Sushi. And so we've always been close to the team.

Obviously Joseph Delong who was the CTO for a while is a is a very close friend who I worked with prior to Dialectic already while I was working on ETH 2 clients and things like that. Um so we've always been close to the team. One of the flaws which you're starting to see in Sushi is one of the issues that I wrote about I think in 2017 2018 which is that DAOs without any hierarchy I don't think they work, right? Um and the problem that happens is that there's no accountability.

Um no one feels responsible for anything. No one's truly in charge. So it's hard to delegate and things like that, right? And and these flaws were starting to to show up in Sushi, right? Um a lot of the core developers are very unhappy with what's going on. They're not being compensated fairly and things like that. And the proposal was actually twofold. First, we wanted to restructure the DAO and create a hierarchical DAO with different sub DAOs which are for the various projects that Sushi does and for like different roles uh along with changing the way the compensation structure works.

What actually happened in Sushi? I think a lot of people from who maybe weren't closely monitoring the chats, like are were kind of surprised when Joseph stepped away in general. And then your proposal kind of came after that. But what did what did actually happen? I got a DM about 4 months before that, which to the was to the tone of we're all going to quit unless this gets fixed. Ooh. We were taking those 4 months to try and figure things out and organize the DAO, which was quite a challenge. We were hoping that people don't leave in that time frame, and I was on calls every day to convince people not to end up quitting. Sadly, that ended up happening. I think

Sushi, if we've learned anything, will probably survive. It's just now a harder challenge to organize it completely, but I think Sushi is going to come back stronger than ever. But are you still involved? Um not as much as before. I don't know the new or the the current leadership as much as I knew the other one. Mhm. And I haven't had much time recently. I'm I'm hoping to get involved again soon. I think Sushi is a is an incredible experiment Totally.

Um when it comes to like some of the crypto ideals. And so just for that sake, I I love being involved. It sounds like I mean this is we know that this is the place where a lot of experiments are happening. This one's DAOs, MEV. Actually, are there any other topics or spaces that you're exploring right now? Maybe within uh project Blank or within Dialectic, other kind of paradigms that you think would be a great place to start experimenting? Uh I I mean we're doing a lot of interesting things at Dialectic. I think one of the things that amazes a lot of people about our team is how developer heavy the team is.

Um so like one of the things we've built internally is a completely automated uh stable income fund. Ryan and I, when we started Dialectic, were heavy yield farmers. At one point that obviously didn't scale anymore. It it made sense when we were doing it. We hired people to do it and it still doesn't scale. And so one of the things we've we've done which we're hiring now for as well to continue is this automated stable income fund where we deploy into specific strategies. There's bots that harvest when gas prices are good or where profits are high enough.

In the future the bots will be able to automatically rebalance depending on the API of specific strategies and things like that. So we're we're incredibly developer heavy on the things that we do. Do you are you playing in NFT land or you avoiding it? Dialectic is a heavy player in the NFT land. I think we're one of the biggest investors in play to earn as well as hold probably one of the biggest like fine art NFT portfolios. What is your opinion on NFTs? I see Dean has opinions.

You're going to get me in trouble Anna. No, I don't get in trouble. I say what I want. I think gaming NFTs are interesting. I find gaming NFTs very interesting. I think the utility is becoming something quite cool especially with some of the gaming investments that we've we've done. I think once we get to an interoperability stage which I'm hoping we see soon, gaming NFTs will be even more proven. I used to think gaming NFTs were stupid until like I sat down with someone working on them and one of the things we had discussed was like imagine you're a World of Warcraft gamer and I know you used to play World of Warcraft as well.

Me? And you have one of I let No way. Didn't you play World of Warcraft? No way. I thought you did. No way. I thought we had this discussion. Okay, you played other nerdy games which count as well. Civilization man. Okay, well just imagine like you know, you have some item that you need to complete a dungeon or something and you haven't grinded to to get that item, but you can now rent it from someone who's lending it on an open market, for example, right? Like, that becomes really compelling when you can start generating yield on the items that you've earned in games, which kind of already exists on the Steam marketplace, but it is within their closed system.

They take a 30% commission and and the financial products that you can build upon that are not very interesting, right? On Steam, you can buy and sell. That's it. You can't like lend out an item and things like that. And so, I think that's where it becomes interesting. When it comes to NFTs as art, I see the appeal. I'm not a holder of any. I bought one NFT, which was the first ever fashion NFT, and the reason why I bought that was purely an investment thesis, which is this is the first ever NFT by a fashion house. It's probably going to go up in value, and so I bought that.

Have you sold it? No, not yet. Okay. I'm still holding. I think it's soon Now, with like Fendi getting into Ledger accessories and things like that, I think we're going to start seeing the top for that NFT. And what about you, Edgar? Do you care? Do you pay attention to this or is this like super off your radar? It's on my radar because of Twitter. But it's um I try to stay hyper-focused on something. So, I've kind of been on the deep end on work lately.

Do you see MEV developing in a direction? I'm kind of thinking like going forward, you know, right now we're talking strategies, you're talking about different networks, but do you see this role, this kind of strategy, this type of work you're doing, do you see it sort of developing into something else? I think it will stay. I think it will evolve. I think it's going to get more nuanced and complicated as real L2s come online, as inter-chain things become more reliable and robust, uh it's going to get even more creative and complicated.

Once you bring on these other chains, is there actually like MEV arbitrage between different chains? Maybe this is completely silly to even say, but like is that something that it exists? Yeah, that's totally something, right? I mean, it's just when I do arbitrage between two pairs that are out of whack, the fact that the two pairs are on the same chain is kind of like implicit. It doesn't matter, right? It doesn't have to be on the same chain. They could be on different chains. It just makes that like all that extra code that has to deal with that nuance become more nuanced and more complicated.

Interesting. Cool. Well, I guess it sounds like a space to continue to watch. I think you're right that like as L2s and all of the multi-chain communication kind of comes online, it's going to be total wild west for a bit longer at least. Is there an end game to MEV? Like is there sort of a point where all the strategies are filled out and like there's sort of a standard? I mean, it's like an equilibrium. Like the equilibrium for MEV is when all prices in the markets are balanced, let's say. Like there is no arbitrage. There's no DEX arbitrage. All the prices are perfectly balanced.

It's kind of like that uh econ market efficiency theory or something like that. Like it's trying to trend towards that, but not all MEV is like DEX MEV. Some other MEV could be like a liquidation that like nobody has put in the trouble to code up against because like some weirdo protocol. Another MEV is like you have a a smart contract that exposes a way to talk to the rest of the world in a way that it wasn't intended to and nobody has claimed it, right? Like that's not going to trend towards equilibrium unless someone finds it. It depends on where and what subset of MEV you're targeting.

Some of them trend towards equilibrium, some don't. Cool. Well, I want to say a big thank you to both of you for coming on the show and sharing with me the work you're doing, letting me ask all my dumb questions about MEV. Really appreciate that. You can tell it's not a thing I've studied much, but I really do love kind of checking in on how this is going, what's happening and hopefully what's coming up. Thanks for having us, Anna. Thanks, Anna. I used to listen to the ZK podcast on 1.5 coming back from Mountain View to San San Francisco on my commute.

Really? Yeah, it's a real pleasure to be on the show with you now. Cool. Do you know you're going to say something nice? Uh thanks, Anna. Hey, I listen to your podcast. All right. Otherwise, I wouldn't have known that Oscar was on. True true true. I'm just kidding. Thank you, Edgar, for the for that nice thing you said and thanks for coming on the show, Dean. I know it's been a long time coming. We've talked about it for I think a few years and so it's really cool to have you on. So, thanks again to both of you. I also want to say thank you to the podcast producer, Tanya, the podcast editor, Henrik, and to our listeners. Thanks for listening.