Bankless - 66 - Crypto’s Existential Threat | MEV Panel: Phil Daian, Georgios Konstanopolus, Charlie Noyes

Episode Date: May 24, 2021

Phil Daian is a researcher and software engineer and is leading the charge into mapping MEV with the Flashbots Project. Georgios Konstantopoulos and Charlie Noyes are both partners at Paradigm. ------ 🚀 SUBSCRIBE TO NEWSLETTER: https://newsletter.banklesshq.com/  🎙️ SUBSCRIBE TO PODCAST: http://podcast.banklesshq.com/  🎖 CLAIM YOUR BADGE: https://newsletter.banklesshq.com/p/-guide-2-using-the-bankless-badge  🎧 Get this Episode's Debrief: https://shows.banklesshq.com/p/exclusive-debrief-cryptos-existential  ------ BANKLESS SPONSOR TOOLS: 💰 GEMINI | FIAT & CRYPTO EXCHANGE https://bankless.cc/go-gemini  🔀 BALANCER | EXCHANGE & POOL ASSETS https://bankless.cc/balancer  👻 AAVE | LEND & BORROW ASSETS https://bankless.cc/aave  🦄 UNISWAP | DECENTRALIZED FUNDING http://bankless.cc/uniswap  ------ Bankless Podcast #66: MEV Panel Guests: Charlie Noyes, Phil Daian, & Georgios Konstantopolous In this State of the Nation, we bring on big-brain experts behind projects like Paradigm and Flashbots to explain the complex but critical topic of MEV - commonly known as Miner Extractable Value, but as discussed in this episode, Maximum Extractable Value is a more appropriate descriptor. MEV is arguably the single most important unsolved problem in the crypto space. In the context of incentives and network security, MEV generally refers to the ability of miners & validators to take advantage of their position as securers of the network. Blocks are verified by single entities, and these block producers have tremendous power when they are the ones adding a particular block to the blockchain. We refer to this instance as ‘God-Mode,’ in which a node has the power to assemble & order transactions at will. The threat here is the potential gaps between what is best for the network and what is best for the node. As an internal and unbounded threat to blockchains, solving the problems that MEV presents is critical to maintaining proper consensus and ensuring decentralization and permissionless access. We take a deep dive into potential solutions, and the impact of EIP-1559 and Proof-of-Stake on MEV. Dive in to learn why despite the challenges ahead, Charlie says: “In the absence of certainty, I’m willing to take the optimistic perspective.” ------ Resources: Phil on Twitter https://twitter.com/phildaian?s=20  Georgios on Twitter https://twitter.com/gakonst?s=20  Charlie on Twitter https://twitter.com/_charlienoyes?s=20  Exploring Ethereum’s Dark Forest with Dan Robinson https://youtu.be/MtpRDqrCfbc  Investing in DeFi Paradigms with Charlie Noyes https://youtu.be/dQOhbb5HmLA  ----- Not financial or tax advice. This channel is strictly educational and is not investment advice or a solicitation to buy or sell any assets or to make any financial decisions. This video is not tax advice. Talk to your accountant. Do your own research. Disclosure. From time-to-time I may add links in this newsletter to products I use. I may receive commission if you make a purchase through one of these links. Additionally, the Bankless writers hold crypto assets. See our investment disclosures here: https://newsletter.banklesshq.com/p/bankless-disclosures 

Welcome to bankless where we explore the frontier of internet money and internet finance. This is how to get started, how to get better, and how to front run the opportunity. I'm Ryan Sean Adams. I'm here with David Hoffman, and we're here to help you become more bankless. David, this was one of those mind-blowing episodes on bankless. I think a canonical episode, who did we have on? What did we talk about today? Yeah, a panel of three M-EV experts. We had Phil, Diane, Charlie Noyes, and Georgios Constantinopoulos, who are really tackling, I think, probably the biggest problem in crypto head on. The problem of M-E-V or maximally extractable value, also known as minor extractable value,

is the one single existential problem that could end this whole entire crypto experiment. It could just really either just completely just kneecap why we find crypto so valuable or also make the whole thing just not work at all. And so those are some pretty dire, you know, characterizations of what the MEV problem is. And that's what these MEV experts really illustrate this problem as. And really the TLDR of what MEV is is that Bitcoin, Ethereum, these are leaderless systems. But the way that they are leaderless systems is that they just allow people to take turns. being the one single leader. And that leader is the person that proposes the block, right?

At some point, while there's a bigillion Bitcoin miners, and in future in Ethereum, there's a bajillion staking validators, only one person at time proposes a block. And so at that moment in time, that one person is, or entity is separate from the rest of the network. And that one person has the power to order transactions that people have signed as they see fit, which gives them basically God mode over one single instance.

instance of time. And that really changes the incentives that the block proposer has in relation to the rest of the system. And this could be, it could be good in some circumstances. It also could be massively devastating. And what these MEV researchers are really trying to do is to try to figure out how do we harness this misaligned incentives opportunity and corral it back into alignment with the rest of the network and the rest of the values of this ecosystem, which is decentralized, decentralized consensus. That's a TLDR of this, the whole entire episode, and it's really the big problem of crypto. Guys, I just want to emphasize once again what David just said. This is the problem that

could sink all of crypto. You know, David, so often, like, we hear mainstream media talk about all of these things that they perceive as threats to crypto, and they're really, like, not threats to crypto at all. Like China FUD, China FUD, you know, nation state regulatory fud, tether, it'll never work. It's too volatile. Like all of these things are such surface layer attacks. And if they were really looking for the problems of crypto that need to be solved, they would be talking about MEV.

But mainstream is not knowledgeable enough to talk about it. These people are, this is the panel that's on the front lines of not. just researching it as a hypothetical, but actively trying to solve M-E-V-type problems. So I think this is going to be like a canonical episode. It's not going to answer all of the questions about MEV, but I see this, David, as like maybe the first in a series of bankless episodes that we do, because we're going to have to talk about this often on bankless. It's not a problem that's going to go away soon. It's a pernicious problem. We use this analogy of it's kind of like squeezing the balloon, right?

Like you squeeze it a little bit in one place and just bulges out in another place. So one minute we think we have it cornered and we really don't. The other interesting thing about this episode, David, is like all of the places that MEV actually exists in the traditional world that aren't visible, right? So like we talked about the GameStop fiasco. GameStop and Robin Hood and the hedge fund manager that called Robin Hood and essentially was a call to say, hey, reorder our transactions. Hey, censor some transactions for a period of time. So, MV exists as a problem outside of crypto. It's not a crypto-specific problem, but in crypto networks,

it's much more visible. And again, the gold standard here of what we're trying to build is a credibly neutral, open financial system for the world. It's never been done before. The existing financial system is not that, is none of those things. things, but we are holding ourselves to a higher standard in crypto. So we want to make sure that this problem gets solved. And I think we ended with like, you, you asked the question of like, how optimistic are you guys about actually being able to solve this problem? And the panel all had really interesting answers to that question as well. Presidency at the very least. Yeah. Stay tuned to the end to listen like to how they answered that

question. But like we talked about measurement of MEV, where it's most pernicious. Phil talked about his flashbots project, which is just like a really cool attempt to take on the MEV problem head on. We talked about this transformation to ETH II, what it means when miners actually turn into validators and maybe there are some incentive changes. Will it really make a material difference to MEV or not? The panel answers all of that. And we get into like some philosophy at the end, which is, I didn't know we were going there, but super cool. It definitely fits the bankless theme. So fantastic episode, guys.

You're going to love it. Yeah, the topic of MEV is such a broad one that I really think the best way that you and I, Ryan, can really help out this problem is just by talking about it more. This is actually the first time that Charlie Noyes, Phil Diane, and Georgios have all been on the same panel. So they were actually pretty hyped to all be in the same spot to talk about it. And there's a number of other people that are definitely one. worth getting on to talk about this.

Like, Hazu comes to mind. I know Hazu pays a lot of attention to MEV. And also, Carl Floresh comes to mind. And you talked about how we ended up on conversation on philosophy. These three panelists are very pragmatic and very concrete. But MEV, like you said, it's not just crypto. It's something much more existential. It goes into just the nature of the universe and how order of operations happens.

As soon as you have just like bookkeeping or even, grander just like ordering problems all of a sudden like topics of who has the rights or power to determine the ordering of things and that's a very existential question and i know that um you know these panelists are very pragmatic and concrete carl florsh i know can take the concept of miv and like talk and connect it all the way to something as like crazy as like human consciousness and so i want to get carl florsher on and also have the same exact m evi conversation uh in that kind of context. And just like you said, there's so many things to talk about. And I think the best thing that we can do at Bankless is just host all of the conversations because the best thing

that we can do to solve a problem is illuminate it. And I think that's what we are trying to do here on this podcast. Absolutely. Guys, you're going to love this podcast. Of course, David and I have some further thoughts for you in the debrief. That is available for premium subscribers. So if you're a premium subscriber, you should have access to that pretty soon as well. With that, we're going to going to talk about the sponsors that made this episode possible. Bankless is proud to be supported by Uniswap. Uniswap is a new paradigm in asset exchange infrastructure. Instead of a cumbersome order book system where trades are matched with other humans, Uniswap is an autonomous piece of software on Ethereum, which is what Ryan and I call a money

robot. No human counterparties or centralized intermediaries, just autonomous code on Ethereum. Input the token you want to sell and receive the token you want to buy. Something brand new in the Uniswop ecosystem is the Uniswap Grants program is now accepting applications for grants. We have been saying this for a while and we'll say it again. Dow's have money and they are in need of labor. If you think that you have something to contribute to the Uniswop Dow, apply for a grant to Uniswap. Just look at the size of the Uniswap Treasury. It's almost $3 billion.

This mountain of capital is looking for labor. Do you have something of value to contribute to the Uniswap Dow? No matter how big or how big or you know, it's worth. or small your idea is, you can apply for a unigrant at unigrantz.org and help steer Uniswap in the direction that you think it should go. That's exactly what we did to get Uniswap to be a sponsor for Bankless, and you can do the same for your project. Thank you Uniswap for sponsoring bankless. AVE is a borrowing and lending protocol on Ethereum and just recently released Avey version 2, which has a ton of cool new features that makes using Avey even more powerful. With Avey, you can leverage,

the full power of defy money Legos, yield, and composability all in one application. On Ave, there are a ton of assets that you can deposit in order to gain yield, and all of those same assets can also be borrowed from the protocol if you have deposited collateral. Here you can see me getting a 200 USDC loan against my portfolio of a number of different defy tokens and ETH. I'll choose a variable interest rate because it's a lower rate than the stable interest rate option, but I could choose the stable interest rate option if I wanted to to lock that interest rate in permanently.

One of Ave's V2 features is the ability to swap collateral without having to withdraw your assets. Trade them on Uniswap and then deposit them back into AVEA. Avey does all of this for you, all in one seamless transaction. So you don't have to repay loans in order to change the collateral you have backing them. Check out the power of AVE at AVE.com.

That's AAVE.com. All right, Bankless Nation, we are super excited about this episode. We are going to dive into the topic of MEV, that is minor extractable value. And we've assembled a dream team panel here. I wanna introduce our guests who we are going to have this in-depth conversation with. The first is Phil Dian.

He is a researcher and software engineer with a passion for computer security and applied cryptography. You might remember Phil authored the original Flash Boys 2.0 project. He's now working in the Dark Forest on a project called Flashbot We also have Giorgio Constantopoulos, who is a research partner at Paradigm. Before that, he worked as an independent consultant. Georgios is a big brain. Follow this guy on Twitter.

Lots of big alpha leak topics there deep in the protocol. We also have Charlie Noyes, who is an investment partner at Paradigm, along with Georgios. He is a bankless podcast alumni. We've given Charlie's full bio in history in a previous bankless episode. episode, so I will not repeat it here, but you know, Charlie, he is also very informed on this topic and all things. Guys, we've assembled this dream team panel. We're going to talk about MEV. How's everyone doing today? Pretty amazing. Excited to talk about MED. Very excited to talk about MEV. All right, guys, so we're here talking about MEV. And as part of the catalyst for this conversation,

of course, MEV has been a topic that has been a longstanding topic. Something, very important in depth. But one of the catalysts for us having you guys on bankless is we read this article from a coin desk. It was called why Ethereum's minor extractable value problem is way worse than you think, way worse than you think. Of course, kind of a clickbait title a little bit. And yet, you know, there are some really interesting points that the author made. And I think points that the bankless nation gets to, has to wrap their head around. This topic is incredibly important. Maybe it's the most important unsolved problem in crypto or in Ethereum. Anyway, I'd like to get your take, and I want to start with that question.

Is MEV way worse than we think? Why don't we start with you, Phil? Yes, absolutely. That's kind of why I started working on MEV and why I talked about it at DevCon and why I've kind of anyone who knows me in the industry going back a few years knows this is like basically all I've been talking about. I think MEV is a very deep rabbit hole. It was kind of intended as a definition to unify all these different problems in the space,

things like what is software security, how do we secure smart contracts, how do we write secure code, as well as things like what does economic arbitrage and financial arbitrage look like in this kind of new domain. So all these different kind of angles on the problem from software exploitation to arbitrage, to system design, to protocol design can all be studied and unified under this one metric, this one measurable quantity called MEV. And it was kind of designed to unify all these very lofty goals. So kind of taken in its sum and in its whole, the whole problem is super important. It kind of underlies everything we're building in the space

and underlies the security and usability properties of every protocol. So no matter how you interact with the cryptocurrency, it's important to understand MEV and understand how it can affect you, your protocols, and kind of the systems that you're using or designing. Charlie, when we did our podcast with you, which is one of the earliest bankless podcasts we ever did, you described M-E-V as the one, the single one thing that could end the crypto experiment. Maybe you didn't use such as such crazy words, but it's something of that nature. Why do you think M-E-V is such a crazy threat to Ethereum and Crypto-Economics at large?

Well, I distinctly remember the first time that I learned about M-E-V. It was at MIT, MIT's cryptoeconomic conference, I think in 2018. And Phil was giving a talk. And I had heard of, I had read the paper, but I had never before heard sort of Phil talk about MEV in the broader sense. And I remember sitting in the audience and like, like, like just like like like a pit of doom formed in my stomach. And I and I was like, holy shit.

the sky is falling. This is absolutely terrifying. To answer your question, though, like with crypto economics, you know, in the past, a lot of work has focused on like externally motivated attacks, like Goldfinger attacks and other types of sort of like thought experiments that you can construct as to how, you know, under certain conditions, like the crypto economic security assumptions of blockchains, of blockchain systems can be violated. So like a simple example would be like, you know, if a Bitcoin miner could access an infinitely liquid put option market on Bitcoins,

then like hypothetically, you know, you just destabilize the chain and Bitcoin is worthless because it can't make progress and, you know, you make a lot of money. And there are actually an infinite number of these such thought experiments. like and and and and sort of for a long time like there are many I think there are a lot of named ones as well and and people kind of spent time coming up with different scenarios like depending on you know financial derivative positions that like minors could take network splits whatever all the stuff but all of them were essentially externally motivated and I think the reason that

nobody was super concerned about them is because there's no obvious reason that they get worse over time or with scale. So like, there's no obvious reason why today, I, if I wanted to go, if I was a Bitcoin miner, like could not practically take a large enough short position for it to be like profitable for me to destabilize the network. And there's no obvious reason why if Bitcoin is twice as large as it currently is, that that will change, right? Now, the thing about MEV that has always scared me, I, it has always scared me.

and when I heard Phil's talk that terrified me, is this idea that it's internal to the system, that it's unbounded, it's internal to the system. And like in the Ethereum context, I think you could expect that it will grow over time. And perhaps even that the growth rate, that the growth rate will outstrip the growth rate of rewards, which is fixed and constant. And I think we have seen that happening. Like that trend is sort of seen in practice. So we've joked that, you know, MED, if it is like the worst case would be like a

crypto-economic cancer. It just takes over the system. To take the contrarian view on this. Thank you, Charlie. I think that MEP can be seen as a bad thing with all the usual consensus destabilization kind of description. However, I think that MEV can also be thought of as a security budget source. For example, if you have a blockchain that does not have a block reward, you could fund

the security of the system not on transaction fees paid via the usual gas price auctions, but via MEV. So I think that it's important that we kind of distinguish that there's also this kind of angle, the problem. So before we get too far into this podcast, I think what we want to do is define what MV is. And let's try the simplest possible explanation of MV, knowing that there's some people listening to this. This is their first exposure to the topic. And we're even throwing out an acronym like MEV and they're like, what? So, Phil, what is the simplest possible explanation of MEV? Maybe I'll give two definitions. I'll give you the historical definition of where MEP came from and then the kind of more modern take, the re-spin of the definition. So historically, MEV, it stands for minor extractable value. It's the amount of value in cryptocurrencies that minors,

especially in proof-of-war cryptocurrencies, are able to extract, aka increase their own balance or send to themselves by taking actions in the system. So miners have a lot of power in cryptocurrency systems. They are responsible for things like choosing what order transactions were included in, in a block in Bitcoin or Ethereum or other such networks. They're responsible for choosing which transaction gets included or not in a block. And their job is to kind of listen on the network and to choose and assemble and act in the system kind of based on what they see in the world.

That being said, they have some power given that they have all these choices of like, which order do I include? Do I include Phil's Transcendant? and not George's transaction? Do I include George's transaction before Charlie's transaction? These are all choices. These are all freedoms the miners have in the system. And all these freedoms taken together give them power to influence outcomes, and that power is valuable. So the value that they can extract by doing things like reordering transactions or inserting their own transactions with themselves, kind of in the transaction, or censoring other people's transactions, causing other people's transactions to fail, kind of any combination of all these things that minor

can do, the amount that they can increase their own profit, that's called the MEV, and it's a number. Basically, the idea behind MEV is that by quantifying this, by measuring this number precisely, which is a very hard problem to do, you can measure the security of blockchain systems, and also how different contracts and different systems interact and compose together and affect each other's security properties and affect each other's UXs and all these kind of complex questions that you can't ask without the lens of MEV. You can also look at, and this is why the MEP was originally defined, the effect on the consensus game. So how does this change and corrupt incentives of miners to actually secure the blockchain,

which is what we're trying to pay them to do? So this extra incentive to do things like manipulate the system for profit actually corrupts the minor incentive to participate honestly and to contribute to the security of the system by sometimes causing them to be incentivized to do things like ignore each other's blocks or compete with each other when otherwise they wouldn't and other such games. So in that way, MEV was also designed to quantify the effect on all blockchain users. So even if you're not using a system that has MEV, you're using the same blockchain, the MEV will affect you and the security of your blockchain is getting from its minds. Just real quick to update the definition.

So a lot of people ask natural questions like what about proof of stake, what about layer two. So we've kind of rebranded MEV to now be maximal or maximum extractable value. I think maximum is actually the canonical. And what that is is basically the maximum value that any permissionless actor in the system can extract. So for proof of work chains, because the miners are the most powerful permissionless actor, and because any attack that any bot or actor on the network can do can be more efficiently done by the miners, all maximal extractable value is minor extractable value. They're the same definition.

Even if the miners choose not to extract it, it's still MEV because it's able to be extracted by the minor. That's what the definition says. Now in proof of work systems, in L2s, you have different actors in different permissioned positions and different privileged positions. And there, MEV kind of looks at the maximum value that's able to be extracted by kind of these actors or any meaningful kind of collusive subset of these actors. So that was a really long explanation, but hopefully it brings some color to what MEV is. It's great, Phil.

And look, keep us honest during this episode, if we say minor extractable value, maximum extractable value is the better, more canonical definition, more expansive and accurate definition, I think. One quick question before we introduce some of the other panelists into the definition of this thing and maybe give a recap of it, does MEV, you phrase it in terms of the blockchain world, right? But it's the category of MEV type problems, transaction, reordering type problems, censorship, some third party having the ability to with that power? Is that exclusive to the blockchain world, or do we see traditional analogs? And what are those? Not at all. I think there's a spectrum. So I think the nice thing about blockchains is it's very

Cartesian, very deterministic. Like you can measure a number. There's not risk. There's not uncertainty because within the blockchain world, everything is kind of deterministic and programmatic. So in that ways, there's some unique characteristics of specifically Ethereum M.V. That being said, it does generalize all kinds of other different settings. So here are a few examples. examples, one of them is a permissioned blockchain. You can have a fully permissioned blockchain where MEV is still meaningful to reason about. So one of the early myths in kind of permission blockchain worlds is like, we're going to spin up this blockchain between like Chase, you know, like Barclays and whatever other five bank consortiums, and they're all

going to validate and we're going to do this round robin protocol and that's going to disintermediate trust between these entities. And in some ways the presence of MEV on that chain like speaks to like a break in that abstraction barrier. We're like just because you've started your validation, doesn't mean you've reduced these problems or taken them away. So in that setting, it's absolutely still meaningful. It's also meaningful in kind of the mixed decentralized centralized settings. So some MEV more similar to slow market arbitrage in traditional markets might exist because of differences on centralized exchanges like Coinbase and Binance that are commonly

market leaders and dexes that are trading kind of within the blockchain abstraction and don't have efficient channels for price propagation other than this MEV fueled arbitrage. So that might be kind of a more probabilistic and more hybrid kind of style of MEP. And if you kind of take that to its most probabilistic and like most amorphous kind of form, you can also view traditional financial systems that way and like quantify for various actors, what is their benefit for being in their position? And I think like had the world been looking at the financial system through a more MEP lens, things like the Robin Hood drama that happened with GameStop, it would have been a lot more obvious. That was M.EV. Yeah, it was MEV. And like it's

something we've been saying at Cornell and like to people I've been talking about in the space even long before that happened that like look at how these people make their money. But why was that MEV? It was because hedge fund managers didn't like what was going on in the market and made a phone call to Robin Hood to reorder things. Yeah, Robin Hood was the M basically, right? They controlled the infrastructure of like whose trades get process and things like that. And they did several things. A, they were reordering and inserting their own transactions. It wasn't their own it was Citadels, right? And like sandwiching people basically, which is exactly what we see on flashbots and on like Ethereum today. So they were doing that like actively the entire time.

And that wasn't bad. That was just their business model and that was like all open and public. And then yes, they engaged in some more active censorship attacks when kind of the Black Swan happened of GameStop and of hedge funds losing money. And that also speaks to how MEV will play out, I think, in permissionless systems. When you see these system shocks, when you see these Black Swans, these systems of trust that we've built are very fragile, and they fail users at those moments, at those really key moments when they need to be working. And that's where a lot of the work in FlashBots is going,

how to kind of shore that up and how to think about that and prepare for that future. As maybe Matt Levine would say, everything is minor excitable value in this system. Something on this, I think that MEP is kind of, as Phil's blogs have written in the past, it's fundamental to having a permissionless system. If you don't have MIV and you have many users trying to touch kind of the same piece of state, whether that's a trade or a liquidation or anything else.

But if there is this kind of race condition where many people are trying to touch the same thing, which has value, if your system is permissionless, it will have MEP. If not, well, your system has permissions and actors and, well, the hierarchy is not as flat as you want. Yeah, I guess just to, I think we're going to get into this with the transaction. fee section later, but I think that there is an argument that like in the permissionless context, like transaction fees as a necessary DOS prevention vector, like are the original sin of MEV and unavoidable in the sense that if you don't give folks the ability to express preference on transaction inclusion, that's a DOS vector. If you do give them the ability to express that

preference that's MEV. I mean, not even a dose vector, right? If you want to design a quote unquote fair system or equity or you know egalitarian or whatever you want to call it, ideally you want people to be able to bid and express their preferences, right? I would argue it's a necessary quality. Yeah. Right. And by the way, guys, for those listening in DOS, you guys mean denial of service attacks, which is yeah, essentially without. Exactly. Right, without the ability to express preference then you can't differentiate spam from normal transactions and there's like a fixed cost of spam although you can't outbid the yeah so so i want to back up and Phil gave a nice illustration of of MEV and i want to do the same but but with different words the way i've heard this is that

you know bitcoin ethereum crypto economics these are supposed to be leaderless systems right there's supposed to be decentralized and what that means is that instead of having one leader that bestows the right to order transactions upon either themselves or they pick out their party Ethereum and Bitcoin, it's kind of ideally chosen by randomness, right? Like random hashes, the beacon chain randomness for proof of stake. And so while these are leaderless systems in the macro, in the micro, whoever proposes the next block has literal God mode over that one specific instance, right? God mode over that one block. And the way that Ethereum and Bitcoin achieve fairness is they they just shuffle around who gets God mode, right?

And like, oh, you get God mode, now you get God mode. But the difference is when you have been bestowed the position of God mode, you all of a sudden have very different incentives than the rest of the system. And depending on the characteristic and nature of the MEV, the MEV could actually keep your incentives generally kind of aligned, or different types of MEV and different kinds of strategies could actually make you very misaligned with the rest of the ecosystem. So, Georgios, I want to turn to you.

Maybe you can shine some more light and get some more clarity on MEV for us. What would you ascribe to be like perhaps good MEV that actually adds to the alignment between the transaction order? And what would you also call bad MEV where the transaction order is fundamentally misaligned with the rest of the participants in the system? Yeah, of course. So a good MEV extraction opportunity would be an arbitrage because, when there's two markets, ideally of trading the same asset, ideally would want these two

markets to be trading more or less at the same price. So what Arbitraiser's do, which exposes MEV, is that they try to make a trade on one place and trade on another place in an effort to make spreads tighter. And generally, in markets, we like it when spreads are tighter. So that's one case of good MEV. Another case of good MEV would be that, let's say that David has a loan on MakerDAO or on AVE. And suddenly, Etherprice starts to move against his loan and he's about to be liquidated. A good MEPV example here could be a minor or a liquidator, however, saving your loan from getting liquidated and saving you from the penalty that the liquidation penalty that you

would get. So there's ways basically to use MEP kind of to our benefit on that end. On the opposite side of the spectrum, a bad MEV would be somebody censoring a liquidation transaction, or rather a debt repayment transaction, such that your system gets, your loan gets liquidated. Many people in the Black Thursday days, they call that MakerDAO, when the MakerDAO loans all got liquidated at zero. Many people were calling that there was minor censorship, for example, around this. This was most definitely not true. But, you know, that would be one example where basically miners start to censor transactions like crazy.

Another example would be that let's say that you are on a layer two system that uses fraud proofs, and the miner just decides to censor a transaction forever. You could say that all the funds in the system could be under MEP because the minor, in a way, would have control of them. And Georgios, how would you also reclassify different kinds of MEP? as it relates to security for the overarching system, right? Some MEV actually adds to security, some MEV detracts from security. How do you draw these lines? This is a hard thing, and I think that if you ask three different people,

you would get a different answer. Probably the ways that the MEV contributes to the security of the system is, as I said earlier, around how do you align the incentives of the miners to keep extending the longest chain. Whereas a case where MEP can be seen as bad for the security of the system would be when the miners are trying to, they see an MEP opportunity five blocks in the past and they would execute the infamous time bandit attack where they would go in the past and they would rewrite the history so that they grab this past MEP opportunity. So I think it's a bit hard for us to kind of say this kind of MEP would be good for security and bad for security because it kind of depends. on how the consensus participant, the block producer, decides to utilize it. Phil, Giorgio said if you get three different people, you get three different answers.

Do you have a different answer? Yeah, I think Giorgio's hit some great points. I think there's a little bit I would add of like one extra dimension that I think is relevant here, which is like just the regularity. So if you have like constant M.V that's like available every block, for example, transaction fees, approximately the same amount every block, block rewards being the most regular. I think that's very different to analyze from a security game standpoint, just because it's much more predictable. Whereas if you have something like Black Thursday style MEP, where it's like,

you know, blocks that just have millions and millions and millions of dollars in MIV in like a very, very short time period because of how quickly things are happening in the real world, that introduces much more variance to like all of your security calculations, much more unpredictability. And in that way, kind of just like erodes your conclusions, whether or not the outcomes are expressed. I think in general, this is very much an open research question. So like no one has a definite answer of what MED is good or bad for security. I think what's clear is that it affects the game. And like it's clear that some types affect the game more or less or in different ways, but it's not clear what to do about it or or if it's fundamentally good or bad in like

a stability sense in the long term yet. I think that MAD shares an interesting property with a long time ago, Goun had written a post called Ethereum is inherently secure against censorship, which essentially makes the argument that because smart contracts can do anything, you can't constrain any specific type of behavior on Ethereum, i.e., you know, if we want to implement uniswap, there are a literally, theoretically infinite number of ways to do that. And so there's no way at the level of the protocol to preclude that kind of behavior. Like the same is true in the case of MEV and different behaviors that generate it. So I think that like there are certain, there are certain base cases like

Phil referred to like transaction fees that are baked into the protocol and like are able to be reason about in a general way. But like sort of as you start getting into the application layer, it becomes difficult to reason about. And in some senses, like, you just can't know what's happening because anything can't happen and no specific behavior can be precluded or even handled differently than any other. So the generality is quite difficult. to reason about. I agree. And for what it's worth, I have a follow-up to Goon's post that you're

talking about on my blog. It's called On Soft Fork Security that I wrote like several years ago. I think maybe two people in the world ever read it. So I'm mostly trying to shill it to Charlie who likes reading my blog posts. It's like way back in the archives. And it kind of talks about this. And I agree that almost speaks to like a fundamental time MEV kind of tradeoff that we've noticed where MEB clearly relates to like fuzziness around time and global time synchronization and subjectivity around time and its relationship to kind of deep computer science problems. And I think that fundamental connection is still being explored. Since you mentioned GUNN, I also do want to say one other thing which relates to our previous kind of line, which is about God mode and kind of leaders in the MEV protocol. It's worth noting that that's how a lot of cryptos work today.

but like on a metal level zooming out, like all we're really doing with like electing a leader, electing a specific minor is trying to come to a group consensus. And Goon himself has a project called Avalanche, which works differently in which it's much more probabilistic and you kind of get an approximate consensus by sampling people in the group rather than a strongly consistent log from one leader. It doesn't matter actually. So you might instinctually say like look at that, that solved MEB. There's no more leader.

there's no more God mode, there's not this privileged actor anymore. But at the end of the day, like there's still a power structure in that network. And the people inside that power structure and that stake distribution and that validator set, they still have influence over the outcomes just by like biasing which transactions they announce first or which they prefer versus others or which orders they go with. And even though they're not the God mode leader for that period, they can still influence the outcome probabilistically. So all you're doing, even when you remove leadership, as the natural solution to these kinds of problems is like making it more probabilistic, which is still kind of MEV, it's still a payment, it still has all the same economic

properties and implications on your consensus game. So just something I wanted to mention that kind of slip bias, that there's no like silver bullet, like, oh, let's just get rid of this one god mode guy and suddenly we're not going to have MEV. It crops up in all sorts of places you wouldn't expect. Bankless listeners, that's exactly why we're talking about MEV. It is such a pernicious problem, something that is one of the remaining unsolved problems, I think, with scaling this credibly neutral open financial systems for the world. And I want to turn this next question back to Phil again, because Phil, you've been talking about MEV for years. And I think a lot of people were unconvinced. Like, yeah, okay, thanks for thinking deep into the

future. It's an interesting hypothetical, but like if it's coming, Phil, where is it? Right. And then there were some posts, people like Georgios and Dan Robinson wrote this fantastic post about the dark forest. I think there's a bankless podcast somewhere about that. We'll include that in the show that's sort of illustrated a real world example of MEV. But now we're like starting to measure it. And I think FlashBots is doing some incredible work measuring it. But let's talk about the measurement of MEV. How much MEV is out there right now?

And let's maybe take Ethereum as our example here, Phil. Tell us how to measure it and how much of it is there. Yeah, that's a great set of questions. So totally agree with you. You know, I have been kind of going blue in the face about MEV for a while. It's actually been an interesting kind of process. I think I would say it's the first like really major kind of research meme that I stumbled on a little bit earlier than the rest of the community. So like I spent a while just working on it, measuring it, building bots.

And it was like I was going a little bit crazy because I like wasn't sure if it was real or not. And I know everyone who went through early MEV also had the same journey because I would like show them what I was doing and explain it and like show them code and like numbers and things like that. And they'd like, wow, this is a really big thing. Let me go like drop my job and go work on this. And then they'd come back and be like, well, there's only two protocols. And it's like not that much here. Where is it feel like you're lying to us? But I think, you know, research is about predicting the long term.

And there's still so many facets of this that we have in. uncovered yet. So maybe I can start by pumping two flashbots measurement products, which is explore.flashbots.net and dashboard.flashbots.net. As you're speaking, I'm actually going to bring that on screen for our YouTube watchers here. Amazing. So the first one, explore.flashbots.net, is designed to measure MEV activity on Ethereum that's kind of outside the flashbb ecosystem. So this is already pre-existing MV activity that was going on, things like bots on the network. placing bids with minors over the mempool. That's what FlashBoys 2.0 was about, the research paper

that kind of started the whole MEP meme, things like liquidations, which Charlie and Giorgos just talked about, things that were kind of going on before Flashbots. And then Dashboard.flashbots.net is kind of what is going on on Flashbots. So that's like very FlashBets-specific kind of MEV measurements and metrics and like what's the state of the network, who are the miners, who are the searching bots, what's the distribution, what is the total profit, and things like that. So I encourage you guys to dive deep. We're super into transparency. To give you some ballpark numbers, you know, like on this Explorer page, you can see $12 million of extracted MEV in the last 24 hours. There's a few important caveats here, number one. We have to manually add coverage for different

types of MEV to this dashboard, right? So because there's so many different types of MEP, to quantify kind of what these bots are doing and things like that is a little bit of an active effort. So we may be lacking in coverage on certain contracts or certain activities, which makes this number kind of a lower estimate for the minimum amount of MEP that's present in the network. Well, ironically, any estimate is a lower, sadly, any estimate is the lower bound. Exactly. Yeah, yeah. That's the other thing that like MEV in theory is, like almost unbounded just because of how the problem is defined and like probably increases over time as like you have more time to like look back at the blockchain and things like that. So like even

what's being extracted is not what's being what's out there. So like 12 million is not the amount of MEV in the last 24 hours, not even close. Like if the bots were playing perfectly, if the miners are playing perfectly, they probably could have gotten a lot more. So we're kind of bounding it and we encourage people to contribute and improve this by contributing code that improves these estimates. It's all open source. We're kind of revamping also the code base right now because, you know, it's a little bit too rust-focused, thanks to someone who's on this call, who's a nice rust developer over here, Georgios.

But we're de-rusting it so that all you find folks in the community can more easily contribute. Context being that when we were populating the initial database, we had Scott Bigelow from the Flashbot team, shout out. He wrote the initial, like, prototype of the M.EV. inspector and then basically what I did was that I took this and then kind of re-wrote the whole thing in Rust which is a great memory safe like very high performance programming language I'm a big fan all the programmers in the show should use it and basically what we did was that we did use that in order to be

able to index like like the last two years of history like in a very short amount of time because when we were like developing new inspectors all the time because we started with just one we started with uniswiswold and sushi swap. But then we started adding more and more and more. So while we were prototyping, we also needed high performance because each time we added a new inspector, we had to go back to the start of time and re-index everything to make sure we did not miss any MEV opportunity then. But it turns out that the language is not as kind of widespread as I would have hoped. Yeah. Maybe, Ryan, if you want to pull up the MEV Explorer dashboard again, maybe Georgia, as you could

walk us through some of the evolution that's happened over time. Yeah, happy too. So basically, we see that until the Defy summer, July 1st, you don't see much. Or well, you don't see much comparatively to what we have now. And the reason for that is that that's when DeFi summer started and all the yield farming madness also started. And when all the yield farming gas fees started kicking in. Exactly, exactly, exactly. And also when all the, exactly, when gas prices start to go up, MEP, like the MEP that goes to miners also went up. And basically what happened over time is that more and more and more protocols started to launch and we now have seen this kind of evolution of things.

What we're missing in this dashboard is MAKER. So a large source of MEP in March 2020 was Maker. And the reason why we do not have MAKER is because simply the auctions or other the liquidations in Maker, they play out more. like kind of options instead of the kind of atomic liquidation where you provide the liquid and instantly you get an asset in return. And so these are harder to classify because they may play out on a multi-block time scale. So this thing, it classifies everything that can happen in a single transaction in a single block. But the reality is that the actual M-EV that's out there, it may

happen across multiple blocks. And so I want to zoom out again and kind of talk about measuring M-EV kind of at a high level. You guys talk about how you guys are constantly finding new ways to classify MEZ, MEV, and then going back and then adding that to the historical data. And if we resume all the way out, and try and measure MEV more accurately, it actually kind of turns into this like subjective qualitative analysis as to what actually MEV is. And I'm reminded of the like the metaphor of like measuring a coastline, right, where if you measure a coastline with very low resolution from very, very far away, you'll get one relatively low number, but if you increase your resolution and if you increase your granularity, the coastline of a country or whatever can actually become like

three times as long. And so what is the actual measure of the coastline? It really depends on the tools that you have available and also what the subjective analysis is for how you measure these things. And so maybe one of you guys could walk us through. Why is it actually not so incredibly like objective about what MEV actually is and and the process for actually measuring these things. Charlie, let's start with you. Sure. So I think, you know, maybe one way to think about it would be it's less a question of like what is and is not MEV and more just a question of like what is feasible to like heuristically

kind of classify or rather identify. So the hindsight looking change would not be, you know, like that we decided that some activity, you know, was in fact MEP. It's that we recognize it was happening. We add the heuristic and then we look back at how much MEP has been collected over time. And so part of the, you know, lower bound estimate today element of it is, you know, as we add more and more sort of heuristic, as we can identify more and more types of these behaviors horistically,

then we can also go back in hindsight and see how much we, how much we missed. So three years from now, we might look back and be like, oh man, we were off by a factor of 100 today. Correct. And what Charlie said was extremely accurate. So building the software, there is no kind of arbitrary rule that said, this is MIV, this is not, like some very generic.

What the software does is that it literally goes, like, step by step in the transaction, and it sees, does this match one of the things that I discussed in my call with Phil last night about the heuristics that we'd want to classify for? There is no, like, automatic thing, which, you know, you pointed to, and it tells you, this is MEP and has a million dollars in it. It's literally, like, very, very manual. So I want to stress basically that the thing that we found right now, it may be like so big, more bigger than we ever imagine. So what might not be obvious to those listening on the podcast is we were showing a flashbots UI and Explorer, which shows one of these hockey stick like growth curves that we are so familiar with in crypto. But this is a depiction of MEV growth over time, right? and George Eos was saying it started in Defi Summer. At that point in time, about three and a half million or so total MEV was able to be measured.

Now, flash forward to today, we're at over half a million, sorry, half a billion dollars in MEV. I want to ask this question to Phil maybe. Is this sort of what you predicted, kind of a hockey stick-like growth of MEP? and what accounts for this growth? Like, where is it all coming from? Is it the case that because the Ethereum economy is booming and Defi is booming, we get all of these new MEV opportunities.

Characterize that for us, Phil. So in terms of what I was predicting, I'm actually terrible at making predictions and, like, my track records of, like, you know, investments and, like, being crypto for a long time is, like, very much proof of that. So I didn't really have any intuition on the trajectory. I didn't know whether it would take, like,

one year, two years, five years. 10 years. It all happened kind of a lot quicker than I expected. It seemed a lot slower at first, but like in retrospect in the last like two or three years, it's been all a lot quicker than I expected, I think. The only thing I predicted was that kind of the incentives were set up that like ultimately this would be extracted and there was no kind of way around that. That was prediction A. Prediction B was that there was a big chance that it would be centralizing on the network. If you had certain specialized actors, people who used information asymmetry, actors like hedge funds in traditional finance or other kind of silos of knowledge,

I thought those were all very possible, and that those would become the most competitive miners if nothing was done to kind of distribute rewards or otherwise redesign the system, because those miners would be more profitable than other miners, and it would be in their interest, if Ethereum especially stayed on proof of work long term, to kind of centralized power to themselves and use that competitive edge

to drive others out of the miner and mining pool game. And we have seen that to an extent. We've seen mining pools like Ethermine kind of picking up projects that would indicate that they're working towards that direction. So I think those were kind of my main predictions and also that all these things would kind of harm user fairness

and cause financial loss and that people would continue carelessly designing systems without thinking about MEP. So I think all those have kind of somewhat come true. I hope that in the future, as the problem is accelerating, and I think the meme is taking off much more than I ever thought it would, that people think a lot more carefully when they're developing DAPs on kind of the ground floor about what MEPB we're exposing and what this means for our users. Other than that, I think things are going relatively on track,

and there's still a big possibility for MEP to be a really terrible centralizing thing for cryptocurrency, which we definitely want to avoid. Yeah. I would maybe, just add on kind of the question around like, will it get worse or like, you know, did we expect this to happen so quickly? I mean, it's always hard to predict the future, but like one mental model that I kind of have for MEV generally is that it kind of acts like pressure on the incentives of the system. Like we've kind of designed this nice little box. And there are, certain valves on it like transaction fees and others that the pressure created by like all the economic incentives can escape out of and so like as MED started to take off this summer something

that we saw happen was back running started which is a normal Ethereum transaction or priority gas auctions you can bid for priority get your transaction in front of another what you what those auctions don't allow you to do is bid to get behind someone else. However, the default Geph config at that point in time this summer basically said that when these transactions have the same gas fee, like it's random, which one ends up getting included behind the other. So we started seeing people sandwiching trades, not just front running them, meaning getting a transaction in front and spamming a ton of them, spamming a ton of transactions to hopefully probabilistically get one in behind as well.

Then because this is not a consensus level kind of constraint, it's based on like the configuration of gas miners and how they handle transactions with the same gas price. They decided to change it away from random to stop the spam. After that happened, a lot more miners started adopting a lot more proprietary solutions. And I think that like the, you know, expectation that things kind of continue down this route for me is like there is a lot of pressure. And every time we've tried to like close off valves that allowed it to escape, we've, it seems like we've just pushed the system for push miners further and further towards just like leaving the sort of default state.

And so like I would say that in hindsight, like the decision to change the GF config to prevent backrunning almost certainly accelerated the adoption of, you know, like custom mining software. And so to the extent that's true, it's sort of like I think, you know, indicative of like that trend likely continuing. We really are squeezing the,

balloon with this MEV problem, just like the air just seeps into another pocket of the balloon. And guys, if you were wondering why gas fees were so crazy this summer, users, defy users, this is part of the reason why this sort of activity goes on. David, I think you've got next. Yeah, the metaphor of the balloon is a really apt one. It actually does like a very good job explaining why we all care about MEV. It seems to be this thing that is intrinsically difficult to capture and harness. Yet, if we can capture and harness it, a lot of potential comes out.

of this. And so I think this is actually a great place to turn to the topic of the FlashBots project that Phil is working on, which in my mind is an attempt to harness and capture MEV and make sure that it behaves in the way that we want it to. And so Phil, maybe you could walk us down like the project of FlashBots. First off, what is the goal of FlashBots and how and why are miners incentivized to participate in FlashBots in the first place? Yeah. So the goal of FlashBots is to Democratic. MEP basically, so to make sure that if possible, and maybe it's not possible, there's still an open question that, as Charlie said, MEV could be totally unfair and torpedo all of cryptocurrency. But if that's not the case, and it is possible, then what we want to do is we want to drive miners to a state where MEP is well distributed, where the ecosystem of MEP extraction is efficient, because that's important for network security. If it's not efficient, that becomes an attack vector. Where the profits are democratized, to all the miners in the system in a way that's not centralizing,

where the system stability is insured, and where the UX of everyone else in the ecosystem, from client devs to layer one devs to layer two devs, to DAP users, to DAP devs, everyone else in the ecosystem is not trashed by MEP, to the point where the systems become unusable, expensive, unfair, uncertain, or otherwise risky. So the goal of FlashBots is to kind of form an open-source research and development,

collective where we create products, we create research, we create other artifacts and other ventures and structures that are intended to shift the incentives of the MEV game towards a more decentralized future that's more aligned with cryptocurrency and with the open source free software movement that backs cryptocurrency itself rather than purely fueling information asymmetry. So as part of that mission, one of the products that we've released so far is called MEPGeth or MEPGeth. And it's a patch to the Go-Eetherium reference implementation, client implementation, that basically allows miners to create a channel by which people on the network can submit bids to them for private bundles of transactions to be mined.

Those bundles containing other people's transactions or their own private transactions. This allows arbitragers and other bots who are running economic strategies on the Ethereum network to have a more efficient channel to communicate with miners than the Mempool. Why did we build this? Because we saw MEV kind of taking off and we saw miners building these channels independently and individually as a centralizing force. So we thought that having one single community channel across all miners that maximized also bot activity on that channel would be the most beneficial equilibrium from which we can

then build a more decentralized, more robust kind of long-term-looking ecosystem. We thought we had a better chance in that kind of decentralized protocol model that we were striving for rather than allowing the fragmentation to kind of naturally take shape, which possibly could lead to the fulfillment of the predictions I made about MEV being centralizing and centralizing to actors like hedge funds. So that's why we built MEVGET. Why do miners run it? Because it strictly increases your profit over running Geth. If any part... of MEV-Geth breaks, if flashbots fails, if whatever happens, there's a strict security property in MEV-Geth, and it is alpha software, so run it at your own risk, it's not fully audited,

but it is coded in a way such that at any point in time, if you mine on top of a block provided by MEV-Geth, you will make at least as much money as mining on top of GEP. So in basically all aspects, it gracefully degrades to the status quo, which we think is the optimum for network resilience and which we think is the optimum for also minors to make it a very easy decision to switch. You don't have to do complex reasoning. You just need to kind of run this software and let the decentralized community compete to kind of maximize your share of the MV game. And a lot of the work of developing the software and staying at the bleeding edge of actually understanding and processing MEV, which is a constant arms race, is outsourced this decentralized

marketplace rather than needing to be performed in-house by each minor and needing to be duplicated and becoming kind of a centralized information silo source. So that's why we built it, and that's kind of what it does. It's not the only thing FlashB is going to build. We have a long agenda of kind of projects and products we want to do in the MEPB space with the broader goal of really aligning the incentives of MEP and cryptocurrency as a whole. And that's really the throughline that I see with Flashbots. And overall, Crypto Economics at large is that it's really,

Crypto Economics is a sandbox for tinkering with human incentives. and with clever people like the three panelists that we have here, if we can figure out a way to incentivize people to stay with the group, then we will figure that out and implement that. That's like the optimistic case. The model for FlashBots to me is getting people to coordinate together for commonly shared good, and it only actually works if there's financial incentive to participate in the network.

And so, Phil, where does this, how do you expand upon this model? What are like the next steps for flashbots to increasing its overall usage and adoption and making sure that the It reminds me of the the model the theory of like the the the Leviathan right where We have this this powerful person with with a sword that controls the body and it controls the body via force But like flashbots controls the tries to keep the body composed via rational economic incentives How does how does the quote of quote metaphorical Leviathan that is the flash bots and grow to compose everyone and stay and keeps the body composed as a whole? Like what are the next steps for growing the Leviathan, if you will?

Yeah, I think there's a million different things we're doing just in parallel. And I think anyone in the community can come be involved. One of the next steps is building a community. So come join our Discord and be part of this question of organizational design and how we build anti-fragile organizations and how we align these incentives. I think there's a lot of question mark, question marks there. Maybe I'll give you some like immediate random sampling of things we're doing at flashbots, but it's far from exhaustive.

One of them is decentralizing ourselves. So right now, flashbots mainly is in a kind of alpha testing mode with our software, where we're very much reliant on some centralized infrastructure that we operate. Not fully reliant in that, like, people have forked this and you can use flashbots without using this infrastructure. But we're not comfortable giving the software to minors

and saying, like, just remove this, because then it opens up attacks that we can't control. So removing our need to police things like spam on the network and police things like privacy, this privacy property that we're kind of enforcing as this incentive cartel that you basically mentioned. That's how we're doing it. Removing ourselves from those equations are kind of the immediate next technical steps for our current product line, both through cryptoeconomic and through cryptographic means we want to provide full privacy of anything going through the flashbots network without relying on centralized parties and without relying on trust assumptions.

We want to also provide spam resistance so that we're not needed on the network whatsoever. Once we've done that, we're also doing a lot of work on integrations. So looking at things like ETH2, layer twos that are coming up, all these other blockchains, the universe of Cosmos parochains and polka dot chains. Sorry, I might be mixing my terminology, but basically integrating all these other systems into flashbots and creating basically the ultimate preference expression engine for financial bots in the cryptocurrency ecosystem across any chain, across multiple chains. This is kind of work we're doing. And then also just a lot of ethical work, a lot of work on governance,

on how to ensure we build an organization that does good research that doesn't lead to runaway information asymmetry, how we contribute back to the open source commons and the standards that we use for doing that. That's a lot of organizational design work we're doing right now. And we do have regular community calls about all of these subjects. So those are, as I see at, the kind of immediate challenges is just making sure all incentives are aligned in a robust organizational way and that no trust is required anywhere in the system. Those are kind of the near-term goals. And the longer-term goals are to create the platform for the robot financial revolution, basically, and, you know, give people this ability to, in a decentralized and completely permissionless way,

express their preferences and express them securely in these auction systems and in these cryptocurrency networks. So, Phil, a lot of what you've been talking about with the success of flashbots happens at a level, I think the typical defy user might not be aware even exists or even sort of notice. But I think a lot of the activity that Flashbots is harnessing was happening on Ethereum Mainchain inside of the Mempool. And ever since the Flashbots project has launch, some of that activity has actually shifted to the FlashBots network, which is super interesting, which I think Defi users might have experience in the form of lower gas fees. Can you confirm whether that's true? Because that's sort of the rumor that's going around,

like some of this MEV activity instead of it happening on Ethereum Main Chain is now happening in this private network. And that has contributed to reduction of gas fees and costs for transactions on the Ethereum Mainnet? To what extent is FlashBots responsible for that? And can we thank you for that? I think it's really too soon to tell. I think people really like that meme and like as much as I wish I would say like, yes, it's all us. I think we want to make cryptocurrencies more efficient and we've done a lot of work on that front. Like we've definitely reduced congestion on chain. We've reduced PGA's. There's less estimator inaccuracy as a result of all these things. And there's just more headroom as well as more minor revenue,

which leads to things like decreases in gas token and chi minting and other sorts of pathological activity. So I think FlashBots' mission is to reduce impact on these commons. And bots were trashing these commons before. They were like trashing the network and trashing the gas space. And like we've brought a lot of that down. That being said, I think a lot of other things also contributed to low gas prices, like an increase in supply of gas and a general stability in price for a while in cryptocurrency. So I think FlashBoss has helped a lot put downward pressure, but there's also a lot of other downward pressure.

And without like really sophisticated economic modeling, it's hard to take credit. The only thing I'll say there is like, you know, we're definitely working in that direction. And one of our goals is to make things more efficient. And I think that's definitely happened. And good side effects of that are like part of what we want to do here because we don't want a situation where there's a few MEPB bots kind of pissing all over the commons and like ruining the network band. and things like that, just so they can make their arbitrage profit. Before we leave this topic and talk about sort of futures and some other things, curious Charlie and Georgios, do you have any insights?

Charlie first. Yeah, I think beyond everything that Phil said, a maximally successful version of flashbots is a future in which Ethereum never experiences a time-banded attack. end of the day. Well, we're going to have to go into what that is. I know we talked about it briefly, but can you just go ahead and give us a recap on a time bandit attack and why that's so critical to not have that?

Yeah, the time banded attack, if we didn't cover it earlier, it's kind of like, I don't know, I'm trying to think of a proper analogy, but I don't know. It's just the doomsday scenario of MEV. It's the idea that, you know, When blockchains are reorged, which happens, small reorgs happen fairly frequently, everybody knows sometimes, you know, it seems like your transaction has one confirmation,

and it turns out that it doesn't. MEV can be extracted from back in time. So if I, let's say, for example, there's five blocks with like a ton of MEV in them, lots of competition, tons of, tons of trades, whatever, and 50 blocks later, it turns out that Georgios on the side here has been shadow mining a side chain and reorgs is the whole thing, capturing all of the MEV available in those 50 blocks. Not only the ME, and importantly, not only whatever MEV was captured by actors on the previously

canonical version of the chain, but he can arbitrarily reorder transactions in every one of those blocks, meaning that it gives you not just like the sort of God power mode as a leader within a single block, but the ability to do that over multiple blocks. So another example, this would be I take out, let's say I'm a minor, I take out a trillion dollars in die debts, and then I just wait to see what happens to the price of ether. And if I don't like it, it's like it never happened. It's gone from the canonical chain. So time bandits are like the, they're kind of terrifying.

It's like the worst possible form of MEV, right? It's like if bankless listeners will be familiar with the concept of MOLAC, right? Like it is literally the instantiation of Moloch in a single instance of being reord. Coordination failure. Absolute coordination failure. Yeah, quite literally, right? Because it's getting reorged. Hey, bankless nation, I hope you're enjoying the MN.

M.EV panel thus far. In the second half of this conversation, we go into proof of stake in EIP-1559 and how they do or do not impact M-E-V. We also talk about L-2s and the relationship between L-2s and how there's both M-E-V on the L-2 and there's M-E-V on the L-1 and those things actually interact. And then we also get into a conversation with Phil about the inherently centralizing force that is MEV, especially as it relates to both L2s and L1s. And then we also get into a more philosophical, broader conversation, which frankly, I only understood, like, chunks of it. And so I'm going to have to go back and re-listen to this after I do some Google

searching and homework. This is one of those episodes where it's not just for you guys, but it's also for me and Ryan to get our learning in as well. So I really hope you guys are enjoying this panel. But first, before we get into the second half of this conversation, we have to pause and talk a moment about these fantastic sponsors that make this show possible. Balancer is Defi's most powerful automated market maker. Typical AMMs just have two tokens inside of one liquidity pool,

which can lead to fractured liquidity across the many pairs of DeFi. With Balancer, you can access the full power of multiple tokens inside of one single AMM, which unlocks an entirely new playing field of possibility. This makes Balancer an awesome building block for so many different use cases. Balancer pools can make asset indexes, but instead of paying fees to portfolio managers, Balancer lets you collect fees from traders who use your portfolio for liquidity.

Additionally, Balancer Smart pools could be programmed to have properties that change according to predetermined rules, such as changing the swap fee based on market conditions, or even liquidity bootstrapping pools, which can help you launch and distribute your token with day one liquidity. At Bankless, we used a liquidity bootstrapping pool

to sell our BAPT T-shirts to much success. Balancer V2 brings powerful new features that makes your money work even harder for you. In V2, idle tokens are capable of generating yield in defy without sacrificing liquidity in the pool. To top things off, Balancer is reimbursing all gas costs with BALV rewards, meaning that all your gas costs are returned to your wallet with the Balancer governance token. Balancer's mission is to become the primary source of liquidity in DFI by providing the most flexible and powerful platform for asset management and decentralized exchange. dive into the balancer pools at pools. balancer.exchange.

Gemini is the world's most trusted cryptocurrency exchange. I've been a customer of Gemini since I first got into crypto in 2017, and it's been my main exchange of choice to make my crypto buys and sells. Gemini is available in all 50 states and in over 50 countries worldwide, and on Gemini there are markets for over 30 various crypto assets, including many of the hot defy tokens, and it's one of the few exchanges that has liquid dye markets. Gemini just launched their earn program where you can earn up to 7.4% interest on 26 various crypto assets. If you're tired of paying fees in defy or you don't want to worry about defy exploits, but you still want to earn interest on your crypto assets, Gemini Earn is the product for you.

Another product I'm stoked to get my hands on is the Gemini Crypto-back credit card, which gives you 3% cash back on all of your purchases, but paid to you in your preferred crypto asset. When I get my Gemini credit card, I'm going to make sure that I get my cash back in ETH, so whenever I buy something, I get a little bit of ETH bonus back to me at the same time.

You can open up a free of $1. free account in under three minutes at Gemini.com slash go bankless. And if you trade more than $100 within the first 30 days after sign up, you'll be gifted a free $15 Bitcoin bonus. Check them out at Gemini.com slash go bankless. So one of the interesting byproducts of EIP 1559 is it actually is a significant MEV mitigator. And so maybe we can talk about MEV and how it relates, or EIP 1559 and how it relates to MEV. Georgios, can I hand this one to you? Of course. Both me and Phil are laughing because I think, David, we've had this discussion

on Twitter before. I don't think that the EAPE159 actually addresses MEV at all. Zero, not at all, not in the slightest. Zero, no, nothing. Bargional, marginally. Marginally, maybe, but not at the scale that we need. I don't think that it would matter at all. Like you could make a very weak argument that maybe, you know, the tips that go to the miners are less and hence, you know, it's less than what they got before.

And hence, it's less M.VEVE, but not really. EAP 1559 is a gas market change. It changes how we bid for transactions in the system. But it does not change the way transactions are ordered in the system. There still is a minor that has the God mode and has full authority over which transaction gets executed when. E.159 does not change this at all. Hence, I do not think that it matters at all for MV. It mattered for a lot of other very important things, but not for MV, unfortunately.

So, Phil, you agree with that? Yeah, I totally would agree with that. So, you know, nothing constrains minors to behave the way they are told to order things in 1559. They still have a lot of degrees of freedom in the system of like where to insert their own transactions, whose transactions to include in this block or not, and all these other kind of freedoms that we mentioned that are valuable are still there. What's changed is a little bit how much, the only thing that changes is how each transaction's cost is calculated, basically. but that basically only matters to the opportunity cost of like reordering things, which is like a very minor part of the calculation of like whether a minor will extract the

MEV or not. And the change of like burning the base fee, because the base fee is probably going to be very small for any meaningful MEV opportunity with regards to the size of the MEV doesn't make a substantial difference in any of those calculations, even if it does economically make a difference in pricing on the chain and like requires a burn. So, yeah, fundamentally, MEP is still pretty much exactly the same flashbots keeps working as is. I think one interesting thing 1559 does do, which has been needed for a long time, is to separate the market out for priority and inclusion.

There are some users that just want their transaction to go in in like the next five minutes, 10 minutes, whatever. They want it quickly. They don't care necessarily that it's the next block or it's at the first position in the next block, right? They just want it in. And so for those users, I think the 1559 mechanism makes a lot of sense because it gives them some good guarantees that they'll be in eventually within some time period. And it gives them predictability on how to choose how much to pay, which they don't have right now, which is a very, like, complex, ugly piece of software that all wallets have to include. So in eliminating those things, it's great. But the market for priority, for who gets in first, it doesn't change that at all.

And I expect flashbots, you know, will run in parallel to 1559. actually, I know that will happen. And in fact, like one of the major marketing pushes behind why FlashB was adopted in the minor community was miners looking at the revenue post-1559 and saying, like, we're going to make a lot less for simple inclusion. And so we need to kind of increase our revenue elsewhere. And therefore, they looked at MEB. So if anything, it's really accelerated the MEV game more than mitigated it.

But really it's kind of a separate orthogonal issue. Right. Bankless listeners will be familiar with EIP-1559 because we talk about it endlessly. But just to recap, we have the base fee, which gets burned. But what you guys are saying is that really the M-EV game happens inside of the tip sphere, which is completely unrelated from the base fee, right? Because the demand to capture some amount of M-E-V will be fully expressed by the tip, not by the base fee.

Would you guys agree with that? It's not just in the tip. It's just in all payments. So the tip is one example. But even if there was no tip, you could just have a transaction that transfers money to eith.combase, you know, or block.combinbase from inside the, uh, the kind of system. There's an op code that lets you pay the miner in a bespoke way in any kind of transaction you want. So even if there was zero tip and 1559 was the only pricing mechanism, you could still just

make a direct transfer in a smart contract to the minor, and that would constitute basically an implicit tip. Uh, the reason the tip was, was added was to kind of stop people from doing this. Uh, so, so the tip is one place, but there's like, that's not the only place flashbots operates, we also take into account kind of these other more bespoke payments. Yeah, it's kind of also like, you know, if it weren't included, right, then like an off-chain market would like certainly be created. There's, you know, half a billion dollars, whatever that have been extracted. I think the, like pretty much anything that you do to try to prevent miners from accessing MEV,

because you can't actually get rid of the MEV itself is liable to incentivize the creation of like off-chain or off-protacle mechanisms to access it. Okay, well, let's talk about another upcoming protocol level change to Ethereum, which has got to affect something. You guys tell us. So that is staking, proof of stake. So basically the way I think about this high level

is that all of the miners, they lose their responsibilities, and those responsibilities turn over to the validators, those with stake inside of the Ethereum network. So you've got a new set of stakeholders. And what staking does is essentially merges the holders of ETH, the asset, with the obligations of validation. So this doesn't eliminate MEV. I think anyone who's listened up to this point, understood up to this point, can understand that.

But it's got to direct it to a new set of stakeholders. What are the implications of that? All right, Georgias, let's start with you. There is some impact this time, but to kind of stress it, MEV is, again, an ordering problem. It is about who orders and how. So proof of stake changes one thing. how validators are elected.

In proof of work, validators, rather, block producers, are elected randomly. You can never know who the next miner is with absolute certainty. In proof of stake, specifically, Ethereum 2 proof of stake, not any kind of proof of stake, because the protocol details may change things. In Ethereum's proof of stake, once you are a validator, you can always know the validators for the next 12 minutes, which is in if two lingo, two epochs. What this implies is that if I know that I will be the next validator and then I know that my friend Charlie is going to be the next validator, and then I know that Phil is going to be the next validator,

I can call them and I can tell them, hey, guys, like, we will have a guaranteed three block opportunity to do whatever we want with Ethereum. So what this allows, is for miners to, or rather block producers, to execute multi-block MEP extraction tactics reliably, while these before were only probabilistic, and because they were probabilistic, they were probably not something we would see. And the very trivial example of that would be manipulating uniswap oracles, Uniswap V2's oracles, for example, require that, rather they snapshot the value, from the last transaction in the block, and they use that to kind of determine the price in that

block. However, if I am a trader, rather if I'm a validator, I can manipulate the price to wherever I want in the last transaction of the block, and the next block, I can then push it back to wherever it was without realizing any loss net of fees. So the implication is basically that minors will be validators will be able to execute these kinds of attacks reliably. So you think it makes the problem worse, George, yes. I think so. Well, let's talk about another implication of this, which is you have completely changed the miners out. You swap them out for eth holders. So if you are a holder of ether as an asset, right, you are a potential recipient if you choose to stake and validate, you are a potential recipient of this MEV versus

minors. That has some effect. So I guess rather than minors receiving sort of the MEP payment, the bribe as some might call it, validators are the new recipients of this MEP payment. Maybe bullish the price of ETH, but not bullish the credible neutrality of the network? I don't know about bullish or not. What I do know is that right now, validators on ETH2 are earning a fixed percentage on their staked assets. Once EF2, the ETH2 merge happens, validators will also start earning transaction fees and MEV. And of course, what this means is that the APR for staking would be higher because as a validator, you now also earn MEV.

So that's kind of a positive byproduct of the merge to contrast the maybe negative out there attack that I mentioned earlier. I think politically it changes a lot of things too in that it changes the power distribution, which fuels how MEV expresses itself in the real world. So zooming out, like it's always important to also consider the social angle of these systems we're building and how they'll interact with society. Miners kind of have an incentive structure, which because there's a high barrier to entry

and a lot of capital required, has kind of taken a particular structure in the real world. So often you have very scoped business entities that exist to extract a certain amount of profit in a scoped way and have like investors and operations and things like that. In the validator world, you have very different politics. So in the large validator community,

you have things like exchanges, which are competing for users and trying to maximize user yields. There you might see things like MEV farming. You might see things like centralization through proprietary MEV strategies or partnership with the information asymmetry driving entities that I mentioned like hedge funds. You also have large holders in that picture who maybe have less technical resource access and sophistication, but a lot of system and incentive alignment. And you also have have a long tail of small holders who are trying to get access to APR and trying to get yields, basically, without kind of needing this full control and without having as big of a piece of the pie as these other actors. So that political landscape is really different than miners on

Ethereum today, where it's much more than even pie chart. A few things I think that says is that number one, it's really important for these small guys to be able to access their percentage of the MEV that they're owed. If you have 0.000.1% of the network, and you can't get 0.0001% of the MEV, you'll go to Coinbase instead. You'll go somewhere else, and that'll be a centralizing force to access these profits. So democratizing this and creating this network where each validator and each miner doesn't have to be doing the work themselves to get most of these profits, I think is critical to the decentralization of proof of stake. And it's like a big part of how we view this transition at FlashBots enshrining that.

And also being aware of these change power dynamics, we don't know how this will play out yet. I think these are all big question marks. So as much as I wish I can predict the future, I can just say, like, these things are going to change and these actors will be relevant increasingly in the future. And there will be kind of politics and drama there that we can't predict yet. I think the through line that listeners should really be cognizant of is we just talked about EIP-1559. We just talked about proof of stake. And in neither of these situations did the EMEV situation really improve all that much and perhaps even gets worse. and that kind of just illustrates the perniciousness of this problem.

And what Phil is really talking about is that if we do not figure this out, then the whole argument of that, you know, proof of stake makes the wealthy or wealthy. There are plenty of arguments as to why that's not true, but one of the biggest perhaps manifestations of a future version of the world is that we cannot solve MEV. And the wealth distribution of proof of stake just becomes so incredibly centralized because of the natural incentives that,

proof of stake, just really just the power asymmetry is just accrue faster and faster and faster to the large stakeholders. I want to maybe bring up two other implications that I'm seeing and ask the panelists about this. One is, and bankless listeners will be familiar with all of our episodes we did with Justin Drake about EVE being sort of an ultrasound money and even like modeling that out with like what does a world look like when validators are receiving issuance, validators are also receiving transaction fees, he did not model out what it looks like when validators receive MEV, which could be an absolute massive portion of revenue flows. And to the points David just meant that has, and the point of this entire podcast is that has some negative effects, negative externalities to the credible

neutrality of Ethereum potentially. And also yet, I'm linking this to a comment that Georgios made in the introduction, which is, that could also be fuel to pay for network security at the same time because it means eth issuance can be effectively lower. Why? Because it's being paid by not just transaction fees, by large amounts of MEV as well. Those are two implications like in my mind and also like just the implication I guess in the short run. Wow, like all of this revenue flowing to ETH holders like, wow, okay. But I want to get the panelist's reaction to maybe that.

Is there some flip side that is actually good from an economic security perspective? And since Georgios, you brought this up, maybe I'll turn it to you first. Candidly, not much to add, Ryan. I think you explained it very well. I think that MVEV like transaction fees can be a source of security budget in the future.

And I think it remains to be seen whether it will be enough of a security budget. Yeah, I think to add a little bit to that, I guess there's like small things we can do that change things a lot, right? Like, for example, just designing systems in an MEV-aware way. Like, if to expose MEV, you need to be mined on top of the particular block. If you add that as a condition to your transaction, it stops how deep you can reorg and it affects things like time band that attacks a lot

without being a substantial UX penalty. So there's a lot of like low-hanging fruit mitigations like this. that could probably change these equilibriums a lot and make MEV a better security budget. I think there are worlds in which MEV is a good security budget and worlds in which it's a horrible security budget, and it depends how we design our mechanisms, and it depends how things play out socially and on all these other factors, which are really hard to predict or even model. But I believe both of these futures exist, and we should try to work towards the one where it's maximally not bad as we can sort of build.

And that's kind of what we're trying to do. Well said. There's also a good, very obscure, but good post back from like May 2020 by Vitalik on Medium that I can say the link to afterwards, where he talks about basically, you know, MEV as like a centralization prone kind of revenue stream, ended in a role of centric Ethereum, although I guess this is probably written before, role of centric Ethereum roadmap is written, but anyway, that we might want, to centralize MEV extraction into roll-up sequencers, given that it's very high fixed cost,

probably more so than mining itself and definitely proven stake, and that if we're not able to harness the MEV revenue stream as a security budget, we could potentially at least quarantine it off, essentially more bifurcated from like the core consensus tension and make in the worst case, you know, kind of, like, if we needed a hard fork to fix something, it might be easier if the revenue stream is quarantined off. The basically the action of centralizing MEV on layer to sequencer networks, it does not remove the MEV, but it moves it somewhere else where it can be contained and basically not leak into the security of the rest of the system.

them. Right. That's exactly where we wanted to go next. And so sequencers on layer two's, entities that are responsible for ordering transaction, is a totally different design scape of MEV that's separate from the Ethereum main net, but is MEV all the same? And there's, you know, whoever gets to aggregate transactions on L2s gets to do extract that value. And then, and then the transactions go to the L1s. And then there's L1MEV miners or validators as well. So let's talk about some of the structural differences or similarities between MEV on L1 versus MEV on L2 is probably under the context of optimism and arbitrarum because I know these two these are the two L2s that are really going head on after this.

Georgios, I'll turn that to you. How is MEV the same and how is MEV different on L2? I mean, in the end, a layer two, when we talk about layer two today, I think that we really refer to a roll-up, an optimistic or a Zika roll-up. That's what Layer 2 typically refers to. Agreed, yeah. And this may also be relevant given discussions in the rest and days around what a side chain and what a layer 2 is.

However, I want to stress that if the roll-up has programmability, the MEV it exposes is the same as if that thing would have been a layer 1. So in a way, M.EV is not a property of whether it is layer one or layer two. It is a property of how much ordering is allowed in the system and how much programmability is in the system, hot tip to Charlie for his blog post, clearly laying out that aspect. So specifically, I think that, again, it matters on the election process for the person who will order transactions. Optimism design specifies the so-called MEV auctions, where the MEV auction basically says that if there is some amount of MED, MV in the layer two, then the sequencer that wants to extract it should be able to pay up to that amount to be elected to become the next block producer.

And that's what the design, rather that's what the implementation will be. eventually it's not there today and there's no kind of commitment to dates on when that would be live. For Arbit Room, I think that right now there's similarly kind of like a fixed or a fixed scheme for how validators, how sequencers are elected. If it's one, if it's many, I think it would be something like a round robin based on stake, which means that if you own more stake in the system, you get elected more often. but what Arbitrum also has been working on, which Phil maybe has some thoughts on, is the so-called fair ordering algorithms, where they say that instead of having one person decide on the ordering, you have, let's say, all the consensus participants to agree on their ordering. And this is basically the same thing as running a tenderment or a BFT kind of blockchain, where two-thirds of the stake first need to agree. on the ordering and then only does execution happen.

Yeah, and you can in fact add like a threshold sign, I mean, if you're using like BLS signatures, then as like part of a BFT style protocol, then you can just have the threshold signatures unblind after final, after finality. But obviously that restricts the set of possible participants given that you have like a quadratic communication constraint on the consensus

process. So I mean, guys, is the story here that we've just squeezed the balloon once again and now move the MEV problem back to layer two's and back to roll-ups? What do you think, Phil? I think this is where the coastline analogy you mentioned before is most accurate. So I think layer two is like a fractal instance of layer one. Like there is no layer two and layer one. There are just blockchains and how they connect to each other. Layer two is like another chain of blocks with its own process for designing ordering. So depending on how you design it, it might have the same exact MEV problems as a similar layer one would have. You could on each say like, oh, we're just going to give all the MEV to the tallic and like hard code that in the protocol and that would work, right? And that's

basically what optimism is doing. He just gives the charity though, Phil. The zero. Yeah, exactly, or whatever you want to do. So with a centralized solution, that's basically what you're doing. In terms of kind of fairness protocols, I think there are a lot of, that's kind of what we're describing with these kind of BFT style protocols and what Arbitrum is trying to do, where they say, like, here's some process that we consider more fair for some reason, and they state why this process is more fair. Usually it's because it makes some claim, like, if three out of the five nodes we have, here's your transaction before mine, your transaction will definitely get in and it'll definitely be before mine. And so therefore, that's a fairness property that takes away from ordering.

That being said, the way they expressed this fairness is in like computer science terms, not in economics terms or really like fundamentally speaking to MEV. So it doesn't talk about things like what if your assumptions fail and those three nodes collude against you. What is the incentive for these nodes to honestly follow the protocol? And like, is it detectable if they don't? How do network assumptions play into this? How do, which assumptions do we have to accept as users of the protocol to kind of engage in this protocol and believe that it's going to give us the output? So these are all kind of very deep parameterization questions that themselves introduce MEV and kind of represent blockchain assumptions. So in all these cases, in layer twos and layer ones and fairness protocols, these are all basically just blockchains. They all kind of have their own MEV that you can think about that you can calculate. And they also have MEV across them. So layer two is going to have MEV that's fractally larger, almost like the coastline, right?

Like when you combine it with the layer one to which it's anchored, because if you're in a privileged, position on both layer one and layer two, you can amplify your own attacks on each one. You can make your layer one attacks more powerful with the economic degrees of freedom you have in controlling value flow on layer two, and therefore how that flows back into layer one. And you can make your layer two attacks more valuable using things like layer one censorship and reordering and other messing with the kind of core security anchoring of that system. So let me ask you, the panel is kind of a follow up to that, right? So we've just chased MEV to a roll-up, right?

We've squeezed the balloon. Now MOLAC lives on a roll-up, but we still have the same M-EV problem, except for the fact, layer twos can harness a thousand different experiments at once, right? So the Ethereum settlement layer, other chains like other layer ones like it, they have to ossify over time. They can't experiment with new fairness protocols to try to solve the M-EV problem as rapidly. they certainly can't iterate as fast. Maybe I'll turn this question to Charlie.

So we've chased the MEV problem to layer two's. Do you think that there is some magic protocol out there, some magic mechanism we will be unable to uncover to actually slay MOLAC and solve MEV, or at least drastically mitigate it on layer two? Are you optimistic or are you pessimistic about that, Charlie? So I have a couple of thoughts. And first, it's a fantastic question.

I think that there will, that there is a wide space of experimentation that's possible in layer two. And I think that it is going to give us the opportunity to kind of try designing more potentially application-specific kind of controls. I think that for roll-ups where, you know, it's kind of, for roll-ups which basically just emulate, like, the EVM, like the problem. is essentially identical modulo, the sequencer selection. Just, I mean, like, given the nature of their design. I think it's possible, though, that we're going to start seeing more, like, application-specific L-2s that start to incorporate, like, well, application-specific transaction fee orderings, or which make it easier to add stuff like VDF receipts or do thresholds.

signing for ordering. I would say that like I don't think that there is any silver bullet. I would guess that there are probably like most applications probably have certain protocol layer like levers that they can pull to help mitigate the problem and find like more stable equilibrium. And for a lot of those applications, it'll probably like they'll probably need to be on L2 to do that. And this is related to, you know, like paradigm has invested in Cosmos and is somewhat interested in, and you know, it's obviously interested in that ecosystem for like largely similar reasons. It's just the idea of like MEV with generality is a very hard problem. And I think a lot of the experimentation that we're going to see happen is, you know,

sort of like more conforming. to the specifics of each application. Guys, one burning question in my mind that I have to ask, because, you know, as David and I say, a lot of this, this podcast is also for us. Like, we're on the journey. We're trying to understand, learn this stuff as well. So one thing I'm unclear about is if this is an Ethereum specific problem or, say, a general smart contract platform problem,

or to what extent it exists in a network like Bitcoin, say on Bitcoin's layer two, which is like, crypto banks, let's call it, inside of a Coinbase or a Binance exchange, like think of these as side chains, versus to what extent it exists in traditional finance. It just seems to be the case that maybe in kind of crypto banks or in traditional finance, you just can't see the MEV, but maybe it still exists. Is that the case? Or is this really a specific problem to you Ethereum smart contract and its design. I think it's a philosophical.

I think it's a philosophical question, honestly. I mean, yeah. We like philosophy. Why is it philosophical, Charlie? Well, because, you know, my bank, like, or broker or whatever elects not to front run me because the SEC is going to come after them, which, you know, you could call a, you know, crypto-economic incentive. I don't know where the cryptography comes in in that part, but whatever.

You know, we could call it that philosophical. philosophically. And you know, MEV like like most things, yeah, it has like fractal, you know, sort of analogous or fractal analogies in in other contexts. But I think that when we talk about MEV in crypto and like with respect to permissionless blockchains in particular, like it is actually a pretty narrowly scoped concern in that. the sense that like it's it's like not a thought experiment it is real it's like fairly easy to define you know what is and is not and I mean and I think that there is like a pretty clear difference to a conversation about for example the incentives of like you know an optimistic

role of sequencer on Ethereum versus like why Wells Fargo may or may not decide to front run me, you know, in their position. So yeah, I guess I would just say it seems to me more of like a practical concern within the context of permissionless systems in the same way that like, you know, custodial systems, permissionless systems just have different tradeoffs to custodial traditional systems. And like MEV practically is one of them. Yeah, I would agree with that. And that's kind of, you know, segues into like a new perspective I've taken recently on MEP after like a long time thinking about it. And that's basically that MEV kind of measures what happens in like sort of the raw economic

security case only, where the only thing that matters are economic incentives. And so yes, MEP exists in traditional finance, but it's not as meaningful because the incentive structure that's built there is not only these raw economic incentives. It includes all sorts of social and technical and political and legal and paperwork-based and human-based infrastructure processes that are outside the scope and that can't be modeled as easily, like Charlie said. So yes, you can measure what happens in like this technical system we built on Wall Street. If all laws break down and everyone's acting maximally maliciously, it's just that that's less meaningful than in the cryptocurrency case where the goal is to build a system where we're like encouraging people.

to pursue their rationality, and that's almost like enshrined as like, you know, acceptable no matter what. Phil, you're, you're getting dangerously close to rationality of self-defeating right now, my friend. Yeah, I mean, it relates to rationality is self-defeating in many ways, but I think it cuts through rationality is self-defeating. So because rationality is like self-defeating, this like simpler metric is like the only means of analysis that like matters. And kind of, you know, like there's obviously a difference between Bitcoin and Algorand or other honest BFT protocols inside an economic model in that like Bitcoin is more robust in a wider range of rational actor rational incentive models where there are fewer honesty and more economic assumptions. That being said, rationality is self-defeating. You can't perfectly model the world. And in many models, Bitcoin breaks down. For example, if you have two equally sized Bitcoin like blockchains, the analysis says nothing about that. And that's like a perfectly valid future. So yes, it is. It is a is self-defeating, but that being said, even though it is self-defeating, you know, somehow the systems we're building, we're trying to build things that are robust in more self-defeating worlds than not.

And towards that end, kind of looking at it through the lens of MEV is a useful tool to see, like, in which universes does it clearly break down versus not? And because we're building these kind of economically secured systems, that's a very immediately meaningful and, and like we just said, practical question. Oh, my God, this panel did not disappoint on on everything we've covered, in particular, that last question. Thank you guys. Super insightful. David, I think you've got the last one. Yeah, it's very, I hope it's obvious to listeners as to why this is such a critical conversation.

And I thank you guys for coming on and giving us your time. And also simply just being public servants or, yeah, public servants for public goods. Somebody has to fight this fight. And so I'm just honored that you guys are fighting this fight on behalf of, in my opinion, the whole entire globe. And so I kind of want to finish off with. this question. When you hear the term MEV, do you feel optimistic or pessimistic and why? And Phil, let's start with you. Ooh, that's a hard one. I don't know. I feel cautiously optimistic. The same way I imagine I would if I was observing like the early days of the internet, you know,

I think we have an opportunity. I think it's easily squandered. So it's very cautious. And I feel still a lot of existential anxiety about the way things will go in the space. And if everyone who's looking back on this podcast in 10 years is like, this guy screwed up and he's the new Blyde Masters designing like terrible credit fact instruments, like, F that guy. I'm sorry, you know. George, honestly,

I hope people, oh, I was going to say, hopefully like they come back and listen to this podcast. They were like, they weren't such bad guys. They gave it their best go. Georgios.

We are at M.E.B. Are you optimistic or pessimistic? I feel very optimistic. I think that the better world of tomorrow, firstly, this is like very philosophical. The better world of tomorrow is built on optimism and we need people to kind of risk working on hard problems like this because without these kinds of risk takers,

we wouldn't be where we are as a species. And secondly, I think that MEV right now, like what I care about, most fundamentally as an engineer, research, or investor, is working with the smartest people in the industry. And I know that the hardest problems in the industry attract the smartest people in the industry. Charlie, MEV, optimistic or pessimistic?

Optimistic. For this reason. The rationality of self-defeating argument or like the girdle's incompleteness theorem for blockchain's argument to me is incredibly intellectually attractive. It's something that I've spent like an individual. or an amount of time thinking about. And this is the argument again that for any set of quote unquote rational

crypto economic assumptions, you know, we can come up with that there is a way to violate and defeat them. Yet in practice, if you go to coin market cap.com, how many blockchains currently like exist and work and how many have been broken? There's a ton that work in practice. very few have been broken. I guess like Ether Classic had like a reorg attack, but like, I don't know, it'd be kind of hard for me to think of another. And so I think that like, I think of MED somewhat similarly, which is there is like I think a really deep philosophical concern that's worth taking

seriously and like obviously, you know, is being more realized in practice today. Like this is something we actually have to confront and we sort of have like the counterfactual that if we don't do something, the default state might not be successful and optimistic. But that being said, broadly, crypto has done pretty fantastically at succeeding even when we have an incompleteness theorem as to why. So I don't know. In the absence of certainty, I'm willing to take the optimistic perspective. Incompleteness theories themselves are rational claims that are self-defeating is what I would say. Yes. Yes, you can technically design a mechanism that breaks any rationality assumption, but like I think it's much more useful to look at the whole and the big

picture and like approximately what you're getting and like the leverage you're buying in various universes. And that's really more where I think MEV can provide leverage. I'm also optimistic because Phil is more articulate than I am. So I just defer to him. Guys, we're going to have the same panel back, and it's just going to be blockchain crypto philosophy, because I think heads would explode as mine did toward the end of this. I hope you've enjoyed this episode, bankless listener. I want to thank the panel for making this happen. Thanks a lot, guys. Thanks so much. Guys, what did you learn from today? Here's one thing I learned. M.EV is a pernicious problem. It's probably one of the biggest last unsolved problems in the world. the space and the stakes are extremely high. This is like the difference between do we create a better financial system for the world or do we fall flat on our face and just create a different financial system with a different set of actors? Super important here. This is not the last time. We will

be talking about MEV on bankless. I guarantee you. But there's reason to be optimistic and I'm leaving this conversation more optimistic having spoken to these panelists today. Of course, course guys. Got to close with risks and disclaimers. ETH is risky. Crypto is risky. DFI is risky. You could lose what you put in, but we are headed west. This is the frontier. It's not for everyone. But thanks for joining us on the bankless journey.