Stephane Gosselin, cofounder and chief architect of Flashbots, and Uri Klarman, CEO of bloXrouteLabs, joined Unchained to discuss everything about MEV, an especially important topic now with the Merge coming.
- what MEV is, how it works under Proof of Work
- some examples of MEV, like frontrunning and arbitrage
- how MEV is a rabbit hole full of monsters
- the role of Flashbots in the MEV industry and how Flashbots enabled a reduction of gas fees for users
- how MEV-geth functions and how it relates to auctions systems
- how bloXroute, Uri’s company, is working with MEV
- where Stephane and Uri disagree on MEV, despite agreeing on many fundamental things
- how MEV changes with the implementation of Proof of Stake
- what MEV-Boost is and the problems it solves
- what proposer-builder separation is and the motivation behind this idea
- whether DeFi would flourish more if there wasn’t frontrunning and why MEV-Boost doesn’t prevent it
- what externalities MEV-Boost or proposer-builder separation (PBS) would solve for users
- how MEV exists in other industries and in traditional finance
- why Flashbots’ own investors would rather trade on a fair-sequenced rollup than on Flashbots
- how the number of MEV extractors has been decreasing and whether this would end up centralizing MEV
- whether exploiting certain types of MEV is illegal
- how other chains are trying to solve frontrunning, like Osmosis with threshold encryption
- what are the levers that can be pulled to solve MEV, according to Uri
- how eradicating MEV can positively affect the price of ETH, and by how much
- whether validators can potentially share MEV profits with users
- how to determine if a network has a good MEV supply chain
Thank you to our sponsors!
Ava Labs: https://www.avax.network/
- Twitter: https://twitter.com/uriklarman
- Twitter: https://twitter.com/thegostep
Previous coverage of MEV on Unchained:
- The Debate Over Ethereum MEV: Should It Be Minimized or Boosted?: https://laurashin.bulletin.com/the-debate-over-ethereum-mev-should-it-be-minimized-or-boosted
- The Other Side(s) of Ethereum’s MEV Argument: https://laurashin.bulletin.com/the-other-side-s-of-ethereum-s-mev-argument
- What is MEV: https://ethereum.org/en/developers/docs/mev/#:~:text=Maximal%20extractable%20value%20(MEV)%20refers,of%20transactions%20in%20a%20block
- Proposer-Builder separation: https://www.alchemy.com/overviews/proposer-builder-separation
- Uri’s blog post: https://medium.com/@uri_61495/flipping-the-table-on-the-mev-game-dc31df8baaf7
- MEV Profit sharing validators: https://www.recvc.com/mev-2-0-the-rise-of-mpsvs/
- Panel conversation mentioned in the episode (39:53): https://youtu.be/UKfBFV1TuEM?t=2393
- Sealed auction design: https://ethresear.ch/t/mev-sgx-a-sealed-bid-mev-auction-design/9677
- Priority Gas Auction (PGA): https://www.mev.wiki/terms-and-concepts/priority-gas-auctions#:~:text=What%20is%20a%20priority%20gas,up%20fees%20for%20other%20users.
- Dune’s dashboard – amount of MEV extractors: https://dune.com/queries/1171509
- The MEV Game of the Crypto Economy: Osmosis’ Threshold Encryption vs. SGX of Flashbot?: https://mirror.xyz/infinet.eth/SFjR1H1-RMnKoIoPjqkxpauVPrLYGqLHQP1dY9FHvx4
- Bank of International Settlements report on MEV: https://www.bis.org/publ/bisbull58.pdf
- Flash Boys 2.0 paper: https://arxiv.org/pdf/1904.05234.pdf
- Flashbots Dashboard: https://dashboard.flashbots.net/
- Flashbots auctions: https://docs.flashbots.net/flashbots-auction/overview
- Flashbots relay open source: https://twitter.com/hasufl/status/1559908331145265156?s=20&t=gtHe5sVetxQXVKfZL2VMYg
- MEV Boost: https://www.alchemy.com/overviews/mev-boost
- MEV in ETH PoS: https://writings.flashbots.net/writings/mev-boost-call-for-testing
Hi, everyone. Welcome to Unchained. Your no-hype resource for all things crypto. I’m your host, Laura Shin, author of the Cryptopians. I started covering crypto seven years ago and as a senior editor at Forbes was the first mainstream media reporter to cover crypto currency full-time. This is the August 23, 2022 episode of Unchained.
I’m looking for a part-time evergreen editor for Unchained. If you’re skilled in SEO and love blockchain and crypto, email your resume and writing samples to email@example.com with the subject line evergreen editor application.
With the Crypto.com app you can buy, earn, and spend crypto in one place. Download and get 25 dollars with the code Laura, link in the description.
Harness the full power of the Avalanche network with Core, your new Web3 command center. Built by Ava Labs, Core is more than just a wallet. It’s a noncustodial browser extension engineered for users to seamlessly and securely experience Web3 like never before. Explore Avalanche dApps, NFTs, bridges, subnets, and more today.
Whether you’re crypto curious or a C-suite decision-maker, you have to check out Web3 with a16Z, the chart-topping technology podcast about the future of the next internet. Listen to Web3 with a16Z on Apple Podcasts, Spotify, or wherever you get your podcasts.
Today’s topic is MEV, which up until recently was called miner extractable value and is now being called maximal extractable value. Here to discuss are Stephane Gosselin, Co-Found and Chief Architect of Flashbots, and Uri Klarman, CEO of bloXroute Labs. Welcome, Stephane and Uri.
MEV is a pretty controversial subject, as I discovered when I started diving into it, but before we get into all the concerns and arguments about it, let’s just make sure people really understand what MEV is, how it works, etcetera. Right now, we’re in a phase where Ethereum has a proof of work consensus algorithm. So, why don’t we just talk about what MEV looks like right now under proof of work. Stephane, do you want to start with a description?
Sure. Yeah. So, I mean as you mentioned, MEV is a very deep rabbit hole. Extremely deep rabbit hole full of monsters and unexpected findings. But I mean what’s been really interesting about MEV and in particular I guess the narrative and the meme of MEV is how it’s sort of come to mean something different to different people, and it’s a narrative that’s been sort of picked up by a variety of different users of blockchains to help advertiser protocol try to explain what they were doing, try to say what’s wrong with other people’s protocol. It’s just become this really deep and interesting technical topic.
In terms of the Flashbots perspective on MEV, we saw MEV as sort of an economic force that came to be with blockchains but exists outside of it. In particular, our observation was that this economic force came from the roles that certain actors in a blockchain have. In particular, miners at the time when the research came out. So, we explored what are the different activities that miners are able to take with regards to the powers that they have in the protocol, and how can they use those powers to be able to maximize the amount of value that they generate for themselves.
Practically, a lot of MEV means ordering censorship or inclusion of transactions and so the power of miners to be able to create blocks and decide which transactions go in the block also means that they have sort of a monopoly on the execution of the EVM over a given block that they’re assigned. So, the question for them is how do I maximize the amount of value given this power? The Flashbots sort of introduction to the ecosystem and the way that the MEV sort of industry has evolved on proof of work is for miners to outsource it to searchers. So, third parties who come up with strategies on how to order transactions, how to insert their own transactions and then compete for the inclusion into a block.
Uri, can you elaborate on that by giving people some concrete examples of how MEV works and the types of transactions that happen?
I was just about to say, coming from slightly a different angle, if you think about a regular user, a regular user buys a Uniswap or makes some DeFi transaction, and you could think that if somebody’s about to buy a lot of ethe, the price is going to go up. When DeFi started to get big, traders figured out they can see other people’s transactions and front run them, get to their transaction before a major transaction happens. Kind of like capture value by buying it cheap just before the price goes up. So that’s kind of like front running is one piece.
A second piece is there is a price up ethe on Uniswap. Just before the block comes out somebody sees that the price on Coinbase just crashed, right. So, they’ll try to capture arbitrage. They know the price is actually different from what’s happening on Uniswap, and they’ll try to get their transaction in to capture that arbitrage. So, traders have been doing that when DeFi started to get serious and then they started competing who would be the fastest and who would be the one to pay the most. So, this is where it moved to MEV to miner extractable value.
People started to say if I’m making a thousand dollars, I’ll pay 500 of it as tip for the miner or the validator. It doesn’t really matter which protocol, but I’ll pay it as a tip and then they’ll put my transaction first before somebody else. So, that led kind of like to the gas wars that kind of were pre 1559 and pre-Flashbots. So, traders started to kind of try to compete with one another capturing these opportunities. In my mind, MEV more than anything is basically front running and capturing arbitrage. These are the two big pieces, as I see it.
Once things have evolved, then miners realized, wait, instead of actually…they started by maybe doing under the table deals and all this kind of stuff. Flashbots came out and said how about we make it public? How about we allow everyone to participate and it’s miner extractable value because you could think whoever is really good at extracting value can pay more and get their transaction executed, and it is the miner that’s being paid for it.
So, if I can make 100 dollars on an arbitrage but Stephane is really good at it, he can make 150, then he can offer 101 dollars and his transaction or his arbitrage or his front runner will get in and mine won’t. This is kind of like where things currently stand where miners are the ones making most of the money. They’re not actually doing it themselves. Other people are kind of like professionalized and kind of found the expertise in capturing this value. Bribing or paying the fee or however you’d like to call it to whoever construct blocks. It’s validators in POS. It’s miners in proof of work. That piece doesn’t really matter. That’s kind of like where things stand.
Earlier you talked about how this happens in DeFi. Does this also happen with things like NFTs or what about single payments?
So NFTs for sure, right. It’s the same story if a big NFT drop is happening. Everybody wants to be the first one to capture it or maybe mint a million, I don’t know, kitty cats or whatnot, etcetera. Because of the financial implications of the NFTs, that game is also being played. But if I’m just sending one ethe to Laura Shin, then this transaction, there’s no arbitrage, there isn’t front running, there isn’t anything like that. That transaction would go through one way or another. It’s not inside this game. The MEV game is very much about DeFi and NFTs and it’s less about somebody making a payment from one entity to another or I deploy a smart contract which does something, which isn’t that valuable at this time or there is no value to extract from it. Then that’s the regular game.
It’s probably worth mentioning that we at bloXroute have been in the game for quite a while. When Flashbots came out, I was actually slightly worried from the implications of that, but Flashbots actually did a really solid favor to the community by siloing this competition. If we had the gas wars and people tried to outbid one another, basically driving out all normal usage, right. My normal transaction won’t go through if I’m not paying really tons of gas and really high fee, although I really shouldn’t.
So, at the time, this was prior to 1559, etcetera, the immediate result of Flashbots siloing the MEV game and kind of like separating it from regular usage was that regular users were no longer forced to compete in this game. So, the immediate result of Flashbots was actually a very positive thing. I was wrong on that. I was actually concerned about that. That taught me to be like I have my own opinions. I’m a very opinionated guy, but we should also all kind of like, well, I might be wrong what my opinion is.
Uri, just to clarify, are you saying that the immediate benefit for users was lower gas fees?
Yeah. When Flashbots came out that was the immediate result. Basically take all the gas competition, put it silo. They’re competing against one another and only the best one is included. Everybody else is not really affected.
Just to quantify this for users, you wrote a blogpost recently where you said that MEV costs users 160 thousand eth per year. That’s roughly 300 million dollars and about 50 percent of that is front running.
That’s kind of like taking it from…the immediate result of Flashbots was kind of like that affect. On the other hand, I think front running is bad. Basically, if a user makes a transaction and somebody else comes and this user gets a worse trade then the outcome is friction. It’s not value. It’s not economic activity. It’s that me trying to be a thousand eth cost me, I don’t know, 1.1 million rather than one million. That kind of thing. I’m trying to look for ways to capture all the good and the upside of the MEV in the system, kind of figuring that piece out without hurting the users, which is hard.
Okay. So obviously we’ve already been talking about Flashbots but it’s been Uri talking about it. Stephane, why don’t you explain Flashbots’s role in working on MEV.
Okay. I mean there’s a few ways I could go about it. One of them is to try to untangle some of the things Uri talked about. But maybe instead I’ll talk about how we see our own role in the ecosystem. So, Flashbots emerged as a research collective, so it was based on the idea that we see this MEV activity having negative externalities on the network. There’s a lot of spam in the transaction pool. Regular users are ending up paying more and more volatile gas prices. Is there a fundamental flaw in the way that the transaction pool works that causes this? That was the initial sort of approach.
So, we developed a solution for this, which ended up being called the MEV-Geth, and it is that the transaction pool was not expressive enough for certain types of users of Ethereum. Certain types of users had preferences on the specific position in the block that they had that the only way they could express that preference was by causing spam or by optimizing for latency or by basically creating some externality to everyone else in the network. So, our view of MEV is…MEV is a complex, systemic problem for blockchains and solutions, depending on how they’re implemented, have some externalities. So, we research what are the externalities of various different solutions and how can solutions be designed to maximize the alignment with the objectives and the principles of the underlying blockchain?
So MEV-Geth is a version of the Go Ethereum client, which is I think still the most popular client that essentially enables these MEV auctions to happen in a separate place rather than, as Uri said, right in the mempool or right on the blockchain. Is that a fair characterization?
Correct. So, it replaces what was the most common strategy, PGA, which has a latency game. Price gas auction and it replaces that with just a sealed bid auction. So, instead of leaking value to investing and to low latency strategies, instead you capture that value in the auction.
Okay. Now, Uri, how is bloXroute working on MEV?
So, bloXroute is a service provider. We’re a networking company. We allow people to see transaction fast, to see blocks fast, to act fast, which for competitive DeFi actors and anybody in this space, basically you want to see any transaction happening before it’s added to the chain because if it’s going to affect the price you want to know how it affects the price, what’s happening at any given moment, especially if you’re a large market maker, if you’re a major trader playing both the DeFi game and the CeFi game. So, that’s at the core of our business.
When MEV became a thing then many of the MEV actors are actually our customers. So, all the MEV researchers want to see everybody else’s transactions. They want to see the blocks as fast as they come. They want to make sure that their transactions or their bundles go directly and ASAP to all the mining pools because they want to make it to the next block, etcetera. So, we are more like picks and shovels as a company. Then when Flashbot came out, they’re running the Flashbot relay, which kind of like connected searchers to the mining pools we’re running. Also, Flashbot relays, kind of like a private version of that, that people pay us to get faster latency, etcetera.
I think what Stephane says, like some transaction has preferences they want to express I think that’s really high level of saying somebody wants to be first to capture arbitrage or to front run another transaction. Not always but almost always or just come after a transaction because it creates some opportunity, which is usually still arbitrage. I think maybe the point where me and Stephane kind of disagree is that I see it as we want to minimize externalities and figure out the affect on gas was very easy to see. So that’s definitely a point to Flashbots’s credit, but what are we paying here? When we define the system in such a way or we build the system in such a way, what’s the price that’s being paid?
If we say that always Ethereum is optimizing for MEV, so every transaction, all the value that can be extracted will be extracted, so it’s really hungry and attempting actively to kind of like capture anything left behind it. What does that mean for users? If I’m a regular user and I’m not, I know high frequency trading firm and whatnot, etcetera, and I see a price on Uniswap and I try to buy there, but I allowed for some slippage because maybe the price would slightly change. I would always get the worst price or I would always get front run. There are some DeFi projects.
I spoke with somebody interesting not long ago. Basically, automated bots that kind of like adjust their positions based on oracles, etcetera. But they can’t be fully automated because all the MEV searchers, oh, here’s an oracle transaction. This is how the positions are going to be adjusted. They always act before this product will work and that product will always get the worst price.
So, this actually affects more than just the regular user that kind of made the transaction. It also means what kind of product we can or can’t build in DeFi, and the question is what’s the tradeoff? What do we want to enable? What do we don’t want to enable? How much it’s benefitting for users? How much it’s giving them a hard time. That’s maybe the point where me and Stephane might not see it eye to eye.
So before we get into the controversial aspects, which I definitely want to get to, let’s make sure that we explain how MEV is going to change as Ethereum moves to proof of stake next month.
Yeah. I’m happy to talk about it. Obviously, proof of stake has been a long time in the making, but at Flashbots we try to keep up to date with as many research topics as we can. We saw that around about last summer that a lot of the core developers are stating to ramp up and focusing on the merge as being the next major release of Ethereum. Immediately after the London hard fork.
So, we started engaging and figuring out, okay, the easy thing to do would be for us to continue shipping the same software that we already have, MEV-Geth, for a proof of stake environment, but we thought we could do better. So, we engaged with various different stakeholders including Ethereum foundations, validators, core devs, client development teams to try to figure out is there something that we can do to improve the current situation of MEV on Ethereum.
The biggest concerns that they had were basically twofold. One was client diversity and the second one was solo-staker participation in MEV. So, one of the big limitations of the way that MEV-Geth works is that it requires miners to see the full content of the transactions that they receive. So, that means that you cannot send these transactions to every single miner. You can only send it to mining pools that have some trust involved in that you can monitor how they operate with the transactions that they receive. Matching that over to a proof of stake world, it would not have been possible to send these transactions over to solo validators.
In order to mitigate this, we basically developed a new piece of software from the ground up, right, called MEV Boost. That aims to sort of solve for these key concerns. So, it’s a neutral piece of software. It’s done sort of in an open source, public good way, and it’s software that any validator from the largest staking pools to solo validators at home can run to get full access to MEV.
One other component that will be part of this is something called Proposer-Builder Separation. Uri, do you want to explain what that is?
Sure. Although I think me and Stephane should switch roles because I think Stephane would actually have…I would explain the first a bit better and he would this one. I want to add something to what Stephane is saying, which I think is important, especially for the greater community. You can figure out where we’re going and what are the implications. What MEV boost does is that it allows any validator out there to receive sealed blocks. Okay. So just like they don’t know what’s inside the block. They receive this sealed block and actually multiple potential sealed blocks, and they choose who pays them the most. So maybe one block builder…block builders are special MEV entities that try to maximize how much MEV they can capture.
They’re each trying to capture as much MEV as possible. Again, similar to the previous setup, trying to outbid one another, who would pay the validator the most and so the validator gets several…hopefully several competing potential blocks and it chooses which one to actually sign blindly and then the block propagated either by that validator or by the relay but basically signed it blindly and then it goes through the network.
Which basically means, and I think it’s a big deal that blocks are constructed to optimize MEV. If something improves them out of MEV it will be included. If it’s not then it wouldn’t be included or lower priority. It’s a major shift, again, from the perspective of the users, as I see it. Knowing that this is how blocks are constructed is a big deal. Now, MEV boost is a step towards PPS. It’s basically the idea of separating the proposer, the validator which proposed a new block to the other validator from the builder, the one who constructs the block.
I think it was Vitalik like two weeks ago in a conference was saying that the motivation to separate these entities is that a small…if there isn’t a separation, so think about the validator, try to optimize and extract as much MEV. You could imagine that a solo validator would earn much less proportionately compared to a giant liquidity pool with tons of resources, right. The solo validator would either construct a vanilla block or would use some open source tool that would help him, while the very large entity will have professionals trying to come up with strategies and construct a better and more optimized block, which means that larger validators would be better off.
Okay, so a validator which is 100 times bigger would make more than 100 times more money. Maybe a thousand times more money. I don’t know the math. So, by separating the validator and the constructor of the block, you’re preventing the economies of scale affect of this off-centralization. So, if larger pools are more profitable, it makes sense that they will be more dominant and this is a centralization.
If you split the two pieces, the actual validation of blocks, and the construction of blocks then the economies of scale and centralization power stays within the MEV block building space, while validators are all the same, whether you’re small or you’re big. You’re going to make the same amount of proportional money. If you’re ten times bigger, you’ll make ten times more money but not a thousand times more money because somebody else will be doing this work, will be providing them the same options.
So it sort of sounds like you feel like some parts of this at least are not a good move. What would you prefer to see happen?
So, I feel there are issues here. Me and Stephane are actually kind of like DMing about it the other day, etcetera. Stephane mentioned the externalities around it, okay. We’re doing all of this to mitigate negative externalities. The question is what are these negative externalities? What good are we doing here? How are we saving Ethereum or we’re improving it? And what is the cost and kind of balancing between the two. The downside of PBS and the downside of any MEV-boost, etcetera, as I see it, is that it optimizes to maximize MEV. Now, I’m totally okay…personally, I feel totally okay that arbitrage opportunities will be captured by somebody for the system to be efficient. It’s not bad MEV. It doesn’t hurt anybody. There is inefficiency and whoever fixes it profits, and I don’t see any reason why that wouldn’t be the validator. It’s actually a good thing, but optimizing for front running and enabling that, so trading is not somewhat adversarial. It’s like the most adversarial. You’re never going to stand a chance. You’re always in the worst position you could ever be in. That is a cost to the users and the system. So, you want DeFi to flourish. If there wasn’t front running, DeFi would flourish more, obviously.
So, this is the cost. The question is what’s the upside? I think this is more of a question Stephane. What’s the benefit of doing it and is it worth?
Yeah. Stephane, I’m curious, did you guys ever look at constructing MEV-boost in a way where front running would not be allowed in some fashion? I don’t know how you would implement that, but kind of what’s your take on the parts of MEV that are harmful to users?
I think it’s difficult to have a conversation under this framing. You know, a lot of the categorization of MEV that Uri has been presenting I think is inaccurate and doesn’t represent how MEV works from research and protocol development perspective. I think that there’s very complex implications for MEV across the application layer as well as the protocol layer. The most important way to think about it is through the frame of where is the value accrual going and how are the strategies for capturing that value impacting groups.
I think there’s also some correction I’ll put to the way that MEV-boost as software was framed. So, MEV-boost does not inherently maximize MEV in any way. It’s actually completely unopinionated. The only thing that it does is it allows validators to outsource the way that they construct blocks to third parties. The approach that those third parties use is completely detached from the MEV-boost software itself.
So, asking a question like why doesn’t MEV-boost prevent front running is, from the protocol developer perspective, just not really a coherent question. There’s a few undefined things here. So first of all, front running itself isn’t a defined concept within the realm of blockchains. So, you have such a thing as a public transaction pool, which some bot operators pay people like bloXroute to get preferred access to, and then they use a transaction that they receive from that channel to then insert other transactions ahead of it and then submit that to miners for inclusion. So, I think those are the types of strategies that Uri is trying to refer to as front running, which is really just a subset of the MEV strategies that are out there. Thinking about how do we mitigate those kinds of strategies is a difficult topic to talk about, and we can get into that, but the sort of perspective of the MEV-boost is it’s not meant to be a layer for solving this kind of strategy or preventing those kinds of strategies. It’s purely just meant to be a layer for allowing validators to outsource and get access to the MEV.
Hopping in for a second, first of all, you’re right. MEV-boost allows…you give it to third parties, not necessarily to whoever pays the most. I totally agree with that. But I remember we were speaking briefly on this in Paris. We were talking about front running versus arbitrage versus other, and you said I disagree with that’s the definition. So, I’m not trying to kind of throw you and phrase a question in a way that looks bad. Rather, I know you have a different opinion about it. I’m honestly interested to get your take. I think arbitrage is kind of like, okay, and front running is bad and we should try to mitigate it. I know your opinion is different. I would very much like to kind of better understand that.
Sure. So, these are very deep research questions on which even I don’t have the best perspective. So, Flashbots has a whole research team that’s focused on answering these kinds of questions. They publish a lot of useful documentation on trying to classify different types of MEV. For anyone that’s listening to this and interested in checking those out, you can see writings on Flashbots.net, and that’s where a lot of the preliminary research and research articles that we post go.
With regards to the ethical side of front running and arbitrage and just types of strategies out there, we have an upcoming post that’s going to be on that topic that’s going to be published maybe by the time this episode goes out. It’s going to go into depth about those different classifications. So, I wouldn’t want to…
Spoil the surprise.
Before it gets posted.
And what about the externalities? For me, it’s super interesting. It’s kind of like what are the externalities that really you think MEV-boost is solving for? It’s kind of like one thing, basically allowing everyone to participate. That’s super easy and kind of understood. What other…again, I asked you something and you said, no, that’s not really what we’re solving for. So, what are the externalities on the chain and on the user who we’re solving for with MEV-boost or PPS?
Yeah. I sort of mentioned the design goals behind MEV boost are to allow for solo validators to participate in MEV. Then also to keep promoting sort of client diversity. Both of those things are…maybe I’d say the client diversity one is mostly just an architectural technical one, but allowing solo validators to participate is perhaps the main one that has implications for the way that the ecosystem of MEV develops, which is you want them to be able to extract at the same rate, the value of MEV as even the largest staking pools and node operators can.
So, it eliminates the economies of scale that comes with being a node operator in a way that encourages even solo-stakers to participate in the network. That’s really the main objective with both PBS and MEV-boost.
Now, one thing that I want to specify is PBS itself sits on a much larger roadmap of Ethereum. So, whenever we talk about these technologies, about these protocol features, they come with, one, a lot of unanswered research questions, and two, dependencies on other parts of, for example, the Ethereum roadmap to be able to sort of fit together in a cohesive puzzle. So, while PBS may solve the ability of solo-validators to permissionlessly participate in MEV, it doesn’t solve other issues that Ethereum may have.
All right. So, we’re going to dive a little bit more into the controversy around this, but first, a quick word from the sponsors who make this show possible.
Curious about the world of crypto and the future of the next internet? Then check out Web3 with A16Z, the chart-topping technology podcast from the minds at Andreessen Horowitz, the go-to destination for discussions on tech as it changes our world.
Whether you’re a crypto-curious person looking for signal versus noise in the day’s headlines or a C-suite decision maker seeking to understand Web3 as part of your business strategy, Web3 with A16Z is the podcast for you. Tune in each week for leading insights from the top scientists and makers in the space through carefully curated conversations with acclaimed podcast host, Sonal Chokshi, former showrunner and long-time host of the A16Z podcast, along with frequent guest appearances and hosting by Chris Dixon. Listen to Web3 with A16Z today on Apple Podcasts, Spotify, or wherever you get your podcasts.
Join over 10 million people using Crypto.com, the easiest place to buy, earn, and spend over 150 cryptocurrencies. Spend your crypto anywhere using the Crypto.com Visa card. Get up to eight percent cash back instantly, plus 100 percent rebates for your Netflix, Spotify, and Amazon Prime Subscriptions. Download the Crypto.com app now and get 25 dollars with the code Laura, link in the description.
Is your Web3 experience hindered by inadequate crypto wallets and browser extensions? Ava Labs has created Core, a free, noncustodial browser extension engineered for Avalanche users to have a more seamless and secure Web3 experience. The best-in-class Avalanche Bridge now offers native support for the bitcoin network. Put your bitcoin to work in the robust DeFi ecosystem by bridging BTC to Avalanche today. With Core you can also easily swap assets, display your NFTs and style, store your assets in a ledger-enabled wallet, and put real dollars into your crypto wallet in just a few clicks. Core has everything you need for a simple, secure, and convenient Web3 experience. Download the free Core browser extension from Google Chrome’s Appstore today.
Back to my conversation with Stephane and Uri. So one question I had about what you were saying before about how MEV-boost doesn’t do anything on its own to maximize the amount of MEV being extracted. As far as I understand, it at least makes that aspect of it competitive, meaning builders will propose…
You’re right. It’s builders propose. It’s confusing.
So they’re creating these blocks and they’re doing it in a way where they’re trying to make it something that is profitable for the miner and for themselves, right. So the more value that they can extract then the better rate they can offer the miner and the better profit that they can give to themselves. So, in that aspect I do think that PBS or MEV-boost does kind of contribute to this, right?
Yeah. Here we’re getting more talking about what is inherent about MEV and less about the specifics of the MEV-boost software. So, one way to think about MEV is like this free energy that’s created by applications, wallets, and users. They create transactions in such a way that they’re giving the power to someone else to be able to extract value. The question is, okay, what is the impact of this and how are third parties able to extract that value? That’s really sort of where the value originates. So, someone is going to be able to see, okay, there’s an opportunity here that I want the capture. How do I go about capturing it?
In the full MEV-boost system, these are searchers. So searchers see some value capture opportunity here. They submit it to a builder. A builder aggregates all of these bundles, and they produce a block. Then they submit those blocks to validators for inclusions. So, the reason why you want a system like this to happen efficiently is because it puts all of each step of value capture in a public auction, a public market. So, the amount that a builder pays to the validator for the block becomes publicly known to the entire network. There’s no longer sort of these special relationships, these special deals where a builder can agree with a validator to say I will produce all of your blocks and we will split the revenue 50/50. Instead, you have this public market where all of the data and the payments become transparent, and it just brings a lot more light to this entire supply chain or this entire industry.
Okay. So, you’ve clearly delineated some of the benefits. When I asked about MEV on Twitter, the vast majority of the responses were kind of negative to the concept. As we’ve been discussing, being able to extract MEV is actually becoming more baked into Ethereum rather than something that’s being avoided. Here’s just a few of the comments. One user, Harry Popper Rosseau wrote, why, as an industry and community, are we not focusing more on prevention instead of just accepting it as something that will always exist? Bitcoindadpod tweeted, why would anyone knowingly consent to participate in a system with MEV? Why not just stick with the fiat system? So what are your responses to this?
I think it’s very easy to look at MEV on the surface immediately and say can’t we just get rid of it? Isn’t there something that we can do? Isn’t there someone that can prevent this from being…
Devs do something.
Exactly. The answer is it’s a lot harder than it may look immediately. So, MEV exists and like exists in the traditional financial system as well, and it’s the reason why you have such large financial intermediaries that are just dominating and doing a lot of regulatory capture. So, it’s not like it doesn’t exist elsewhere. It exists elsewhere. Just the ways in which it’s dealt with are maybe not as transparent and maybe the public is not as aware about where the MEV resides.
So, step one is just making sure that people understand how the system works, people understand where the value goes, and there are definitely a bunch of different solutions that can be developed for dealing with this value capture. One of the ways that Flashbots is sort of a strong believer in is in reducing the amount of MEV that actually gets exposed.
So, a lot of the MEV on Ethereum is exposed from the way that the applications are designed. So, Ethereum…I mean Uniswap is a great example, and it’s sort of the one that was used in the initial explorations of MEV. Uniswap works by exposing a ton of slippage for its users, but Uniswap could also decide to design an alternative system that doesn’t allow the users to expose that much slippage.
Similarly, wallets could also figure out how to produce transactions and send user transaction in a way that doesn’t leak as much value from their users in a way that is capturable by other people. So, one of the important things about like, oh, wow, MEV is bad. If only we could solve MEV. Well, users have the choice. They choose which application that they want to use. They choose which wallet that they want to use. The reality is a lot of them continue to choose wallets and applications that expose a lot of value that they could capture. But instead, they send that value to bot operators and ultimately to miners. So, the Flashbots’ job is like what do we do with all this value that users and applications have decided to expose? The most stable thing you can do is submit that all the way down the chain in a transparent way to the validators of the node.
I actually also want to ask about one specific comment that someone made where they pointed me to a talk that was given by people. It was a panel. Vitalik and Hasu and people from Paradigm. At one moment, they were asked would you rather trade on a Flashbots version of Ethereum with MEV-boost of MEV baked in or a fair, sequenced rollup, which is just what it sounds like. It’s not something where the transactions have been reordered in any fashion. He pointed out, and I did watch the segment, that even your own investors, people from Paradigm, would have chosen the fair sequence rollup. What is your response to that?
I mean I can’t speak for the people at Paradigm, and I don’t know the exact situation involved, so it’s probably best to ask them what they meant by that. Maybe they didn’t understand exactly how the system works. I’m not sure. Maybe they have some other interest in mind. I think it’s worthwhile to ask them.
So you wouldn’t go with a fair sequence rollup? You would choose something with the Flashbots method?
I don’t think there’s any…I mean there are some solutions that I think approximate what a good trading experience would be on Ethereum today. The batch auction sort of approaches to building decentralized exchanges is actually quite good, and helps mitigate a lot of the externalities. So, I think it is a solvable problem at the application layer. It’s not necessarily something that needs to be solved at the protocol layer. That is sort of compatible with any chain. It’s a system that you can build that helps find better prices for ender users of decentralized exchanges.
That’s not dependent on the protocol in which it’s built, which is always a good way to build these kinds of applications. So, would I prefer to use Flashbots versus fair sequencing, to me isn’t really the question. The question is would I prefer to use a decentralized exchange that gives me the best price or one that exposes value to someone else. My answer is always going to be the one that gives me the best price.
I also was wondering just kind of if we’re going to look at sort of other consequences of implementing this more fully into Ethereum, I saw that the number of MEV extractors has decreased over the last two years. It went from 80 down to about 27 now. I wonder, as we get into this space where it’s sort of more baked in, won’t over time the larger extractors simply become more competitive, therefore winning more auctions, eventually leading to centralization? Or if this assumption is incorrect, what do you think will happen or why have we already seen that the number of extractors has been decreasing?
I think it’s an important discussion to have. I will ask is that based on the Flashbots Explore numbers?
It was a Dune XYZ.
Yeah. I think it’s difficult to know exactly the number because as strategies get more complex it becomes harder and harder to define exactly what is MEV strategy and what isn’t. So, we can look at a specific strategy like atomic arbitrage on Ethereum, and we can see that over time there are fewer parties that are sort of excelling at it. I think that’s normal. For a given strategy space, you’ll have specialization over time and certain parties will be able to sort of reliably capture more opportunities. The ones that aren’t and eventually follow are the market. So, it’s interesting to look at it as a pure strategy space.
Definitely I think MEV, just like every other industry, sort of as it matures you sort of see very clearly who are the winners and who aren’t able to compete. The goal, I think in MEV and any other industry should be to foster maximum amount of competition. So, you want to be able to have a system that trends towards more pure competition and low barriers to entry for new entrants rather than something that trends towards being more monopolistic where a single party is able to capture all of it. That’s perhaps the conversation that becomes more interesting is does having only a few parties acting monopolistically have any externalities. Flashbots’s perspective is, yes, definitely it does. So, that’s something that we want to avoid and we want to maximize the amount of competition.
Another question that I have, which is highly relevant given what happened with Tornado Cash recently is that, again this was suggested on Twitter by Rebecca Rettig of Aave, and she mentioned that the Bank of International Settlements recently wrote a paper on MEV. In it they said in most jurisdictions, activities such as front running are considered illegal. In the paper, they suggested that exploiting MEV and blockchains could be illegal in those jurisdictions. I also saw a paper published last week in SSRN by two researchers at the University of Surrey. They assert that sandwich attacks, which is having a transaction that pushes up the users price and then a transaction on the other side that is able to benefit from that, profit from that. That they run a significant risk of being found to violate the chief anti-fraud provision of the Commodities Exchange Act as effectuated by CFTC Rule 180.1.
So, what do you guys think about this potential illegality of exploiting at least certain types of MEV, especially as we’re seeing these kind of freak outs about what the sanctions might mean for Ethereum at the protocol level?
First of all, I always take with a grain of salt everything coming out from the International Bank of Settlement. Not even regarding to this. This might actually have a point, but as a source for whose announcements we should be listening to. I think this, and especially with the Tornado Cash and everything around it, actually brings us to some serious questions we need to answer. Also, in the context of MEV and front running in general, what do we as actors in it, what do we do? Specifically Tornado Cash is a complicated question.
If bloXroute is a block builder and we’re trying to add transactions and construct them into a block and give that to the validator, and that block includes a banned transaction, an OFAC banned transaction, are we facilitating that transaction? Maybe the answer is yes. Maybe the answer is no. We think the answer is no because we’re helping the validator order the transaction based on their preference. They can accept it or not accept it, etcetera. But it’s definitely a major legal question that everybody in this space were looking into. What are the implications of both for major staking pools whether they are Coinbase and exchanges, whether they’re Lido. Do they comply? Do they not comply? Do they include build blocks with said transaction? Do they not?
Yeah. But I want you to answer these questions for MEV.
It’s not as serious as OFAC just because of the implications of it and how…is this national security or is this like a slap on the wrist? But the same goes here as well. It’s kind of like if you’re participating in it, are you in fact breaking the law in your jurisdiction? There are two parts here. One of them is completely legal, which I think both me and Stephane are not the people to actually answer about it. It’s kind of like, well, it really depends when you create the system. Is what my entity doing, is that participating or not participating. If I’m accepting from somebody else and doing front running, am I participating in it?
Also, front running isn’t necessarily the right term for it. We just borrowed it from a different field. So, it’s not exactly the same. From my perspective, it’s more like I think Flash Boys, the people who connected New York and Chicago. If you’re one millisecond faster than everybody else and you manage to reach price crashes in New York, you’re the first one to reach Chicago and sell before everybody else. Are you front running? There are legal questions around that. I think the more interesting question, especially for people like us, actually not the legal ones. It’s actually the ethical ones. Like is this a bad thing that we’re doing here? Is this a bad thing we’re enabling here?
It might be a bit of a non-answer because it’s a legal question. Me and Stephane are not lawyers, but I think we disagree on is it a good thing, is it a bad thing. If Stephane calls MEV like it’s energy out there that MEV actors or others could extract. I see it slightly differently. If somebody makes a transaction and he wants to buy eth for a thousand dollars, and the price of eth is a thousand dollars but 15 seconds later the prices has changed. The price went down to 900 dollars and now, oh, look at this sucker transaction. I can buy, sell him for a thousand and then sell after him and capture the value that way. Is that a bad thing? Saying that the application layers to just solve it, I’m not sure I agree with that. Maybe some applications cancel some of it, but if we’re talking about Ethereum, what should Ethereum as the base layer do? It’s an open question. Should it optimize without that and just leave it to do applications on top of it or provide a systemwide solution which aims to mitigate it?
Yeah. I didn’t get to mention this but as Uri just said, it looks like this will be integrated at the protocol level. I guess we alluded to that earlier. But given these risks about the potential illegality of…
Even if not talking about included in the protocol, which is the next big thing. The fact of this is how validators are going to run very, very soon. This is happening right now, and it’s a major change. Maybe it’s a good thing. Maybe it’s a bad thing. I don’t know, but it’s definitely a major change, and I’m not sure all the community would agree with Stephane that this is kind of like a neutral situation. How does this affect users, etcetera. I think it’s something that users should be aware of. This is how Ethereum is going to act very, very soon.
Stephane, what do you think, given what I just mentioned about this potentially being illegal in some jurisdictions? What do you think about the fact that the way Ethereum is moving is to integrate it more?
I don’t think that’s true. I don’t think that’s the way that Ethereum is moving. I mean MEV-boost does a very specific thing. It allows all validators to be able to extract at a proportional amount to their stake, eliminates economies of scale. It doesn’t have sort of an opinion on legal issues that you mentioned. I just want to avoid sort of entangling separate questions or concerns together in this conversation because it definitely causes more confusion than it does help the public understand how these systems work.
So you’re saying that if this does get integrated at the protocol level then you don’t think there’s a problem with what at least these legal analyses say?
What is this and what does integrated at the protocol level mean?
I mean building PBS into Ethereum.
Yeah. So PBS is a commit reveal scheme at the block level that allows for a validator to accept a block from the network. Then it does a change to the way that the fork choice rule works to be able to accommodate the fact that you have this commit reveal at the protocol level. What it doesn’t say, it doesn’t say anything about the way that the blocks are constructed or sort of anything that happens before that commit reveal period. So, PBS, while it is sort of a protocol change, it really does not impact the way that MEV itself gets extracted from how it does today.
One of the research paths that is interesting is sort of further transaction pool privacy. Is there ways to provide that in a way that is not trusted? So today, users can use something like Flashbots Protect to be able to send transactions directly to a builder and get those transactions included on Ethereum, but they still have to reveal those transactions to the builder. So, ultimately the builder can, only in a trusted way, sort of promise I won’t operate on any of the information that you revealed.
But using privacy technology is like threshold encryption, SGX, or others. It would be possible to implement technologies where it’s not possible for a builder to see the content of the user transactions. So, I think those kinds of solutions, those kinds of research topics are sort of completely separate from PBS and PBS getting included at the protocol level and are definitely sort of active areas of research. It might be more relevant to questions of what options does a user have when they send their transactions to a platform like Ethereum.
About the post that you mentioned, the report by the BIC and the reports on…
The CFTC rule.
Yeah. On the CFTC rule and sandwiching. I think this is a new topic that’s being explored academically, which is what are the ethics of transactions. How does that relate to MEV and how does that relate to the law? I think those are good initial explorations but they do miss some aspects of the problem, as far as I understand. We have our research team that’s specifically looking at this. I don’t know the details, but my understanding is that parts of the analysis may not consider exactly how it actually sort of ends up working on Ethereum.
For example, presenting a similarity that front running in the traditional finance sphere is the same as front running in the blockchain context. They end up being very different activities that probably would benefit from being called differently because otherwise it involves a lot of mental models that are borrowed from a system that is completely different.
Interesting. About threshold encryption and some of these other options that you mentioned to conceal the mempool and make it effectively impossible to exploit MEV. As far as I know, I think Osmosis at least is working on baking in threshold encryption. Are we seeing that other chains are trying to do this and if so, would that eventually make Ethereum less competitive against them or is this something that Ethereum is also looking into?
So, I don’t think that they make it less possible to extract MEV. Again, all MEV that exists will be extracted and the question is just what are the externalities of that extraction? So, if an application…
Wait. How, with threshold encryption, can people exploit that if they can’t see the transactions?
I mean there’s multiple different ways. It pushes at the edges of the encryption. So, it depends on what is the sequencing model of those transactions. It depends on what happens before a transaction is encrypted. It depends on what happens when a transaction is decrypted. So for example, if you have a bunch of opportunities that create an arbitrage behind them and you push those through a threshold encryption system, whenever the threshold encryption decrypts is when the MEV gets exposed. Then whoever is in the best position to capture that MEV gets to capture it. So, it transforms what has been sort of an auction-based system around the capture of the opportunity into a latency game. So whoever is collocated and is the first one to see the decryption and able to submit a solution to it gets to capture that MEV.
But front running wouldn’t be possible?
No. Similarly, you can build other systems on the other end of the threshold encryption. So, because users are sort of submitting transactions to this threshold encryption, you don’t actually solve the problem of payment for order flow. So, order flow in itself has value that comes from the information contained in the transactions submitted by the end users. So, threshold encryption doesn’t solve this problem and so you can build a system in front of the threshold encryption that says, hey, if you route transactions to me I will give you the same execution or better execution than you would get through the threshold encryption system and you can see it because I’m willing to pay you for it. So, you can still see that the MEV that gets exposed can either be captured at either end of the encryption scheme.
Although in that part, it means the user captures part of its own MEV. If whoever builds a system, show me the transaction. I would bid you to send them through me versus through a different system. It means the user who created this value with his transaction would capture a bigger part of it, and whoever built that system would capture a smaller piece of that. If it’s competitive you could…similar to the argument about miners or validators eventually capturing all the value. If you have a competitive system that everybody tried to outbid one another to see the order flow then it would be the user who would capture the value, and in my opinion it’s a better equilibrium. If the value being extracted goes back to whoever created it, it wasn’t extracted. It stayed with whoever created it.
Yeah. The big problem that exists with order flow systems in this way is that oftentimes a value is not attributable. So, it can be difficult to tell exactly who generated the MEV that ends up being extracted. For sure, I think in the best design applications and wallets…again, I’m saying applications and wallets because a payment forward or flow type of solution to me exists at the application/wallet level. It doesn’t exist at the protocol level. The best applications and wallets will be the ones that are able to return the maximum amount of value to the users effectively giving them the best price. So, what this also means is exposing a lot less MEV by these users. So, one way to think about it is a good order flow system or a good order flow solution is one that minimizes the amount of MEV that’s exposed.
Uri, you have a theory about different levers that can be pulled to affect MEV. It also goes into kind of how much MEV is impacting the ETH price or if we were to reduce it, how much that would benefit the ETH price. So, can you talk a little bit about that?
Sure. So, the high level is something along these lines, taking from my comment earlier, I think the major pieces of MEV is basically arbitrage opportunities and font running opportunities. Arbitrage is kind of okay. It doesn’t hurt anybody. Somebody should capture it. It might as well be the miners or validators or whoever produces quality. It actually might be a good thing that way. But front running meaning hurting the user. You send a transaction and you get a worse execution. Either time passed and maybe you made a bad transaction but also you might have done your due diligence but the market has changed and kept you kind of exposed because you were trying to buy ethe at a thousand dollars even while waiting the price crashed, and you will get that.
My take is that in order to achieve that we have two interesting levers. None of them actually solve it by itself, but I haven’t given up about finding a better solution to the current situation. One of them is understanding that now that we’re moving from proof of work to proof of stake, there is a new actor playing. That actor are stakers. So, in a proof of work world, you have miners. They’re using their hardware. They’re trying to get as much ethe in our context, etcetera.
With POS, all of a sudden whoever construct blocks also has a giant bag of ETH. In today’s price, 25, 30 billion dollars or something. So, if you look at the MEV game and you say wouldn’t it be nice if arbitrage still existed and people would be able to capture it and even validator capture it is fine. But do we have a magical way to eradicate front running so it doesn’t hurt users? If you look at the numbers, if you see if MEV pays the validators and therefore the stakers, something like 100 million dollar per year for front running.
So, something like 100 million dollars for arbitrage opportunities and 100 million dollars captured through front running. This is a really big number up until you compare it to how much is at stake. So, if you look at all the stakers together and they have 20 billion, 30 billion dollars’ worth bag of ethe and you could say how much is that for you? That’s something like if you do 100 million out of that, the TL;DR going back for a second. If there was a theoretical world in which front running doesn’t happen and you’d ask yourself how does that affect the price of ethe and how does that affect the bags of the stakers. You’d say you know what, if we magically eradicated front running, the value of ETH in five years will be, let’s say 10 percent higher in this theoretical world where DeFi flourish and there is no front running compared to the current world with front running.
If the difference is only five percent or 10 percent, then stakers would actually rather not allow for front runners, okay. Stakers would make more money if front running is eradicated somehow magically and the value of their ethe goes up. Because 100 million dollars is nothing when you have a giant bag of it. So, the math around it is kind of nice. Basically, if it would only increase the price of ethe something like four percent over five years and the audience is going to think what would the price of ETH look like. Anybody could have an opinion. Would it be five percent more, 10 percent more, 50 percent more, 2X? If we somehow magically allowed DeFi to flourish without front running and so this is a super powerful lever. Stakers decide what validators would do.
If stakers stake with validators that don’t do front running there will be no front running. The problem with that or with this particular lever and why this single lever doesn’t solve the problem is that if there was just one staker in the world and he would have to choose, oh, do I eradicate front running or not? Then he would definitely choose to eradicate front running. Eradicating front running would pay him something like 300, 400 million dollars per year even if we’re talking about a small price change of 10 percent over five years versus the 100 million that he would get from front running.
But if you have many stakers, it’s still in their collective favor…this is a classic tragedy of the commons. Each one of them would want everybody not to do front running but each one individually when it’s their turn to propose a block, they would actually make more money if I include front running. I’ll make slightly more money. So, as a collective, they’ll make more money if they eradicate front running and only stake with validators which don’t do front running but each individual one would actually be slightly better off if it does front running, which kind of breaks this entire utopian dream. But it’s a start point, okay.
They would make more money. Can we push them in this right direction? Which brings us to the second lever that we have, which is users. Stephane alluded it earlier that users actually control the order flow. Whoever makes the money from front running actually needs to fee these transactions. Currently, people just send out the transaction to the peer-to-peer network being seen by everyone and get rekt. This is where I kind of disagree with Stephane. It’s not a good thing that you send your transaction out there and you’ll always get the worst possible execution for you transaction. But they can choose where to send their transaction, so you could imagine a world where there’ll be an MEV searcher or a block builder. Again, if there was just single user who was doing all the transacting it would tell them, listen, I give you the transactions but you don’t front run me. You can capture arbitrage, you can make money however you’d like, but you don’t front run. If you front run me, I take my order flow and I go elsewhere.
That MEV searcher or builder would be the most profitable one. He has access to the order flows. He can capture arbitrage. He can make money and others can’t. So, users actually have a lot more power as a group than they’re actually utilizing or people realize, but again it’s a utopian dream. How do we actually do it? How do users actually leverage the fact that they control the order flow. It’s an interesting and hard question.
MetaMask is a bit of king maker in that regard. You would say, well, MetaMask can choose whoever. If we all direct a transaction to a trusted entity we have a deal with that’s not really decentralized. Basically, we just move to a centralized setup. So, there is a tradeoff here. You want it to be decentralized and send it…anybody could send it everywhere. So, users are kind of faced with a hard choice. Users or wallets, depending on how you look at it. Do I send it to everyone and the users just through the broad peer-to-peer network and everybody get rekt or we send it to just a few entities who promise us that they won’t front run us but kind of lose out on the decentralization piece.
What I’m trying to solve for and I’m trying to figure out through bloXroute and in general in discussion with the community is how can we utilize these two levers to get the best outcome for users? I think it’s the users we should care about, and I think this is where me and Stephane do not necessarily agree. Taking it back, Stephane’s looking at me in a funny way. Again, it’s not that I think Stephane is a bad guy. It’s just that I think it is possible to reduce MEV by playing with these levers and designing the system right and kind of like at the Ethereum level.
So, we make the base layer provide some economic guarantees and some mode of operation and not go to every application and so figure your shit out. Build it in a way that doesn’t extract MEV and going back to Stephane’s first point, if there are more serious validators who can extract more value and there are the solo validators who extract less value then MEV-boost basically says now everybody are super well-armed. Everybody are good at extracting. So, we took a somewhat adversarial setup and we made it super adversarial in terms of everybody are fair but this is very, very adversarial. Your transaction would try to take advantage of and my angle is more, okay, so the angle we should be exploring is how to utilize maybe these two levers, which I mentioned that it’s actually in the stakers benefit for front running not to happen. If MEV boost allows validators to connect to any third party that produces a block then maybe stakers tell their validators, listen, accept the blocks but only if they don’t have front running. If they accept front running send it to somebody else. So, it’s a potential. If user transaction could be used combined then maybe we stand a chance. Maybe we can find a good solution.
Yeah. So we’re over time. Stephane, why don’t you respond to that. While you’re talking about it, let’s bring in this idea again about how validators could potentially share their MEV profit with users, because I find that an interesting idea. Go ahead and respond to Uri.
Yeah. I didn’t like that framing. I think, I’d encourage everyone to look at something that’s called the MEV supply chain. I think it’s probably the most useful concept to think about how can this system evolve and what is a good direction for the entire thing to evolve and what is a bad direction for it to evolve. The simple heuristic that I give in that it’s both a presentation and an article is that you should look at where the value accrual is. If most of the value ends up accruing with users or validators it’s probably a good MEV supply chain. If most of the value ends up accruing with someone in the middle like a searcher, a builder, or a relay, or some other roles or an application, a wallet then it’s probably not a good supply chain.
The problem with MEV is that there’s a lot of incentive for value capture, so it’s easy for someone to say, hey, here’s a solution that solves MEV. You have to pay me to access it. We’re both going to be better off because you’re going to receive more value. I’m going to receive money, but the problem is usually that causes some centralization risk in the supply chain. So the order flow problem in this question that we’re discussing is one of the main ways in which the supply chain ends up becoming super centralized. A block builder that’s saying, hey, if you exclusively send your transactions to me, I’ll give you a better deal also means that this block builder is not operating on the same amount of transactions as all the other block builders in the system.
They can then use this advantage to integrate further down the supply chain and sort of reinforce a monopolistic position. So, it’s not that simple. It’s not as simple as just saying here’s a product that’s better for users in this scenario. If this product also helps a certain entity get sort of a monopolistic position then it will be bad for the whole set of users as a whole and sort of entrench this. So, the Flashbots perspective on this problem is there are multiple different sources of centralization. The main one initially was access to block space. It’s sort of the initial monopoly that Flashbots unbundled but we see other two big sources of monopolies that are unsolved for as of yet. One of them is order flow and the other one is cross domain execution, so execution on multiple different chains.
The solution that we seek to this is what we sort of loosely define as decentralized block building. It’s really just a vision of the future that says this supply chain that we have is going to be maximally competitive in such a way that there is no single entity in the middle that’s able to extract a huge amount of the rent, and therefore the validator gets pushed back to the user under the form of better order flow systems or applications. I suppose less MEV, I should say, or efficient than transparent markets for block space, which return the value to validators and stakers.
All right. There’s so much more to unpack. I had a number of other questions that we’re going to have to maybe touch on another time. One last fact that I do want to throw out there is that right before we recorded, Hasu of Flashbots did also announce that Flashbots is going to be open sourcing its relayer ahead of the merge. I think they had originally planned to do it at the time of the merge, but given what happened with Tornado Cash I guess they felt the need to move that timeline up. All right. You guys, this has been so fascinating. Where can people learn more about each of you and your work?
You could always Google bloXroute or me Uri Klarman basically on Twitter. I think that goes for a lot of interesting discussion. Stephane, shoot your way, but basically I’m trying to keep everything on Twitter in the open because I think the community should be a part of the discussion and not just left for us MEV nerds in our channels.
Yeah. Flashbots, we still don’t have a website. We’re two years into this project and we still have failed to put together a website, but we do have a writings website. So, writings.flashbots.net is a good place to go. We have a Discord community where people can go and hang out. We’re also super active on Twitter. There are a lot of various Flashbots participants that are engaging in that.
All right. Well, it’s been a pleasure having you both on Unchained.
Thank you so much.
Thanks so much for joining us today. To learn more about MEV, Stephane, and Uri, check out the show notes for this episode. Unchained is produced by me, Laura Shin with help from Anthony Yoon, Matt Pilchard, Juan Aranovich, Pam Majumar, Shashank, and CLK Transcription. Thanks for listening.