198 - SUAVE Explained with Phil Daian & Andrew Miller

welcome to bank list where we explore the frontier of Internet money and internet Finance this is Ryan Sean Adams I'm here with David Hoffman and we're here to help be become more bankless bankless Nation have you guys ever heard of flash Bots if not you certainly should have flash Bots is the silent protector of ethereum it's a shield against the forces of centralization that silently sneak into our crypto systems and turn them more into banking systems more like trafi and we think flash Bots is one of the unsung heroes of crypto the guests we have on today

are from flash Bots Phil Diane is a co-founder Andrew Miller is also a researcher at flash Bots and they believe the biggest threats to centralization in crypto still lie ahead and they've got a new platform that's being built to defend against it it's maybe their most ambitious platform yet it's called Suave and David and I have been hearing a lot about it recently in the context of this is how we solve me and decentralized block building and things like this but honestly going to this episode we weren't exactly sure what Suave was so we find out during this episode This Is Us open- Source

live streaming our exploration of this me technology called Suave a few things we talk about number one The Flash Bots Doomer thesis why crypto is doomed in a world without sve or something like it number two why crypto needs Suave what it actually is how it works and what the wins are for the various stakeholders and then finally number three and this was maybe a personal favorite part of the episode for me why the rest of the world needs SW too not just crypto David why was this episode significant to you

first this episode requires a bunch of prequisite knowledge so if you go and listen to this episode and you don't know what me is or Searchers or block Builders or proposer Builder separation you'll actually kind of still be able to catch a Vibe if you don't know what these things are there are historical episodes that we've produced with Phil and with others in the meev researcher space so you can kind of catch up to the frontier of meev which is where we are in this episode the frontier of the Arc of what is this progression of

understanding and harnessing me and that is more or less synonymous with The Arc of flash Bots and also the Arc of Phil Diane one of the two guests here there's there's this research Arc of like first identifying me then figuring out what to do about it and there's been slow incremental progress in actually harnessing the power that is M and making sure that it doesn't corrupt destabilize our systems now we have arrived at this new frontier of suave which is an evolution beyond what was

this most previous Frontier which was me boost also produced by flash bot um one of the questions I ask Phil is is this the end is this the conclusion of me research is this the final meta and his answer was like Perhaps it is uh so I thought that was pretty interesting there's one thing that we will Define going into this episode something that we haven't talked about on Bank list before and that is sgx sgx is a part of a broader term called trusted execution environments sgx specifically is

software guard extensions which is a product out of Intel it's a computational security technology it creates an isolated and protective Enclave inside of Intel processors that allows for verifiable computation to occur so sjx encrypts data within en enclaves to prevent unauthorized access and ensures code Integrity by only allowing authorized code to execute within it it's basically a way to have trustless comput computation from your own computer so that even you the person running the computation cannot tamper

with it and also you can prove that you haven't tampered it with it to the rest of the world it's a pretty cool corner of innovation that's going on in parallel to crypto that's highly conducive to some of the goals that we have in crypto and it's also a component that is pretty critical to sve so you'll hear Andrew talk about sgx so just want to define those terms all right guys let's get right to the conversation with Phil and Andrew but before we do we want to thank the sponsors that made this possible including bankless Nation I would love to introduce you to Phil Diane a cryptoeconomic researcher Phil

is the lead author behind the landmark paper flash booys 2.0 where he introduced and defined the meev problem in the ethereum landscape over four years ago he introduced me that's kind of a big deal he's also the co-founder of Flash spots which is a research and Dev organization with a mission of mitigating the negative externalities of me Phil welcome back to Bist thank you so much for having me and new to bankl we have Andrew Miller an associate professor at the University of Illinois

sgx Bull and visiting researcher at flash Bots and also a big believer in Cipher Punk values through cryptoeconomic mechanisms andreww welcome to bankas hey great to be here so there's this thing going around in the crypto world called Suave if anyone who's paying attention attention to me knows this thing that is suave Suave to me is this confusing concentration of a bunch of nebulous ideas all instantiated into one single place so here on this episode today we're going to try and parse out

exactly what is suave what Suave wants to bring to crypto and also what Suave wants to bring to the world Phil Andrew you guys think with that we can get that done this episode let's do it easy so I think in order to really understand Suave we actually kind of have to go back in time and kind of run through the idea maze that I think is Flash Bots flash Bots has this crazy awesome Arc to it and it's been one of the cooler stories that I've been following since my early days in crypto and so Phil I think in order to really uh justify why

we why the world needs Suave why we need to bring Suave into existence we actually kind of have to run through the history of flash bot and the history of MV in order to you know showcase the decisions that were made that will ultimately bring us to sve so I'm wondering if you can kind of um run us through the history the speedr runners through the history of me and Flash Bots yeah so I think my particular journey through meev everyone has their own took many many different directions but the tldr is about four or five years ago um I noticed that there was kind of a common problem that was coming up in

different things I was researching at the time I was a PhD student thinking about the question of what does it mean for a smart contract or a blockchain to be secure uh what is different about securing these systems than securing uh things we've seen before in maybe traditional software um and this problem kept coming up that there was essentially this uh breakage in the abstraction that people kind of held Canon which was of the blockchain as a bulletin board kind of secured by these rational actors um and actually if people were purely rational actors there

were many ways in which properties people took for granted about this blockchain or bulletin board as they were thinking about it this public channel uh kind of breakdown um potentially leading to the collapse of the entire system so um as I started investigating this more and more I was kind of more and more convinced that like this is what the math says and this is the inevitable outcome of what we're building um and Flash Boys 2.0 is kind of my attempt to summarize all of my knowledge and write a position paper to

kind of get people in the community thinking more seriously um because I think there are a lot of deep questions here about balance of power inside the system about robustness um censorship resistance um regulations etc etc that I thought were like super understudied so this was me and my co-authors attempt to kind of put a flag in the ground and say look this is not just a theoretical issue there's huge consequences and here's some numbers to show you how important um it was um I think once people are aware of the problem the next

question is what do we actually do about it um how do we solve this how do we fulfill the ideals of crypto um I think for me decentralization is super important I'm a decentralization maxi hopefully that will come out in this episode um and I think for me uh I I saw me and the concentration of power through me as kind of uh the way decentralization dies in some ways um and I think for me Humanity at least deserved a chance to explore and try to

integrate these decentralized uh approaches into various places where today we have trust um kind of to fulfill what we were talking about earlier about the the kind of original ideals um that that that bring for example Andrew here of uh of this movement so I think decentralization is super important um and to me the question is how do we get there from here um I think since then there have been a number of steps the founding of flash Bots is one of them um a number of products we've released to kind of try

to nudge the me ecosystem um into what we think is is a direction that thinks about these problems and handles them uh as kind of first class priorities um and yeah happy to talk about any of the steps on that Journey but I think Suave is the next one I I'm I'm wondering Phil if you could kind of describe the the problem that you saw that people at the time didn't so people thinking about blockchain as a bulletin board and and you're seeing it in a in a different way and uh you are decentralization uh Maxi self self ascribed uh so you know I'm

taking that to me you saw the centralized properties of the system under threat from this particular class of problems can you describe the problem like what is the problem of me and in the I mebe the easiest way to understand for for a crypto native for a crypto listener what is the class of problems that you are you saw then and you're still on a quest to solve well I think the fundamental problem is that what blockchains try to and need to do for them to work is to come up with an order

of events of transactions that they're processing and kind of communicate to everyone that this is the order that we all accept for these events or these transactions um I think people hadn't maybe taken into account how much discretionary power there is in the mechanism that actually comes up with this order so in terms of reordering transactions or inserting your own transactions or censoring transactions these are all options that essentially the infrastructure providers of these networks have the option to do from a

technical perspective and there's nothing in the economics that prohibits them from doing these things um but they're very much not the expected behavior and they lead to this very complicated problem of okay if I'm one of these infrastructure providers what is the most profitable thing I can do in the system uh the core problem there is this creates basically an asymmetry so people who are more sophisticated in being able to leverage these Powers being able to reorder transactions to insert their own and to censor have an advantage in the system over ones who

don't because they get paid more per action basically um and if you're a game theorist what that means is when you play the game many times basically these are the set that will play the game because they're the ones that are profiting more than the other ones they can afford to drive the other players essentially out of the system um if you look at the act of actually building a block and what it entails when it comes to most of these transactions being Dex transactions being Financial transactions um the the the players and the entities that have the most isation

in actually maximizing their own Revenue against this kind of optimization problem of how do I actually process these transactions how do I build a block um the players that have an advantage in that game are essentially hedge funds proprietary Traders and people who have maybe uh other reasons to be trading uh on the blockchain other than purely providing infrastructure things like Arbitrage um and so my fear was essentially that our system was enshrining a concentration of power on

the infrastructure level to these players essentially the more they could um kind of play these games and do so in a way that was allowed essentially uh that was more sophisticated than other parties so I think that's one problem and it could potentially at least at the time one of the big concerns was you you you can concentrate the validator set of eat this way where if you're doing a proof of stake analysis um this was before we moved to proof of stake you're relying on having many parties with many incentives to kind of come to this

robust consensus um if those incentives aren't as clear or clean as what you're putting on paper then your consensus maybe won't actually come to fruition so that's kind of how the system could ultimately break as a result one of this the reasons why I think many people who are in crypto for the tech come to crypto is they see the Equitable foundations on which this these platforms like ethereum like Bitcoin provide to the world I think for what you identified in me is like oh we do have this Equitable Foundation but me

can tilt it into the balance of the Tilt power into the balance of parties that already have it right people that are already well capitalized to Market make right like technical uh meev bot coders who have this power and they can capture that value and that can be expressed in uh an increasing concentration of eth stake at a faster rate than than non-technical people uh and so I you know being a decentralization maxi I think this perhaps like one of the the arcs of flash Bots is like how do we preserve the Equitable foundations of these platforms one of the big Flagship

products out of flash Bots was this thing called me boost can you talk about meev boost and how it produces and like helps maintain the Equitable foundations of ethereum yeah so meev boost came out of a line of research called PBS or proposer Builder separation highly recommend reading up on that uh if you're not familiar with it um this was you know many people's kind of memes spearheaded mainly by vitalic I would say but with many kind of contributors in the community um and this is the idea

of basically separating this sophisticated activity of optimizing blocks from the activity of proposing blocks on the network so the actual proof of stake protocol that was being designed the reason for this was we had a system called Mev which worked on proof of work it was a very different auction um and all we were trying to there is basically show people one ordering matters and two um in a permissionless auction for ordering uh you basically get more value for all

parties if that auction is permissionless and competitive than you do in a in a proprietary or like side deal style auction um when proof of stake came the question was okay what's different here I think one big thing was the desire to allow solo stakers so parties that were not as sophisticated to participate um and Mev boost and PBS was a basically a way to try to contain the specialization to this layer and this role that was called the Builder um that would essentially make these complex trade-off decisions in an

auction that was competitive um the idea being that if you were a validator even if you were not sophisticated and you were just proposing blocks and validating ethereum you could still make as much money as the sophisticated parties um by essentially receiving bids from those parties for um your action space um and if you make that comp itive it's kind of better for the system than having it be side deals or or opaque so that was kind of the the background with uh with Mev boost of course there's

problems with it uh key among them is it relies on centralized trusted parties um to achieve a lot of its guarantees and I think that's really what we're trying to move away from in the next phase the Rel layers right being one of those trusted uh parties and I believe it's what 90 to 95% of um validators are now using me boost at this point in time so you guys have been pretty effective with this and one of my favorite parts of this story Phil is not only were you and the Flash Bots team responsible for making the

entire crypto community and the crypto industry available uh aware I I should say of the power of block ordering which we were previously unaware is you guys actually put together a solution so it's it's kind of rare to have like somebody uh surface the problem and then also architect the the solution to that problem whenever that happens uh it's it's just a fantastic thing before we get to Suave because I think we're now almost caught up in history to where we are right now we've got this thing called meev boost it is helping to

Corral and mitigate uh some of the me in all sorts of ways I I just want to bring Andrew into this conversation so the problem set that uh Phil teed up Andrew what about that world is appealing to you like how did you find yourself being being sucked into the uh the rabbit hole that is solving these weird me crypto coordination problems because you're an assoc you're an associate professor aren't you Andrew yeah that's right I come from this from a a different uh uh perspective starting from privacy

Technologies I've kind of spent the most time uh doing research on I mean the whole stack of things like zero knowledge proofs and multi-party computation and so on and i' I've spent the most time uh working with zcash and their um I mean to me my my perception of me was just it's something that has nothing to do with privacy it's something that's very financial and that Phil has it under control so is just something that I you know I was aware of kind of during the past you know several years but not um you know hadn't piqu my interest so much because I didn't see that you know that connection to it but

um the path that I've been going down has been basically realizing kind of where the limits of zero knowledge proofs are and kind of making this kind of observation that most of the interesting applications have some kind of shared private State that's needed and so you really can't just do everything with zero knowledge proofs you have to use one of these kind of alternative techniques and so I spent the most time working in multi-party computation and in um trusted Hardware enclaves like sgx kind of recently and so um I mean from my perspective like we wrote This research paper called Eiden

in 2018 and that kind of became the basis for Oasis later and it basically has the same design you know structure as secret Network which then um uh I followed that really closely um basically in terms of identifying okay what are the differences between what we wrote in the Iden paper and what they actually implemented that went down this kind of uh uh interesting Rabbit Hole onine of its own and but a really relevant one of like what it takes to build you know a credibly decentralized system that relies on on trusted hardware for its shared private State and so um that all kind of came together at the moment that

I I heard about the Suave proposal and suddenly a whole bunch of things kind of Click so first of all the need for privacy not in the long-term user privacy like protect dissident you know a user privacy view for meev there's this very obvious but short-term kind of privacy need so you care about pre-trade uh privacy and then that's actually really important not just for users own sake the Big Challenge in working with privacy Tech is that users won't pay for privacy they don't understand it very well so it's you know very challenging to get on you know adoption of good

privacy technology and so this is a setting where you really need the Privacy it's really well motivated it's well motivated in a way that you can um see and understand the consequences of it a lot better and so so um I just became clear that this was going to be um a really great way of approaching that problem combining these two technologies and then it fits the kind of other goal that I've had which is to basically um you know Elevate the quality of uh systems built using trusted Hardware enclaves so um I'm not an sgx bull per se but I do think that

this is technology that's kind of misunderstood and and underrated and you know gets a negative sentiment for misunderstandings as much as for you know real reasons that that deserve it and so so OB to me is an opportunity to basically you know do this the right way for a perfect problem you know while Everyone's Watching and the incentives to get it right are really clear and uh yeah that's my path to it so Andrew safe to say Phil doesn't have it all figured out with respect to me he needs some help particularly on the on the Privacy side he needs a team around him to actually help him achieve some of these

objectives yeah that's the case can't just like stare at Phil and wait for the entire problem that's what I've been doing that's we've been doing I'm stunned that Amia is still an open problem at all but uh yeah but now now we're we're kind of staring at at at both you guys and I'm wondering Andrew if you share this intuition that that Phil stated earlier that you know if we don't solve the me problem right um you're talking about crypto uh values and kind of libertarian values and you know identifying with the Cypher Punk Manifesto but if we don't solve this

problem it could potentially lead to the collapse of the entire system do you think that's what's at stake here do you share that intuition um yeah I mean I've heard this described as the flash Bots Doomer thesis I find it fairly compelling I I think um I mean all credit to Phil for you know like you said you know accurately prophesizing a problem just in time to see it you know actually take place and then be able to you know kind of identify paths towards Solutions of it you know that doesn't come around so often so that's like a really cool thing to to foresee and specifically foreseeing when things will become huge

uh Trends towards decentralization I think that's a really great thing to try to identify and then try to miate gate um so I mean whether it leads to just a suboptimal outcome where there's more centralization than than needed um or whether it you know ends the entire experiment um regard whether you take the soft Doomer thesis or the full-on one um it's a really important and worth while go so I do think that these are areas where um uh auctions in general all of these uh things that are like orderflow auctions all sorts of components in the me you know supply

chain they implicitly rely on privacy that comes right now only at the expend through the mechanism of just trust and that is fundamentally a centralizing force and so trying to build something that avoids that is um yeah hugely important we've been setting the stage what I think is going to be a good place to bring up Suave here uh we have the history of flash Bots and meev boost and the mechanism designed that that has created to contain some of the meev then Andrew we're we're bringing in some

privacy and some further mechanism designed tools to to add into the mix of what it takes to fully solve this problem Phil I want to um give you my take about like this this Arc that is Flash Bots and the fill Arc and really just the harnessing of me arc where you know it starts with identifying it with the Flash Boys 2.0 Paper then Mev boost starts to actually contain it like puts a wrapper around it and it pushes it away from the centralizing forces the big you know jump cryptos of the world

the big market makers of the world and it actually harnesses it and pushes it into um e staking actually which is a more distributed set of people but it's actually not doing the thing that I think Suave wants to do which is nipping it in the bud in the first place and I think that's kind of the big golden solution that we are looking for that flash Bots is looking for is it's great that we were able to harness me through Mev boost and instead of giving it to

just one centralized very expert very well capitalized parties that are you know meev Bots and market makers instead pushing it towards a very wide distributed set of people in the eth holders the eth stakers that's an improvement but it's still not what is the golden standard which is just like not having me extraction at all and I think that is where this Arc is pointing towards and what Suave is trying to do that's my understanding how does that Vibe with you I mean I think the distinction between solving and

harnessing me is actually very small like I think ultimately um you know um me is basically about what is the power distribution in your system how is information propagated and who profits off that I think you can always very easily solve it through centralization essentially um like this is actually what when I first even years before flash blaz 2.0 when I first discovered this in the context of Dex design uh this was before Unis swap launched on chain back in The Ether Delta days uh

when uniswap asked me like hey what do you think of this design I was like well you should really have like a centralized order book that you all run in a permissioned way and like don't cheat because otherwise there's no protection you know um so I think there are solutions right now but there are also limitations to those Solutions it's this kind of like devilish problem where like because it's about Power Balance and and Arbitrage and these complex information asymmetries um it's not clear we like even know how to like solve it or what that means um like

there may be one structure that's one set of people consider a solution because it has less of like some type of toxicity that they subscribe to that another set of people totally don't even want to use because they consider it so unfair um I think this is like the trillion dollar you know like divide basically so that's all to kind of troll you into saying like I get the impulse to solve me but to me the only solution is to like have a fair decentralized permissionless way to harness it um and

you want to yes be thoughtful basically not to be needlessly sending too much to the validators or sending too much to to jump crypto or anything like that in your Market structure um but uh yeah it's a very complex problem with with many knobs I think what Suave is trying to do is to add decentralization to as many of those uh systems as possible just provide a new tradeoff Point that's more decentralized from a an infrastructure point of view and to me this is like both the solution and the minimization of me because then users

will be able to harness it because they have a permissionless competitive environment they don't have to be kind of at the power mercy of whoever controls the uh the pipes the way they are in trafi for example so as we get into what is suave and we asked that question I'm not not going to ask it yet I I just want to want to set this up with one other question which is what outstanding meev problems still exist that Suave is really teed up to solve like give us the problem set Here Andrew

said earlier oh I think Phil's got the me problem well it turns out he doesn't there's more work to be done what is that work yeah so I absolutely don't have the me problem which is actually why I wrote the paper like for two or three years before the paper I tried to solve it by myself and with a few collaborators on the whiteboards um and I went down all sorts of different directions including social Choice Theory Fair ordering auction Theory block space Futures these are all like papers that I had half of and then decided it didn't work and like we threw

it out so I don't actually know the solution I think the whole um you know at least for me the like crafting of Flash Boys 2.0 is how to make the problem simple enough that people are convinced that it's that they can just solve it but actually it's like devilishly difficult it touches on things in like law economics Market micr structure you know um protocol politics in terms of distribution to Holders valid daters Etc consensus research um distributed system security you know

economic mechanism design using settings that we haven't even seen yet in in Academia in terms of complexity so there's many many verticals I think every person who kind of approaches this easy problem finds some path to start walking um and it's and it's a journey and to me that's like one of the most interesting things because every every person has a different set of unsolved problems I'm curious Andrew to hear yours also so um I I think at least for me my biggest thing that keeps me up at night is like can we come up with a a

power distribution that doesn't collapse to centralization can we maintain permissionless uh what are the best economic mechanisms to kind of um encourage that competitiveness and distribution versus like uh you know more predatory rent extraction or zero sum games or sniping or or something more socially wasteful um so that to me is the B question uh but we don't know how to do that in trafi so I don't know it's hard to do in crypto also uh so much of that kind of space about uh

where this you know what you can do with Suave or any system connect with uh you know open problems in Market structure that that's all think open even in the tradire world and so um yeah and I don't have so much Insight on on kind of those so I mean I have a kind of simplified mental model of you know what a sufficient number of problems are in me for which there's already like not a good solution you know except for something like sve so to me it's like you can already see kind of the the critical effects of like sandwich attacks and lost Arbitrage opportunities

even in just like a little toy model with like a couple of amm and some kind of fragmentation of liquidity suddenly then the ordering you know matters and then you know all the relevant phenomena kind of show up uh so to me I imagine the problem that is still open but that Suave is addressing in the simplest form of this problem even before getting to you know all the other more open-ended ones but um is really just you know how to do a you incredibly decentralized job of um you know applying some sort of sound auction mechanism to all these

bids coming in from users it's the same thing as finding you know in ordering for trades against an amm and um you know the ways to do that that give any sort of you know guarantee of uh quality for users or satisfy any sort of you know fairness objective you could come up with pretty much they all rely on um you know privacy in some way and so uh without without Suave there isn't a good alternative way of um addressing all of those and especially not in a generalizable way and so the idea that um to me one of the things that clarifies the scope of suave quite a lot

is to say and now actually I didn't know that that story the way that Phil put it but kind of clarifies a lot so after you know banging your head against the wall trying to solve meev through doing the mechanism design yourself different perspective is to say you know this is going to be a game that continues through metagame phases and and shifts The Meta shifts over time there isn't going be one mechanism that wins them all so it's necessary to come up with more of a platform view so a platform for building me relevant mechanisms all the interesting ones are going to need privacy so that's kind of you know the

key feature cred you know execution accuracy as well um so then those are kind of the things that the platform has to provide um but yeah to me that's you know basically what frames the the problem that swap solves I think we've done a pretty good job actually uh defining the spirit of of suave and the the goals of suave the Contours of suave but I think it's about time now to actually approach this conversation headon so Phil what's Suave Suave Suave is the single unifying