Is MEV a Problem to be Solved or Reality to be Lived with? - Lakshman Sankar

um lakshman please feel free to directly share screen and give the talk on is maybe a problem to be solved or a reality to live with will you be able to share screen if you have trouble i can share it with for you sorry one second yeah and uh having some i should have tested this earlier but having some like accessibility issues one second um [Music] cool are we you guys seeing my screen yes excellent you may want to yeah perfect all right uh so hi everyone um my talk is is that maybe a problem to be solved or reality to be lived with um i think it's it's uh uh pretty clear that this is a rhetorical question and the answer is sort of both the you know the the truth is more nuances as hasou pointed out at the end of his presentation just just before um in some sense we need to be working to both democratize med extraction and mitigate within protocol and application where it makes sense and so really i guess my talk is sort of uh a further a further argument this is the case that there is nuance here um as well as a few thoughts around how i'm thinking about how the future here will play out um because it is also interesting to think about where you know where will naturally energy mitigation come into play like what types of chains and and where will democratization be inevitable so by way of quick introduction um i i work uh as a researcher with the ethereum foundation um on a bunch of different stuff uh it's it's i think if i had to classify where what research questions i end up involved with it's stuff at the kind of boundaries at and and of of our protocols and where things are changing um so things like layer twos ibb eth2 staking pools um and other things as well and so this is this is kind of like i guess the the the why the talk i think um you know maybe is an old topic in the sense that you know it it's been on people's radar since the early years of ethereum but it's newly something that everyone kind of cares about and as a result of that i think there's naturally been this sort of bipartisan split and talking about it and as as an american i can kind of speak to uh the fact that that sometimes leads to an aversion of truth uh and it feels like both both of these two parties that have that have been propped up in these two sets of beliefs that i've sort of you know kind of like put up here are sort of are sort of missing the nuance or at least some nuance will be get missed if you just read uh what these two groups are kind of putting out as truth and so kind of some kind of what i see the evidence that the future will be more nuanced uh i think there are three there are three things that really come to mind so one is um and this is this has been talked about elsewhere but there's sort of a limit to scalability on a single synchronous block space actually vitalik wrote a good blog post about this that maybe i should have linked but um and as a result if if we're if we're thinking about what the world looks like when all all finance or all of you know some super set of finance exists on chain um we're gonna have to have you know synchronous block uh synchronous block spaces communicating asynchronously so uh we're gonna have a multi-chain world um there's gonna be difference in choices in different places and there's gonna be meb to be captured between between domains um additionally it's interesting that and to observe that chains that exist today that seem to be serving the same sort of you know market and big quotes here uh are making different choices with respect to med you know i think the the example a lot of people think about is you know optimism kind of like really championing mv auctions and arbitrary kind of pursuing uh fair sequencing fair ordering so maybe that's evidence that you know like that seems like pretty direct evidence that we're going to have multiple choices out in the wild um and then finally i think i think this is an area that needs uh more research and more formalism and i'll suggest some thoughts on this a little later but interchange change you know inter synchronous block space mvv is quite different from interest in from those blocks based mpd uh in terms of what is possible and how kind of big it is and and and there there are a bunch of different ramifications of that that i think we'll figure out in the years to come um so yeah so now so two factors i kind of think about when i'm thinking about um what is the future of you know how chains organize with respect to mbb um as i mentioned earlier i think the inter versus intro chain mvv question is is is really interesting um and and sort of underexplored um and then finally there's a bunch of trade-offs engineering and security-wise when it comes to many of the mitigation techniques people are exploring and so uh it's reasonable to think that you know different different chains will make different different choices there so uh yeah this is this is a kind of interesting topic so like we we haven't first of all like uh so so inter chain mep it's worth it's worth sort of talking about what that is so let's say you run there's there's a few different regimes there's you know like say a couple l2s and maybe like side chains um and you are running um the proposer or validator whatever the thing is that has sort of a monopoly on transaction sequencing on multiple of these chains um then you might be able to capture kind of opportunities that exist between these chains um but one of the one one thing that's very different about asynchronous block space and things across synchronous block spaces is that there's there is a time delay there's always some time delay between two blocks occurring in two different systems um and and as a result of that there always is some period of time where uh an outside party with a large amount of money can observe uh you putting something in a block in chain a and and intervene in chain b uh this is not true for intra chain med and and i feel like this this means that uh intra chain md is is considerably scarier like interesting for this block space that maybe it seems like there are much deeper things that can happen there and and and i but it's also kind of like uh reassuring because you know as i'll get to later um within a single chain it feels like many things can be mitigated if if the chain chooses to do so um and then yeah like i think uh phil phil's talked about this at length but basically uh you know between many different chains it's it's it's inevitable that there will be kind of these mvp applications between different chains because there's no universal notion of fairness um as i'll also talk about later like uh you know that with regards to sort of like the cryptographic and uh like privacy oriented mitigation techniques um there are trade outs there that that make make me believe that not every chain will choose to to adopt like say the best technique there um cool um yeah and and so yeah the other the other piece of this puzzle is that like the like all of the mitigation techniques you know uh seem to introduce trade-offs that uh it's not obvious that every every single regime will want to choose um and this is sort of evidenced by my point about you know optimism and arbitrary earlier but yeah like you know anything involving time lock encryption introduces some type of latency into the use of the community into the use of your system anything using some kind of fair sequencing or like a threshold encryption thing adds some security assumption to how the chain works um and so it's it's just not obvious to me that there's like a universal best solution here or it's even possible to that there will be one um yeah and so this is kind of uh this is sort of like the if there is a conclusion at this time this is sort of like what what i think is true uh and it's it's sort of self-evident but it's it's worth restating over and over until you know it's kind of kind of common like it's kind of if it's true um and it's that like you know entertainment is inevitable it's probably not as bad uh as intro chain but it's something that we should work to kind of like democratize intro chain uh is going to exist um but mev mitigation research will mitigate some of this we i doubt it'll mitigate all of it um there will be like kind of uh there's heterogeneity heterogeneous choices will be made here um and yeah and so that's that's the conclusion mitigate where we can democratize elsewhere this husso made this point as well um and i will echo it and that's that's all i got