Finding rough consensus on issuance by sacha | Devcon SEA

[Music] I'm just testing the clicker okay okay hi um can you hear me okay okay I'm sasher I'm a researcher at Lido um I just recently read the blur for the talk and it's not at all what I originally set out to do so I apologize it's not as funny as I wanted it to be and it's not as simple um so bear with me what this talk is not about this talk is not about making the case for no issu and change um this talk is about it you know there are good reasons for that to do with uh mainly neutrality um and I think like a pessimism towards attracting new solo stat but that's for another talk this is really to try and find some sort of consensus some sort of bridge between what Casper andar um and Anders are doing and how um I guess a subset of lighter contributors think about things so what this talk is about an attempt to find a symphasis as consensus that all sides can put up with it does take as axiomatic that most EF researchers have already made their mines up um on an issurance cut although there are disagreements on the details okay so what do we agree on already I think we are all aligned that issuance should concern itself with like macro level risks to ethereum um this is anders' framing if you look at number two there are sort of exist poal macro level risks and I think that's within the scope of issuance to try and anticipate and fix um one you know one thing that I think Casper and touched on is that the Crux of this is having a viable social layer that's uncorrupted by any staking provider right that's ultimately what matters to Credible neutrality okay so where do we disagree we disagree on three things I think right the first is what the purpose of issuance should be and this is kind of we maybe disagree the second is our confidence in the shape of the longrun supply curve and this is you know probably um disagree and the third is the relative chances of macro level risks under different issuance rates and I think this on this is what we definitely disagree on right this is the Crux of it so I'm going to present kind of one side thesis the other side antithesis and try and come to some sort of symphasis um so kind of the you know one side of the argument goes not touching issuance at this stage increases the likelihood that an LST like Lio ends up dominating e as the most liquid and widely used currency within the ecosystem this would make it much harder for airm to recover if that lstd were to turn malicious at some point in the future much harder to Fork something out that most users depend on um this is kind of like you know well summarized by Anders here uh you know a Keyon for mvi is precisely because of the fear that all e would be staked when some Black Swan event takes place and this is kind of sort of a story uh mode in anders' Twitter which I think does a good job of kind of articulating uh that event okay what's the counter what what's kind of the anti is here um so the other side here is that um cutting issuance increases the likelihood The decentralized Exchange dominates staking this could have dire consequences for aer's ability to remain censorship resistance and hence E's value a coinbase Tailwind is increasingly likely given Trump's election win um there is now an increased probability of staking ETFs so why are we considering adding fuel to the fire but perhaps most importantly even with less than 50% of each stake the malicious coinbase could prove impossible to socially slash and to see this um small story kind of inspired by Anders but kind of the sort of the the other side consider the following right okay a significant isurance cut is implemented um let's say in you know 6 months two years time um maybe even sooner um with margin compressed and Eve staking ETF given the green light Stak centralizes pretty quickly within a few years over 50% of validators are run by coinbase um 2030 the Dems are back in power the US government puts a tremendous amount of pressure on coinbase to refuse to process blocks that contain privacy preserving transactions in them or transactions from a continuously updated government Blacklist that could in you know include human rights activists journalists uh whoever whoever is kind of Public Enemy Number One at the time by this time Brian is no longer o he stepped down to focus on Long levity research he's kind of burnt out of all the politicking uh his replacement is not as OG or aligned and caves in in order to appease both the government and coinbase shareholders um most of the FM crowd is now on farcaster I don't know maybe they already are already on for farcaster um while farcaster has been you know great for the ecosystem it strong ties to coinbase make it especially difficult for anti- coin based narratives to spread across the social lay and that kind of has to do with like how an initial network is seeded right if you look at um I'm not going to go into this now but it does uh you know if there's momentum of initial followers and uh in that's kind of hard to undo later on um okay major Basin farcast influence argue that foring out coinbase at this stage would be extremely unfair to innocent users and you can think of sort of I think Anar has mentioned teachers with Pension funds Etc like you're not going to just fol them out and and and and stash their eath um so this leads to some sort of Crisis right so you have on one side if your ois and researchers who are threatening to leave the project if the censorship doesn't stop and on the other hand um you have these kind of uh you know coinbase aligned people that that have a lot of social influence that that say this is kind of uh a really bad idea for AUM so the AUM Community splits into two camps a contentious for fork occurs and circle surprise surprise sides with coinbase so at this point you have two sides in their own separate universes with their own chains with no practical way of coming back together um ethereum a global permissionless platform created in part to be a refuge from Nations and geopolitics instead ends up cleaved in half by a large corporation that is ultimately accountable to its shareholders and government so coming back down to earth like what what the hell is the point of all this the point of this anti fuses here is to show that it's too simplistic a view um it's too simplistic to to sort of view the macro level risk through the lens of solely the amount of E tied up in staking um which is you know a lot of the which is the foundation for a lot of um the issurance discussion um one of the main reasons given for cutting issurance so as we see there are complex social dynamics at play here that are very different very difficult to untangle right um this is one hypothetical scenario but you can dream up you know millions of them it's it's like uh it's not difficult to so double clicking on on the coinbase risk um today it's I mean this these stats are slightly out of date but it's the largest theum node operator um nearly 15% market share could be over 15% now um staking ETFs likely to happen under Trump there's a good chance coin M Will C on the market here um yeah there's you know us-based custodians are more likely to be whitelisted uh if you look at Bitcoin out of 11 Bitcoin ETFs used coinbase as their custodian this is a pattern that is likely to repeat with E the switching cost of moving capital on chain is extremely low compared to the switching cost of moving capital from an ETF sex um if staking which essentially means if staking yield drops below a certain point onchain Capital will leave higher yields in in defi that's unlikely to happen on an exchange um in more sort of technical language centralized custodians are likely to have a a much lower supply curve than decentralized pools all this means that there's a real possibility that staking ETFs combined with issuance Cuts could think of this as an accelerant uh leads to one custodian controlling 51% of e staked a few years down the line uh if that happens you can probably say goodbye to 's monetary premium um and since this custodian has a large if subtle influence over the social layer it will be very difficult if not impossible to Fork them out did something go wrong so before I come to some synthesis right I've presented one thesis short because I think uh there's like Casper Ansgar Anders have done a good job of presenting that well I've presented the anti fees in a long way because I don't think it's been presented before before we come to a synthesis I want to do a double click on how a strong LST adds value to ethereum the main I mean there are multiple lenses through this but the one I want to focus on is jurisdictional dispersion so so almost 20% of light owned operate owned operators are um domiciled in APAC countries that's sort of Hong Kong Singapore Australia South Korea um and while Europe compromises nearly 60% of the 60% of the total two-thirds of these are actually non-e right so this includes countries like BVI Cayman Islands Etc so 60% in total we have 60% non-eu and non us nodes which is a pretty resilient distribution especially compared to an eum without Lio if you look at sort of a distribution of ethereum nodes you see a real concentration in the US and Europe especially the EU um sort of according to eepa noes like more than 66% of AUM nodes are based in either the e or the EU so that's significantly worse than Lio and very prone to geopolitical uh sensorship um I think if Li is doing well but if it can get to five to 10% representation from both laam and Africa that would be perfect and I think we should make that an objective this is a picture from uh a couple days ago from a Lio Community staking Meetup these are kind of 11 leaders within the Lio Community stakers from 11 different countries all running their own loads okay so how do we find a symphysis here I think we have to start by by um pushing back on the myopic focus of solar stakers so I mean um one way of looking at this is I agree with Anders his point and his point is um focusing on if solo stakers might solo uh might stop solo staking to a higher degree than delegating stakers at some specific yeld is more important than if they change modality yes so stakers are indispensable and we need more of them but the the overarching emphasis on this I think is missing half the story and one way to reformulate this is we need to if we care about using issuance to improve geopolitical decentralization of the underlying nodes we need to pay just as much attention to who or not onchain delegating stakers might stop staking to a higher degree than those holding their e on exchanges at some specific yield I know that's a mouthful um bear with me in other words the answer we're looking for really balls down to the shape of the respective Supply Supply curves of different types of operators in question again technical jargon but I'm kind of addressing two audiences here [Music] um this way of looking at things I think was extremely well formulated by Barnaby back in February so the sort of the way he broke this down into the questions we should be asking ourselves I think was um was beautiful and and extremely on point [Music] um having said that I don't think anyone has quite answered uh these questions and there are different groups working on looking at the supply side but none of them to my knowledge are actually um specifically looking at answering the way barab has framed this problem and I think without this sort of analysis we can't really be sure whether or not um enforcing an equilibrium according to uh you know here I'm focusing on and as's practical end game we can't really be sure whether this will have a significant negative impact on the composition of the staking set okay so what's a possible path forward I have six minutes um barring enough confidence on the shape of the supply cast of different types of operators one possible path forward as I see it today is to build off anders' latest practical endgame proposal which I think is actually semi-reasonable um very well articulated um very beautiful read if anyone has a spare week to uh to get through the post um but very inspirational for me um in terms of seeing how how his mind works and how he thinks this is kind of how uh what he proposes okay this is initial latest proposal where you sort of you never have a yearly issuance rate of more than 0.5% um and there are sort of these different curves that could satisfy this sort of the three X's are on the bottom are I think his his preferred curves and then there's also like uh sort of a a a curve ball um solution he outlines well okay we could just cap it at 0.5% sort of that flat Gray Line and revisit it in in a few years time when we have more uh insights on orbit ssf and and and and me burn um and as you notice sort of they Peak around sort of 20 30 million e staked um and then they kind of smoothly descend down and they do follow the current reward curve up to uh you know let's say 10 to 15 million e uh stake but then they uh they diverge uh Peak earlier and then come down and right yeah um okay so I think to find a synthesis we can work with something that is mostly aligned with this um modulo the constraint we need to try to depart minimally from the current reward curve and this additional constraint comes from three things um one improving the geopolitical distribution of know should be a property of issuance two the macro risk of a dominant coinbase is at least as important as the macro risk of a dominant LST three we don't have a good read on what the long-term supply curve looks like even if we agree the supply curve still has some room to shift downwards from here concretely I think this means finding a mathematically elegant curve that one departs minimally from the current reward curve up until this point two has a good chance of keeping total leave State close to 50% so this kind of you know again aligns um if we're going to find consensus here I think one thing that that uh is kind of axiomatic from the other side is that there shouldn't be another money that dominates uh uh Eve um and so if you're trying to keep keep if you're trying to adhere to that you want to keep the total State close to 50% and three does not increase uh significantly increases chances of calization so that means like doesn't taper too steeply and that has to do with kind of the validated Set uh you know deciding to censor new deposits because um you know as new people uh come in uh their um uh rewards their revenue as a whole goes down so um so you don't want this to sort of go too sleeply this is kind of like a really uh simplistic back of the envelope sketch that has uh uh no mathematical properties at all but it sort of kind of gives you an idea of um this is the brown line of kind of the design space I'm I'm getting at here um although there are mtic properties I think the Brown Line sort of uh Peaks at 0.69% issuance rate uh if so you know there are ways of doing this that are not violating the mtic properties of of of an isurance rate upper limit okay another possible path uh simple pragmatic um I think a way to Fred the needle here is to add uh a fail safe to the current curve so in other words to choose a reward curve that sticks to the current reward curve for longer until we get close to 40 to 45% of each Stak say before dropping down this can be either aggressively in the spirit of sort of Ansgar um or more smoothly in the spirit of Anders um up for the bait the idea behind this is it acts as a pragmatic break on staked eeve surpassing 50% of the eve Supply um so with the current churn rate limit realistically it will take another four or five years for amount of e stake to reach 50% so this weight and C approach would allow us to gather more real world information on what the equ equilibrium is likely to be under the current curve um more information on what the sort of the long run Supply uh uh curve looks like before making drastic changes which are inevitably going to be based on um less than complete information I think the more time we have to sort of uh get real M uh data on this the more confident we can be to make the change um okay and sorry I didn't have time to sketch that curve this morning um but it basically just something that follows the current worker for longer um then sort of comes down uh at 45% stake in terms of you know being uh a fail safe to avoid worst case outcomes okay three questions to think about I have 56 seconds one what are the set of reward curves that could build off anders' pragmatic endgame post while departing minimally from the current reward curve um two how steeply can the reward curve fall without significantly increasing the risk of calization attacks is this something we need to be worried about given that the social air can correct for it um I don't know that's open question three what are the set of reward CES that could satisfy kind of the simple SL pragmatic path uh what are the downsides here again haven't you know haven't really discussed this deeply with the other side so we need to sort of uh start a conversation okay now I think you can all throw tomatoes at me I'm pretty sure everyone from the lighter side is annoyed at me and everyone from the e research side is annoyed at me but thank you um [Applause] yeah can hear me thank you Sasha um now the stage is open for questions I think we should start with the first one do you agree the issue is not in the curve but the fact we don't have a way to increase incentives to active participants solo stakers and reduce passive ones no I don't agree I mean I don't disagree but I don't agree either it's it's one factor but why don't you agree well the curve matters a lot but so there's um you know if you could find a way to increase incentives to active I mean what do you even mean by passive participants I guess whose question is it who question is it maybe you want to elaborate it a little bit more on your question is the person still here well if not then we just go to the next one why do you think 50% of e should be Stak it seems a very high percentage so first of all I don't think that 50% of e should or should not be staked I'm just trying these are kind of not necessarily my personal thoughts this this presentation that kind of trying to find some sort of consensus I think like the way you find some sort of consensus is to say okay the most important factor from one side is if should be the dominant money the most important sort of factor from the other side is hang on we need to care about the geopolitical diversity of nodes like how do you find a synthesis um 50% avoids um having something else dominate EA's money while still like giving you the max maimum room to to to uh decentralize the validated set so that's all next one is there a technical solution to identify solo Stakes for example someone onchain Z zkp and preference them in issuance I'm not an expert on on the latest there so I don't know okay then we move on to the next what do you think will actually happen with the EF decision well it's not an EF decision it's a community decision that's why we were here otherwise we wouldn't be here the decision would have already been made and ethereum would probably be worth zero um not because of this decision being made but because of the EF if the EF is a SLE decision maker the etherum is worth nothing so I have no idea um what will happen but it's not the F's decision to make it's all it's all of us I think we still have more time um should we also think of changing the entry que to be closer and close to zero instead of changing issuance curve um I'm not sure I understand the question sorry sorry with um sorry oh stop letting in new valid I think that's that's kind of dangerous cuz you have like then you cement like a an OG class of of people people that just end up uh you know living off uh yields forever and uh yeah I don't think that's healthy for the system I think you need it needs to be more fluid I mean I think vitalic had an idea if you were a cap right you could randomly eject and randomly come in I think that's interesting but I think you do need some sort of fluid uh movement um are there any more questions questions from the audience which maybe you have not found time to scan it and place it online we still have more time yeah the gentleman in the white shirt has more questions okay um the zkp question was me the technology doesn't matter I guess the question was like if one of the goals is increase geopolitical distribution um and or just like solo stakers if you are able to positively identify Solo stakers in a non Cil way then you could design the protocol that way so there are solutions I think to this like choose your stack but there's like some way to identify people so like is that part of the conversation of you know preferencing that goal by identifying Sol sters I'm not sure if necessarily identifying solers is part of the conversation but like I think part of the conversation is how optimistic are we that we can bring in new solar stakers right and this is kind of like a a higher level than that but it touches on it and I think like I would say I'm like Lio as a whole is very optimistic that we can bring in a lot more solar stakers right and I think like maybe and I don't want to speak for um you know Casper or or or Anar but I think like the the vibe I get is they're actually much more pessimistic about ability to bring in new solar stakers right and so if you are more pessimistic about enlarging this set of actors um then you might be more prone towards reducing issuance um than not but I mean people can disagree on this right thank you so much time is up