Ethereum Sequencing and Preconfirmations Call #9

Transcript

hello and welcome everyone to uh the pre-confirmation and sequencing call uh number nine today we have an exciting topic uh which is credibly neutral collateral for the pre-c confers as you know if you have a decentralized sequencing mechanism with pre-confirmation there needs to be some sort of collateral um so that if the decides to reneg on their pre-confirmation they get uh financially slashed and that then begs a question how what collateral uh do we do we do we want um and in order to achieve credible neutrality and for everyone to agree on

uh on collateral really there's there's only one option which is to use if uh unfortunately um if is not super you know collateral I necessarily there's other options there's llts there's uh lsts there's you know also reputation some people might want to put forward their reputation there's there's uh all sorts of things other than Eve that can be pledged and so uh last week I was talking with Matthew uh and we came up with this pretty kind of simple but neat idea of of of an underwriter um and Matthew has uh prepared a write up and a

presentation and so Matthew um the floor is yours to present all the all the details thanks J I'm going to share my slides real quick and then I'm going to post in the zoom chat and on telegram all of the the links to the full right up all right so yeah well let's start with a quick personal intro so I'm M team I've been a Searcher for a few years um in the M boost ecosystem and now I'm researching ethereum um I recently joined as a co-founder at Spire labs and I'm working as the chief product officer currently we're building a based app

chain framework to enable developers to create base rollups um and we are planning to have native support for base PR confirmations we're also hiring so reach out if you're looking for a job that's not what we're here for we're here for incredibly neutral pre-confirmation collateral so this idea Justin gave a great intro um is just a smart contract that can be deployed on chain to facilitate the hosting of collateral from pre-conference so first a quick background U the fundamental constraint of proposer commitments is that proposers need to post some collateral and this is before we have execution tickets or APS burn or something like that um and the reason for that is that proposers always have the option to self-build which could mean they they reneg on their commitments so we're

going to make some assumptions for this post which I think are reasonable to the kind of based pre-confirmation landscape right now the first one is that only the proposer is making commitments for a given slot no one else is involved this is a kind of simplifying assumption but should extend to like delegated pre- confirmations things like that um and of course there's only one proposer and then all Faults Are attributable So if the proposer does something wrong then we know that they did something wrong and that can be verified on chain for slashing or penalizing purposes so here's here's kind of the problem and Justin gave gave a a great overview um we incredibly neutral and safe and dynamic collateral so when I say Dynamic here I mean collateral that is capital efficient and works well with large operators and solo stakers so um there's a bunch of you know very large

node operators on ethereum who run many many validators we want this to work for them be Capital efficient but we also want to to look out for the little guy the the solo sticker who may have one or two validators but still wants to provide precc confirming Services now doing this is is uh pretty straightforward we also need to be Capital efficient and easy to use and that's kind of how we're looking at this problem now so here's our solution is we call the prec confirmation registry like I said before it's a layer one smart contract that handles three functionalities these are registration delegation and penalties um we have this this role C registrant who posts pure eth collateral to the smart contract registrants can be anyone and they're um

just a ethereum address and then delegation so this allows registrants to delegate collateral that they have posted through proposers so a registrant could delegate to many proposers and then their collateral would be um shared across all those proposers and we also um have the registry handling penalties this can be done in like a modular agnostic way so um you choose like a another slashing contract that can handle when to freeze when to slash Etc and then we have this um kind of conditional if a the total unpenalized collateral backing and proposer is greater than a parameter minimum collateral than a proposer is considered a pre-offer or a sequencer uh so this is like kind of our minimum bar that you need to arrive at we've discussed this number being um a very very high number

of eth so maybe a thousand eth even more than that and kind of the idea here is that registrants can share collateral across all delegates so if I am a registrant and I put up a th000 eth I can share this collateral across um my 100 proposers and each of them has a collateral balance of a th000 eth um individually and they can all become pre-conference so I have this little picture here and this kind of shows how this um onchain design works we have the registrant submitting eth to contract and then choosing some delegates and then those delegates become pre confers and then everyone else who does not receive delegation is just an includer in the based preon model can keep going so like like I said

before the onchain construction is simple but the offchain construction which is what just mentioned the underwriter is very important to to this design and ultimately is going to be what really makes this Capital efficient for for users and getes us some of our um like adoption so the underwriter is a market maker uh it's their job to convert impure collateral to pristine eth so like Justin mentioned we can take collateral from any form from proposers so think eth think lsts or lrts Equity even reputation um there's a lot of things that that could be collateral the underwriter is not opinionated over what type of collateral they're receiving um and then they submit pure eth to the the smart contract um and delate to proposers that had given them collateral

so if a group of proposers um let's say a a very large known operator sent um maybe an LSD to an underwriter the underwriter would convert this into eth internally and then submit that eth to the smart contracts collateral and then delegate it to the proposers involved so we think it's pretty cool um the underwriter acts as a registrant when acting with the interacting with the onchain contract um so there this regist role that we talked about earlier um and then what I think the most interesting part of this is is that the underwriter could be a trustless defi application an offchain Market maker um or even LSD or LRT so there's a lot of options for how to implement an underwriter you can do this in a completely trust display with some um

signature exchange protocols and then you can also do this like a off chain way there's there's a lot of design space for here and this is definitely something we're we're thinking about um the next concept that we introduce for um specifically solo stickers is this proposer pools so proposer pools are formed from a trusting collection proposers so all the proposers trust all the other proposers in a pool and then all reaking they e so they restake to a underwriter or a smart contract and that registers for them on the pre-confirmation registry and then that gets them up to a minimum collateral balance converting them all into pre-conference so this essentially allows a group of proposers to team up to reach the collateral balance that is required to become a pre-offer now of

course there's a problem where if one pre-offer defects and and does like a griefing attack um and ren EGS on pre- confirmations then everyone is penalized for this everyone's collateral is penalized so as a proposer if I Jo on a pool I need to to trust everyone else there um the way we can kind of disincentivize this behavior and make that trust like not um such a high requirement then we can have proposer pools also act as smoothing pools um which haven't seen a lot of adoptions so far but um the codee is out there for that and that just means redistributing all rewards to every proposer in the pool so if there's an execution layer Rewards or a very high Meb block that would occur because someone reneged on pre confirmations then the rewards from reneging would be redistributed to all

proposers so instead of one proposer having the um the the right to reneg and then keeping all the profit for themselves um we just allow any proposer to make that choice on behalf of the entire pool so this kind of like the the short-term thing but once we have DVT um matured and a Max EB increase then we can have all proposers just become one validator with TBT um and then combine their e together through the max effect balance change to something more than minimum collateral so if Max EB is changed to 256 eth then proposers could just become a single validator and um just do the a single registration there

it's kind of a Natural Evolution and very compatible with the the rest of the system so here's kind of the picture for this set of things um AR operators submit they're they're non-e collateral so um well it could be E I guess but it doesn't have to be it could be um like I said LSD reputation things like that underwriter Converses to eth and sends it back up to the pre confirmation registry the proposer pool um has a a group of proposers all restating their eth so they have this like form of collateral called restake eth which is not quite as pure as eth but it's getting there and it's going to be converted to eth um through an underwriter which can also be done like a smart contract on ather so here's our number one parameter and I talked about this before but it's

this minimum collateral thing so Jonah had a great post about setting honest users set like requirements for pre-confirmation um this we kind of want to extend this idea so setting minimum collateral X Ane so before we know what pre- confirmations are going to look like may lead to inefficiency because for example a user um a very small user who just want to send some meth to their friend doesn't have to have thousands of each backing their pre-confirmation they don't need that and quite frankly they shouldn't have to pay for that so the idea here is we kind of allow users to Define what penalties they they actually need for their um collateral and and we go from there now the caveat to that is that if we set minimum collateral to to something low then there can kind of be

this choice where um a a proposer with very little collateral may become a pre-c confer um but there's still a proposer with a lot of collateral later in the look ahead so you have this problem where even though there's a lot of collateral available later in the look um and a proposer who has less collateral will still get sequencing rights and then therefore um that collateral that is available will not be used this is um also kind of inefficient especially for people who need a lot of collateral so this is an open question and we're we're definitely accepting input on this I can message be on telegram we can chat about this all right that's it it's pretty simple um like I said Spire is working on this stuff we plan to play a big role in the implementation and and further

research on this subject um we want to look into being an underwriter into playing some of these other roles in the ecosystem um and that's it yeah so you can contact me at mt. space and my Twitter is Mt thank you amazing uh thank you Matthew um I guess we can open up for questions uh yeah I had a quick question um this kind of touches on like reaking in general but um if you have an underwriter and they're under they have like a thousand eth and they're underwriting say 100 different parties I guess uh how do we envision like the slashing happening or like when does the slashing happen because I could I guess

I could Envision if you have like 100 preconference that are being underwrite underwritten um and the first one Ren eggs on their pre-confirmation but the slashing doesn't occur until like several preon later and the preon fers later in the line basically say like oh well the first one is going to get flashed so I can kind of do whatever I want without any penalties uh so I don't know if you've given any thought to that yep thank you so you can find the answer to this question in the in the full okay that has a there's thing at the bottom that has like considerations and we we touch on there um so we actually call this problem like force ejection if there's a proposer like in the look ahead later than a current one um then we kind of have a liveness failure so if you get slashed early and

then your collateral is also backing future pre-c confers um then you know there's obviously that collateral is no longer there it's already been penalized um and so that those prec confirmations are no longer valid we're hoping that slashing events are are very rare um but in in the case that that does happen then you know obviously there'd be a problem this goes back to kind of our initial assumption that there's only one active pre-con at a time um and so we're hoping that we can know as that there's been a fault and then we can giving out to confirmations before that's like understood and and have a force rejection there I hope that answers your question uh yeah thanks I mean in some sense there's two layers to the slashing there's the onchain slashing that you

know might happen with some amount of delay which is not great but what is a a slashing basically it's acting upon a fraud proof and this fraud proof can be gossiped with extremely low latency in the peer-to-peer Network and so if I'm a wallet and I'm listening to this PTP Network for for for proofs I can just preemptively slash and you just not send my user flow uh to uh to the next uh proposing look ahead if if that proposal happens to be uh about to be slashed on chain yeah that makes sense any other

question I have another question sorry uh going to the underwriter uh could you kind of explain like it makes sense that someone wants to stake their reputation like that totally makes sense I guess I'm a little bit unclear about how the underwriter would take that reputation and turn it into eth so like I guess they would then just have to front the eth and then they would have some like smart contract that says like oh this reputable person is staking their reputation um I I guess is that the correct understanding of how like the reputation turns into pure e yep exactly so the underwriter in this case would um let's say be an offchain Market maker um and they would convert reputation eth on behalf of a you know maybe a large staking protocol or something like that um okay it's definitely it's like a

trust very trusted relationship because obviously there's a lot of collateral um on line here I don't know how like how widespread reputation will be used as collateral especially um in the end game but it it can be used as a bootstrapping mechanism to get us like 90% of the way there and one concrete example could be for example you know lisl who's on this call like his reputation is probably worth more than a, and so someone could be willing to put forward a thousand e and Trust lisl to to not cause a uh a fault um it could it could also be like um you

know contract based like the collateral could be a you know a house or something um and and then you can um make use of the legal system maybe this is beyond the scope and maybe I'm just not understanding fully but how do you guys sort of size the amount slashed or is it just there's there's an amount and it doesn't matter what the impact is like is there a dynamic function of let's say it

impacts you know how do you measure the impact but let's say theoretically you could and it had an impact of a thousand e you know let's say like I was liquidating someone I couldn't get my transaction does it take into account any of that or is it just kind of like this is what we're going to slash yeah so this is generic way to put up collateral but the actual slashing mechanism so how much you're going to slash where you're going to slash um if you're going to freeze it or whatever can be can be done um by by another smart contract it's not like itself yeah great question though right in some sense the amount that's slashed is a bilateral agreement between the user and the the sequencer but the the minimum balance to register that is

something that we want to have a very clear shelling point so that all the roll apps use the same number so that all the rollups agree on who the next prec confer is and we can have shed sequencing okay so the standard is on the minimum required and then the maximum is sort of like great relationship between the two parties in some form yes but even if the minimum is let's say 100 or a th000 the amount that is actually slash if there is a fault doesn't have to be 100 it could be just one if if that's agreed with the user Conor yeah I guess the uh like all of this and every all the assumptions here

or is based around your I guess maybe your original assumption that it has to be 1,000 100 right there has to be some collateral on the line like is there any consideration for like no collateral like because obviously if if we're talking about reputation then like that's that's what rollups are basically doing now right so like with the system you obviously think this system wouldn't function with just pure reputation but is is that is that possible you to be unified interface for right for all sequencers I guess yeah I think can answer this question but I just want to say like the proposer always has um can always reneg on on their formations and because proposers

are such like decentralized um so that's like any if we make it permissionless to become a pre-c confer and and therefore um anyone can do it then prec confirmations essentially have like very little value and it's very easy to for especially because you know it's all permissionless and and very distributed as a part of like ethereums proposer sent um and and of course the I guess the underlying Assumption of all of this is that um pre-confirmation are going to end up securing like a lot of value um and if that becomes you know increases in the future we we also want to have like a robust framework in place to to do that I don't know if that answers your question yeah I guess like so you're seeing that we uh we we need to have like a standard framework for

slashing um but like in the case where like let's say you're the person who doesn't have doesn't have enough collateral you're obviously a highly reputable person so that probably should be enough for most people to trust you to do pre informations anyway um in the case that you're like a pool again you have the same thing you can just like say oh we're actually just GNA in this case we're gonna we're going to put up Stak e like the the thing that you're slashing the reputation or the tokens doesn't necessarily feel like a fundamental thing like is in a fundamental thing to to pre confirmations we're just putting something in the middle here to standardize the interface but I'm not sure if that's actually necessary so in terms of the bilateral agreement between the user and the

there if let's say it's 100 e or a thousand e that doesn't mean this this thousand e is going to be the thing that gets slashed uh it could be like side collateral or it could be as you said just pure reputation so coinbase for example could register with a th e so now we all have an understanding as to who the next pre-con fer is and when the pre-con fer happens to be coinbase coinbase can say you know just trust me and tough luck if you don't trust me um I don't I don't want to be exposed to any slashing risk um but they do need to put forward a thousand e just to start playing the game so that we we know who who is playing the

game why why is that acceptable I I think this the reputation side of this is I'm getting a little lost with yeah I mean from the perspective of users and rollups there's always collateral and whatever this min collateral parameters there's always that much collateral backing a given pre confirmation unless there's like slashing event that happened in the past so it's not reputation based on like the I guess the slashing side I think what Justin's trying to say is that there's um like this reputation applies to to uh like rollups um specifically I don't know J can you can you explain on that a little bit right so in order to register you

need to put forward like the strongest form of of of asset which is which is Eve and this is from the point of view of infrastructure that is enshrined on chain which is you know the the rollup smart contracts and the registry that determines who the next pre-con fer is but then there's basically a dynamic which is fundamentally offchain so once the pre-conference have this fundamentally offchain bilateral agreement um and they can do whatever they want it's totally permissionless and you know if they want to use some other asset other than e as collateral uh that's that's fine I mean it's not something that we can prevent

right but it's it's built into this registry contract to have slashing um if you have like a unified way to do slashing for your roll up right yeah I mean from the point of view of avoiding fragmentation and having a standard yes it would be very nice if every single format and slashing conditions but you know it's just like paying tips and paying fees like in theory it is always possible to have a a a separate agreement that doesn't go through the the standardized rails all right cool

Connor yeah so like I guess what you mentioned that the wups need to have some sort of something put up by the the pre-confirmation like paying up front or paying paying a fee should should possibly be enough

right yeah that's that that's interesting um maybe instead of like paying with the cost of money you mean uh you just pay up front with actual money so yeah just burn burn one e and there you go you registered you're take you're taking the risk that you like people won't send you pre- confirmations but I guess you you would probably only pay in the long term if you're it's a bit like execution tickets but potentially without needing execution tickets I mean the nice thing of having let's say a thousand is that if we can have a shutting point that everyone should not only use the F NE for the

registration process but also for the bilateral agreements with users then then kind of dual purposes this this and it's actually maybe more efficient right because you don't burn the one e um so in some sense the the Thousand e is quote fundamental and so you might as well Dual Purpose it as both registration anti cble mechanism but also as a as slashing collateral in the case of of a fault let me ask your question I I I was thinking about this because I think this is the most one of the most important things that should come up and every conversation is is trying to make these systems as

censorship resistant as well as possible and I'm just wondering if uh the reputation side lends itself as credibly neutral from a censorship assistance standpoint or if that's the purpose of having you know decentralized permissionless collateral right that's the whole point of the underwriter is that there's from the perspective of the the onchain system there's always 1,000 e or whatever minimum collateral is um for for every proposer right so there's always um this kind of collateral pristin collateral backing in the underwriter has to take that like reputational Risk off chain um if we do something like Justin sestic where there's like these um almost out of

protocol agreements for reputation after the registration and then like selection phase is over um then yeah that would be that wouldn't be maximally incredibly neutral but there's ultimately nothing that the users can do to um you know if they want a pre-ert then they have to go to the the preer so the pre- confer um is enforcing that then it's just social consensus to avoid that but there's no U there's always there's always a thousand e on chain backing your pre confirmations anyway Conor you have your hand up so I guess I guess that's probably a

legacy sorry Sam did did that answer your question so in some sense for for sensorship resistance we want to have like a very diverse set of pre-c confers now if the barrier to entry to becoming a proner is a thousand actual like pristine e then that might in artificially increase the barrier to entry but then we have this idea of using Underwriters that can do the translation between non pristine e and and pristine e so that if if for example LL comes forward and he doesn't have a, e but his reputation is worth a th000 e then now he can become a a a

pre-conference so you know L is la is adding to the decentralization of the set of pre confers and it's it's improving uh sensorship resistance yeah thanks Jus the other side of this is the proposer pools right so if there's a solo sticker who doesn't have a lot of reputation but still have um their 32 e they can use they can also um get their teaming up with other proposers yeah I just figured that I mean in a sense like it's a little bit like with the hedge fund or a mutual fund right like you want to invest money because you have this great idea but you don't have the assets so you go out and say Hey you know give me your assets and I'll go and do this work for you right and then I'll return you the returns

like my guess is that the you will have a lot of cooling that will lower that threshold such that you don't need to have reputation but I could be wrong yeah one of the issues with with pooling by the way a lot of these issues just go away completely once we have execution tickets or execution auctions because now the the validators uh don't have to put forward collateral so it is a temporary problem uh and once we do have these execution auctions then the only entities that do need collateral are the the sophisticated gateways or whatever you want to call them and realistically there's going to be a

small number of those um yeah I forgot what yeah oh yeah the problem with pooling is that it requires maximum trust so it requires you know every user in the pool to trust every other user in the pool to not get them slashed um I mean one thing that you can do for example is if you have 30 validators each putting forward you know 32 e um and and there is a slashing that happens more likely than not the slashing that does happen you know will be less than 32 e right because almost all the time um the value of the block

is less than 32 e but like there is technically these like really um longtail uh Edge case events EV where really a thousand a full thousand e gets lashed and now everyone in the pool gets lashed this becomes a little bit better with DVT so with DVT instead of having to trust everyone you only have to trust two thirds which is so much better but it's still a honest majority assumption uh speaking of execution tickets um I was actually thinking kind of differently so instead of this kind of problem going away with execution

tickets I could almost see a similar Paradigm arising with execution tickets where you have these Underwriters and these pools that allow you know uh like solo stakers to pool together to like purchase an execution ticket um I don't know what your thoughts are about that so the value of one execution ticket will be like the expected value of one slot um and that's on the order of 0 .1 e uh at least today and it should go well it's hard to predict if it'll go up or down but uh my guess is it will go down for various reasons and so like purchasing a ticket is is is you know 10,000 times easier I guess

than than putting forward the f e makes sense I'm sorry maybe I missed this does does the underwriter post any collateral themselves yes so they are actually the the one posting the collateral for the pool though so like let's say I I trust Justin to manage it or he's like the the manager of the pool I put my 32th in everyone does you now manage a thousand are you is it between us whether I require you because it's offchain I was just thinking like puffer's model like could uh we don't want to over complicate this but could the underwriter have to buy an execution

ticket so like you know I guarantee there's not a principal agent problem where if the underwriter fails at his job I actually don't have much of an impact kind of I was like think through like rocket pool is this similar or like you know how to puffer address this to get the the alignment right I don't want to over complicated just thinking through is that a a potential idea yeah so like I said before the underwriter is like a trusted entity and um is is you know has to be trusted by their proposers that they're not going to like scond with her eth it's also possible to write an underwriter as a trust list of defi applications you can do a smart contract on ethereum as well um and then you like I said use signature exchange protocols to um and just you have this

your EAS gr so you can essentially say you can write a smart contract that says as soon as I have um 30 proposers opted in and they all restake to me then I will register with the pre-confirmation registry and delegate to all these proposers you can just write that into the code so there doesn't necessarily have to be a trusted relationship in the proposer pool example um but if if the obviously if the under writers off chain then you do need trust to um for them to like handle that eth responsibly sorry I'm a little confused um the underwriter is the one that provides e it's not the one that receives E From Another entity right I'm saying in this example the underwriter could receive eth in the form of restate eth so another Series

yeah right but if if it's a onchain form of collateral then that can just be made Trust with a a small contract exactly yep so I guess one one example could be look I have a punk which is worth a th000 e more than a th000 e and now I have a a smart contract whereby uh if the underwriter gets slashed I lose my punk basically the punk goes to the underwriter because I caused the underwriter to get slashed uh and so now the underwriter

who's lost a thousand they can take my punk sell it and recover their thousand Eve and all of that is trustless so in some sense you can think of the underwriter as just being an over collateralized like loan thing like I have some form of collat which is worth more than the minimum amount of collateral to become a registrant um and uh and yeah the the underwriter is providing the the the ebased liquidity and and taking in as collateral app forms of reputation so in the case of St

um you know maybe you could you could use I don't know a or something you don't necessarily have to go through the under writer but if you're dealing with reputation or punks or whatever some whatever Market is not liquid enough for these over collateralized loans yeah you can go through the underwriter uh Matthew you mentioned that the registrant can delegate to many proposals so you highlighted this one too many um um just as an example in my head so I would have thought maybe we have a um a proposal pool let's say in justtin examples um 30 solo stakers

putting up 302 e collateral through the well by the means of an underwriter so they become a registrant as a pool that's my current understanding but let's say they operate 60 validators in terms of proposers so they would so they would have they would delegate in my understanding uh um the one the one registering to 60 proposers is that's sort of right in this one too many what you're thinking or am I'm is is there like a larger design space I'm yep exactly so the the registrant can delegate to as many proposal as they'd like um it's just any

proposer that they delegate to if that um proposer does a fall and ends up getting penalized or slashed then the regist collateral would be slashed as well so I think the regant is easiest to think about is like kind of a manager of a pool of proposers and they just um are kind of like a logical entity to to think about this collateral here um I will note that the like kind of Delegation mechanics are like very flexible you can do delegation offchain um through signatures you can do uh delegation through like um you can change your delegation over time so if I wanted to add a proposer after a certain am of time I can use a Smart contract to like have a release of a transaction in the future there's a whole bunch of things you can do to uh make the delegation like

pretty smooth and straightforward um just kind of built on top of this this layer Matthew you mentioned that um the exact slashing conditions can be programmable when you can use a small contract um but we we do need to make sure that uh there isn't basically an attack Vector where um a registrant just delegates to every single proposer and configures their slashing contract to be the the never slash aable contract um because then basically now you have a bunch of proposers who don't even know that they're proposers preon first that don't even know they pre-c confers um and uh and you defeat the anti-bible

mechanism so I think what you maybe one possible solution here is that um there's always at least some default or standardized way to to slash if you if you if you do uh delegate um and like my favorite approach for this is just to use um evm bite code um as uh as the language to to to to no maybe I retract that but do do you see the problem uh Matthew that we need to make sure there's at least one default way to to slash if you do

delegate so I think this is where I kind of um models different for a little bit I'm imagining that the registrant would choose um like what slashing contracts they are are willing to except slashing from this is similar to like kind of opting into an ABS you're choosing what slashing contract and therefore what slashing conditions you're like um enabling your your proposers to be slashed on um so and this is this is you're going back to saying like my model has all of the slashing and penalization happening through this like collateral smart contract so I think for you you're imagining this is like registration and taking this a step further and saying what if we just do all the slashing and all of the um penalization through this this eth that we already have um and I want to say

like if we do this this model where there's like a a slashing contract um for for each rollup and then reg can kind of choose what what rollups they want their proposers to to be involved in um and this is kind of be the way to to opt into new rollups because obviously you can't um make permissionless slashing so you can't um require a pre- confer to to set up their um like like slashing with with in a permissionless way or else you could get into slashing conditions that you're not aware of so I don't know if that makes sense and we can we can discuss this um in the future offline if you want um I I guess that's how kind of how I think about it right so so you're thinking of having

the the slashing and the registration be all in one place and I think that that makes sense um but uh we want to make sure that if I'm a registrant and I delegate to someone that the person that you've delegated to um can actually slash the registrant because if they can't then now the registrant can just dos the system by delegating to everybody none of which are actually slash right so I guess what you're saying is you can get to um minimum collateral without choosing any slashing conditions um and this means that you know you have Gadget collaterals but the collateral doesn't actually mean

anything this is where I think I disagree I think minimum collateral is like um I determine what a offer is but it's up to every rollup to to know if um a a a proposer or pre-offer has opted into their um slashing so let's say you could have a proposer that has a collateral balance of more than minimum collateral if they are not opted into the slashing conditions of the rollup then from the perspective of the rollup they're not actually actually have any collateral so I guess minimum collateral is like a shelling point and in practice um should be you know if you have more than minimum collateral and you're opted into a rollup slashing conditions then that should be enough then hopefully all pre-conference are like almost all slashing contracts whatever they feel comfortable with um so there's this like a dynamic of of are you a pre-c confer

for a rollup and the goal of this is to um coordinate that all proposers are are pre-c confers like or the same and of course of course C multiple did I answer your question I mean one thing that we do want to make sure is that every single uh rollup agrees on who the next pre-con is so that we have this notion of of a shed sequencer um and and decreed like this can be permissionless we can't have um it can't be permissionless for a new rollup to opt in um like having their leader done by this program or else they could have

their own slashing editions um and then then there would be like their their pre confirmations wouldn't mean anything they could they could of course still use a leader but they wouldn't be able to offer base pre confirmations right so it's permissionless for a rollup to do whatever they want but presumably there will be some some shelling points some standard um and and we do want to make sure that this this standard is reasonable and I think it would be an unreasonable standard if it was possible to register and delegate and not not not be slash in any way right what I'm saying is that even if you do that then the um delegates that you have let's say you

have um a delegate to to Justin then then Justin would have you know a lot of collateral but from the perspective of the rollup he would not be eligible to to be elected so the leader for for a um for each rollup might differ based on that but of course that's an edge case and and something they should be avoided so if my if my rollups are all let's say have all of the optimism rollups they all agree that you need to be um opted into like the the op stack slashing conditions then those can be a a um the requirement for to like interrupt with OBX I guess it's a the register can choose who they're going to delegate to and then based on how they choose this is what determines like the um I would

call and then there can of course be a minimum so if you do want to opt into the the slashing conditions of a um of a rollup or you want to reach minimum cler then you also have to meet some like minimum requirements and but I don't think the the preformation rry should be like opinionated or permissioned um on like what slashing contract you can use that is it shouldn't be permissioned but there should be at least one slashing condition so the the the the set of slashing conditions that come by default should be non empty I don't see why I mean if you it's okay if you have more than a thousand e of collateral if it's just doing nothing you wouldn't be well no because a sing

if if you can literally delegate to all the validators all million of them with just a thousand e then you've just completely dosed the system except that even though your delegation like exists in his onchain and the the collateral bounce of these rollups is is greater than th so they would be considered promer in like strictest sense of the term every rollup that's actually you know achieving interoperability and being like based and having the shade sequencer effect that you're mentioning um wouldn't can wouldn't elect these proposers these like right but that that requires kind of social agreement amongst the rollups but yeah I agree that can be done but it's kind of a a higher lift from a coordination standpoint how do you mean

why do we need social coordination from theups because they all need to agree on what is you know an acceptable uh slashing condition to be considered to be registered as a as a prec confer for that rollup right so yeah okay so let's say rollup a and rollup B both have their own like different slashing conditions if roll of a sets slashing conditions to a Smart contract for for roll of a and then roll of B sets it to a Smart contract for roll of B in order for these to elect the same leader and kind of be um like form this coordination then a a reg would need to delegate um while saying that these two smart contracts are both allowed and this is like very easy to do if you just

have the register and um choose those to but if the reg doesn't then neither of these rollups I consider those pre-com I guess does that answer your question understand kind of a complicated way I think we need to have a conversation off off chain off offchain offline um yeah any other questions before we move on to the private part of the conversation and close this Public Presentation okay thanks oh there was a

question in chat I was kind of wondering think same thing someone asked about the risk to the underwriter I don't know if we covered that yeah I was just typing your response so the underwriter actually has very little risks um we talked about like collateral dpeg risk but they only have like long-term Cal devra so let's use a LST example an underwriter accepts an lstd from a large node operator um and then they they convert this to eth and and register um so they have kind of this this dependence on their collateral it's it's a LSC um if this LSD DPS goes through some sort of dpeg event maybe a bug um then the a th000 eth which exist onchain but the underwriter obviously wouldn't have access to um the the offchain Coller or excuse me onchain

collateral that um that they would hold so the problem with that is that you know the underwriter um is then kind of like in the hold is um doesn't have like enough value to fulfill his commitments but it's only it's only temporary there's no like liquidations the underwriter can't be liquidated um they're just they're just taking like a short-term um excuse me long-term dpeg risk there's a short-term dpeg um and an lstd dags you know 50% for 20 minutes then they wouldn't be liquidated there's no problem they just have to um hope that the collateral you know goes back up to um cross the the threshold again so kind of the solution that we think about for this is just have the underwriters be um like opinionated about what they um what collateral they

accept so don't accept some LSD that has a high dpeg risk and then you do your kind of due diligence on that but yeah it's a risk to the underwriter and not to the pre-confirmation collateral protocol right thank you any other questions before we close amazing so uh that's the end of the public part of the conversation opening it up to the off the Record uh part of the conversation