EigenLayer: The Endgame Coordination Layer | Sreeram Kannan and Soubhik Deb

Transcript

in one world you have 100 protocols Each of which have1 billion State this is world one in world two you have1 billion Stak which is reaked into all the 100 protocols which world is better the benefit of world two is to attack any one protocol you now need to attack the100 billion in world one to attack any one protocol you only need to attack $1 billion the second world brings the benefit of pooled security the first world doesn't so what we very careful about this convergence hey everyone Sam here if you're like Dan and I you probably don't pay a ton of attention to the institutional slmr side of things in crypto and our laser focused on the dgen

shenanigans happening on chain however in order to on board the first trillion dollars of capital on chain it's incredibly important to hear from these institutional players and get them involved that's why we're excited about hosting the digital asset Summit in 2024 in London the land of tasty pastries be sure to join us at the event next year and we will tell tell you how to get 20% off your ticket a little bit later in the show what's up everyone welcome back to another episode of zerx research before we dive into today's episode a quick word from our sponsor hexin the most hardcore security team in web 3 pioneering in ZK and novel cryptography hexin is trusted by tier one projects like polygon mantle risk zero Lio 1-in newbank and more check them out on the ground at defc connect and be sure to mention ZX research when requesting a

quote and you'll get a free web tube pen test with the purchase of your audit all right everyone welcome back to another episode of zerx research we have an interview episode lined up with the uh gentleman from igen lay today so it's going to be an exciting conversation about some Cutting Edge technology getting built in and around the ethereum ecosystem we're joined by Shri Ram the CEO and shabek the head of ABS research uh so doing a lot of the groundw work there on the R&D side of things so thanks a ton for for joining us today got an exciting conversation lined up um but again when we have these uh conversations about like this really novel and nent Technology I think it's always best just to start a little bit high level before you tear back a few layers and get into the weeds of things

so you know you launch back in June and already have over 200 million in TBL between about 21,000 unique depositors which is a very very impressive Milestone showing some serious demand for what you guys are building um but if we could just start from the top and kind of say all right we're building I igen layer to solve these subset of problems one way to think about it is imagine you're back in 1995 and you know you look at we application development 995 um you'd have maybe two options you could go build a website you know most of you here are too young to know this but you could go host a website on geocity which would be a static web page you just have your homepage put your photo up right something basic you don't

have to do much work to actually uh do that but if you wanted to build anything more complex you want to store complex State you want to have identity login payments any of these things things in 95 you would have to build everything yourself you have to build your own service stack you have to build your own identities payments you have to build your own database tag and then you build whatever thing that you actually want to build so that that's the state of web development in 95 if you look at it in 2023 it's absolutely not the case what you would do is you would say hey I'm not building my own servers I'll use AWS for you know a lot of it you'll say I'm going to use um o for authorization

stripe for payments bongod DB for a database and then you know focus on the V new thing that you're building you know a stat is that like pretty much know an average consumer application uses 15 Tas Solutions on the back end right not one not two and people talk about oh you're going to have many layers in the modular ecosystem and you have to pay each of these they haven't seen how web development Works each consumer application you 15 SAS services in the back end this is already the way the world works so the the thing about these SAS Services is each of them something very narrow very powerful right just payments I want to solve payments I want to solve

authorization for social I want to solve something very very narrow but solve it in the most powerful way possible right you know if you look at basically the Arc of civilization it's always bends towards specialization and fast Services demonstrate you know an example of that and if you look at the crypto economy today we are basically you know between 1995 to 2000 so if you want to go build an application which will fit as a smart contract on an existing blockchain like ethereum you could go ahead and do it right you could just go like customize your contract and you know deploy it that it's easy it's simple you know but you're constrained into the into the

language environment into the set of functionalities into the gas limits into the uh you know uh the particular distributed system architecture that you know underlies these systems and then you could say no I don't need to do it I can actually go be my own thing and let's say build a cosmos chain you have to build your own Val Network you have to build your own underlying token of value you have to build your own Tack and customize it heav there is open source software you know but software doesn't is not the same as just a bunch of SAS components that you just choose custom parameters and integrate right so you have that and then you can go and deploy it so there is a powerful option which is to go build your own chain uh there is an S powerful option just

deploy a contract but can we get to the SAS type world where you know there is a common place there's lots of protocols on top and an end user application simply concatenates a bunch of these protocols and then you know deploys it so that's our vision for what we're looking forward to and the the first difference you know when you think about like the difference between the cloud era and the blockchain era The Cutting you know hyperplan the dividing line is decentralized trust so if you're running all these things you know any one of these components depends on a centralized validation then you lose the whole benefit of decentralized trust so

in order to build the crypto economy all of these have to be run on decentralized trust Network and the the thing with security security is the weakest of all the things that you depend on so our vision is shad Security on the base layer so you have you know the ethereum validator set I that makes it more flexible and more programmable that's the base of Trust on top of which anybody can build permissionless Le new protocols as a service right so basically you just write a protocol you know the the way you deploy a s Service you you you write your code and then you send it to AWS and AWS takes care of everything right that's roughly the world that we want to build you write

your protocol decentralized protocol tip it to igir ire takes care of where the trust is coming from making sure that the trusted notes keep to what they're needing to do and then you can now start building everybody can start building their own specialized you know highly powerful services and then now you concatenate all of these services and then build an endo application that's the world view and vision that we are actually kind of building towards that's awesome and so what you're powering this from a security standpoint is with stake eth and restake eth so uh just I mentioned that 200 million is in tvl uh just looking at some of the statistics

here and breaking that down about half of that is lto stake deth about a quarter of that is rocket pool and CB stakd so coinbase is uh LST and then the that last chunk there the last remaining quarter is native Native staked eth which is a really interesting mix there uh one question I have is what is it why only limit this to stake ether because ultimately what this is really doing is just putting some um like respected value at risk it's sort of like in some ways it's kind of similar to a pro proof of stake chain but what is like why not open the door into a larger basket of assets maybe like a usdc or if you wanted to remove some of that centralization maybe like a die like what is it just about e that kind of makes you excited to build around that

asset I think the way we think about it is the it's not so so important whether it is eat or usdc or something else but it's important to build a category of asset on top of which you know the same stake Powers all the different Services I think that is very important the convergence is very important it is because you know convergence is what gives you pool security so the way to think about it is you know you can have so I'll give you two like different worlds in one world you have 100 protocols Each of which have1 billion do state this is world one in world two you have1 billion do state which is restak

into all the 100 protocols which world is better okay if you ask this question then you know what is the benefit of world two the benefit of world two is to attack any one protocol you now need to attack the100 billion in world one to attack any one protocol you only need to attack1 billion dollar the second world brings the benefit of pooled security the first world doesn't so what we are very careful about is Convergence and you know the natural place because we're building on ethereum we're aligned with ethereum values the natural place to build this on is ethereum you know staking and so that's why we're focusing on that as a kind of a natural shelling point because what we don't want to have is some some protocol

saying I only take Ste some protocol saying I only take like usdc somebody saying I want to have bonds St you know these will all happen in due course and time but what we don't want to miss is that big chunk of steak which is then Shad security for all the different services that is very important so not so much you know that's why I think you know even the word reaking is a misnomer really permissionless programmable staking that's what we're building when you're staking right you're normally staking into a condition that the protocol kind of enshrines into the protocol and what we did is just break that open and say it's programmable and it's programmable permissionless anybody

can come and program what the sticking conditions are and I think that is the core thing that we building so it's absolutely true that we could open it up to all kinds of assets but it is basically this Shad security versus dispersed security that we want to avoid that's why we have the Dual staking model where if you're building a new service you can rely on you know some portion of security coming from Eve and some portion of security coming from your own but we do want to keep that portional like Shar security coming from a common pool that that's what like makes this really partial all right everyone I'm excited to tell you a little bit more about crypto's Flagship institutional conference the 2024 digital asset Summit in London the UK's

open regulatory environment has laid the groundwork for growing institutional adoption Das brings together over 600 institutions that manage a collective $800 billion in assets so there's no better place to hear perspectives on digital assets than straight from the world's largest Capital allocators this year's speakers lineup is fantastic and we're going to hear insights on macro Trends rwa adoption and utility the fluid regulatory landscape the evolution of stable coins and more between the All-Star speakers and the relevant topics d224 is a can't miss event for anybody interested in the growth and maturation of our industry hit the link in the show notes to secure your tickets and be sure to use the promo code 0x20 for 20% off I look forward to see seeing you all at D London in 2024 I had a sort of follow-up question to that concept of shared security right so today Ian layer

has deposit caps for Stak de rocket P CV and sort of native Stak de I guess there's two questions there first will we ever reach a point where there are no deposit caps and individuals can deposit as much as they want and the second question is in that scenario what if for example for some reason there's some Market dynamic where like 100% of the restake Thief was only like Lidos Stak Thief would you see that as something that is either problematic or requires intervention yeah um you know these are complex uh you know dynamics that we cannot predict so it's possible that

something like that happens um and we have to steer the protocol the you know ethereum is you know as as a protocol I think igen layer is somewhat different from ethereum you know right now for example ethereum wants that you know Lio should have a you know good share but not like more than a certain percentage that's what ethereum wants but ethereum has no tools that it's disposal to actually like get that outcome whereas you know I think I like the goal is to create a Shad security mechanism for decentralized trust so we will have time mechanisms to shepher

this you know in in in various ways because there's going to be like a decentralized governance process which can actually exert judgment on these decisions because it is really you know from our point of view the igen mayor governance over time as it becomes more and more decentralized and involves various stakeholders has to balance the different sides of the marketplace right so we have stakers on the one side we have node operators on the other side we have services on the other side it's and all of these three actually you know you can even think of it as takers operators and AVS you know service creators AVS is this idea of actively validated service which is anybody can build a new protocol you can even think of it as

like protocol as a service right just like the SAS economy in in in the cloud which is basically protocol as the service you can take your protocol you know throw it into I and IG makes sure that it's being run through the decentralized validation mechanism but the these these three segments you know sters operators and abss are coordinating to offer a service to the consumer of these avss right like imagine if the AVS is an oracle somebody's consuming the Oracle imagine if it is some other thing that is a different consumer but the important thing is I think governance serves as a place to coordinate between these different

sites right so that is what the governance' goal is and to maximize the long-term utility of this platform you know the governance has to engage in complex you know uh negotiation between these multiple sites and that's the nature of the platform and you know in fact we had already stated that this is the nature of our I platform is whether you know when somebody's building an ABS or a service whether they treat all liquid staking tokens equally is up to the service right that's that's an example of how we let that side express its preference you could say I only want 50% of the state to come from SE somebody could say that right as if they're building a service they have the power

to express or somebody say I don't care I just need capital capital is capital right but the the point is capital is only one side or one aspect of decentralized trust decentralization itself is another aspect of decentralized trust and for a given service you know I given this example before but I think it's a pertinent example imagine I'm doing Secret sharing I want to take a secret and then spread it into small portions that are held by different notes I want these notes to be as like independent as possible it's not enough that they have sted a lot because you know what if they all meet in a private room and dis disperse that secret to each other I wouldn't know and to make it as collusion resistant

actually having many distinct parties as possible who are normally not called ordinating in any way is what brings us that property and so somebody building an avas might say I I want very very decentralized notes and that's the right kind of thing that they should ask for it so over time like the uh ID becomes a coordination mechanism for people to express these complex preferences and to actually you know that creates a certain derivative for or certain pressure for certain amount of decent Iz ation and you know and it it depends whether people want decentralization or not and if people are happy with their 30 no then or like

100 nodes or you know maybe next year Lio becomes thousand nodes and that's good enough we don't know these are all complex Dynamics for the marketplace to figure itself out what's up everyone as we explore today's blockchain landscape let's take a moment to recognize hexin the premier cyber security provider in web 3 hexin is trusted by tier one projects like polygon including a Security review of the new polygon DK evm mantle risk zero Lio 1in newbank and more get a deep dive into your technology stack with the most comprehensive analysis in cyber security Consulting with over 555 billion secured they cover everything from Smart contracts to blockchain to web two pentests there has been nearly 77 billion of total value hacked in crypto's nent history so it's safe to say that your team has a lot on the line

don't skip out take your security seriously and choose hexans don't forget to mention zerox research for a free web2 pentest with your partnership and you can reach out to hexin at hex.io or find them on the ground at devc connect without further Ado let's get back to today's episode so I I want to dive into kind of some of the interesting research you guys have done around the ABS landscape but before we do that one of the things you mentioned was kind of the parties involved in this um in this Dynamic here so you have operators stakers and the abss themselves uh I think stakers are pretty straightforward and then abss are like these applications that getting built on top of this technology but can you expand a little bit on the role of the operator and what exactly that role entails as well as who kind of plays that role

yeah um I know you you can think of you know internally we think of igir as the coordination engine the coordination mechanism for open Innovation so the the principle the goal that we optimizing is open Innovation we want people to be able to come and build complex things with very low barrier to entry so innovate because you know Innovation is I think one of the most prototypical nonzero some games it's it's the way in which you know you you take something that you know that doesn't exist you know it's not a resource and then make it into a resource like taking oil and making it into energy taking you know air and making it into spectrum and so on so you have you know that's the goal to be the coordination infrastructure

for open Innovation and um if you think about the role of operators right so so broadly if you think about what is needed to build a coordination mechanism for open Innovation think about oh what are the V who's bringing the resources who's bringing the capital who's bringing the ideas right like that's the three different components that go into you know the stakers are bringing in the capital to underwrite and promise that the service is working correctly The Operators are bringing the resources and the infrastructure to actually run these things the computational infrastructure and then the um um services are the the AVS creators are actually coming up with the ideas and the code that actually

make this work and all these three are coordinating they're working together it's this code plus this stake plus this node operations together is what makes the service so again there being the coordination mechanism for all these three sites to converge and then offer a service so that's the core uh core principle so if you're looking at who can be an operator an operator is a permissionless role in the I there ecosystem so anybody can be an operator you can be a Seer and you can be your own operator and you delegate to yourself but you know you may you may be like yeah no I already trust ABCD for some other purpose you know for custody for some other thing that I'm already using whether it's a coinbase whether it's a figment whether it's a P2P like

whatever is the kind of operator that you like you already using them you may say yeah I'm going to delegate to them and uh the the trust for delegation is not in protocols so there is no mechanism that says that this delegate's going to be nice to you or not nice to you or anything like that it's like you know you find like a service provider who is actually like trustworthy and then you give them to actually run your services for you so that's the delegation mechanism other people can build more complex coordination mechanisms on top and you know we'll see some of them come up um for example there is uh people building um trusted execution environments on top of I layer

so you can run inside a hardware Enclave the the software you know particular l a software we call anti- slasher which basically simulates what the slashing contract is going to do and then tell you that okay you know if I sign this I'm not going to get flashed and this is done inside a trusted execution envirment that means that as a consumer right as as a Staker I can actually trust the a node operator without knowing who they are so there are all these ways in which you can reduce the barrier of trust between stakers and operators but this is a market mechanism this is not in protocol for I so that's so the role of an operator is a permission list anybody can you know

volunteer to provide like nodes and run various services and you could be your own node you can delegate to somebody some AVS some services will have caps on how many nodes they want to participate because of you know if you're running a secure multiparty protocol like the communication complexity depends on how many nodes and so on some protocols may have no limits it's it's kind of taking the like a Middle Road between like the cosmos approach of like in protocol delegation to the uh ethereum approach of like I I don't even know if you're delegating I know you're delegating and I want to bring out so surface information and optionality that is the igen lay view is you want to surface the

information that these are different node operators and then like as an AVS you can decide you know how to reward for example a service may say I I'm going to cap like a per operator reward to like you know 5% 10% whatever that is so that already drives different kinds of behavior so you can have very sophisticated Market Discovery for what kinds of things work as opposed to enshrining these in protocol gotcha yeah no I think it's really interesting how like I layer seems very simple yet at the same time it's just very like complex like coordination game between various different stakeholders right um and so one obvious question that I have to ask is that governance is a very important part of coordination and so the question

is how will Ian governance work you know I think right now at least for let's say deps we're very used to a one token one vote sort of governance model and sort of what does governance for I look like whether that's now three months later three years later I we are in early stages so I I don't think we are in a position to dive deep into it it but it is something we want to give a lot of thought to about how to bring these different parties together what are the inputs we need from these different parties how to make sure that you know it the um the coordination mechanism remains solid so but but it is too early

to to do a deep dive you know hopefully in I or somebody else can come back on and do a deeper dive on governance but it's too early for us to to talk about it it's it's in R&D right now okay that that that's fair and uh yes absolutely we'd love to have you back on to kind of dive into things we it was always a pleasure getting getting some of the the latest and greatest with aan later um but diving into the you know what can be built with this kind of side of thing so maybe shopic if you can give us a rundown of some of the cool research you've been doing on the ABS side of things and what that landscape is looking like as well as some of the potential things that are either ideas at this stage or actively getting

built yeah so one of the major category of services that people like people are trying to build is something called rup services so basically there are like various aspects of rollups uh which you want like for which you can inherit security from ethereum vler so for example if you want data availability so we are building igda inh house uh which gives like hyperscale data availability while uh while inheriting security from ethereum at the same time then there is decentralized sequencing for example espresso is L building a decentralized sequencer by like also having a dual staking model on top of igen layer so that's one more uh there are like certain other protocols um uh certain

other teams who are building bridges on top of I for example uh lran they are actively building the outbound bridging so that's a very cool project and and like they have consume they are also trying to get consumers for that bridge so this this is the aspect of the AVS consumer side the fourth aspect that yeah that remains hidden from most of the discussion but that's another category of actor that people we need to care in general and then there is a something called Fast finality layer basically in ethereum right if you want to get finity you have to wait for uh twoo in on average but can you can you get finality much faster so with I you can get crypto

like an economic crypto Economic Security uh towards that kind of fast finality and so this is like a role of services uh there's also people are also building Watcher networks uh there's uh so there's this team uh from Princeton called witness chain who are building u a watcher Network for uh for generating challenges to optimistic role of services if there is malicious execution and and something it's just one aspect but there can be watch Network for many other proposals many other uh protocols then people are building kind of Applied cryptography oriented applications like FHA or secret sharing that shiram was

mentioning so uh how to distribute shares across a decentralized set of actors and while also trying to prevent collusion among them then uh there has been like a recent uh interest in trying to run AI inference engines on top of uh Ian so so here you need to guarantee so since it's a decentralized like there are like lot of actors who are running the inference engine so you need to guarantee against program integrated like they're running the right model that you want it's not like something other some other model they are running with some other parameters so there is this uh program Integrity you need um so some people are building projects around that so that would be really cool like

how a how Ai and crypto can like like the marriage between that Ai and crypto it will be really awesome and hopefully in future people can extend it to Fed like Federated training uh not having like uh one entity do all the training and get all the models but have data sets coming from different people having a market model around data sets also because if you have data you want to make money out of your data right you supplying data to someone to train your model so you want to make money out of that so having like economics around that figuring out it'll be really exciting also I love how there's a couple different categories that you're all working around and the AI stuff is pretty interesting the I just saw like Kos we recently talked to Greg Ai and they're uh training a model fully on a

fully obviously on aash so it's like a fully decentralized model being trained for the first time which is super super exciting so it'd be really exciting to kind of see how this pairing between crypto and AI really evolves but uh one of the other interesting things you mentioned was what espresso is doing with their uh the Dual staking model can you like zoom in on that specific example and explain what the Dual staking model looks like and why the this is really beneficial for espresso so uh in their dual staking model so they have a different a parallel consensus protocol going on they call it hot shot so now in Hot Shot there can be two types of nodes one is the uh one is the native e Staker nodes and another could be their native token uh nodes so suppose you have put X you

have restake X8 and opted into into the uh espresso into the espresso protocol you have some corresponding state an a stake weighted in the consensus protocol if you put X native token of espresso then that will be like some um stake weighted around that and the consensus protocol will like consider that stake weighted stake and and whenever they it's finalizing the CH uh a proposed block so it will consider that weighted stake as the like the value for deciding that has the 66% threshold has been achieved or not on that newly proposed block so this is the Dual staking model at a like a

architecture level so what is the benefit they're getting so if uh with with getting by just having e reaking there like you can get benefit around like uh if downward spiraling of tokens so it's a protection against that so you want to protect the so there are two aspect there's a protocol there is a token right the native token of the protocol so you want to secure say make the protocol safe so you can use the E to make the protocol safe and there's the token aspect of the protocol which like by using it as part of the Dual staking uh you can like bring additional utility to that token so so both the benefits are

there following up very quickly on that dual to token staking model right um sometimes it feels to me that value of curo in crypto is a zero some game you know um for example espresso only use fre s deep then all of like the sequen of profits would go to for example espresso and for example if there was $10 million of sequencer profits between like the shared sequencer that connects to for example arbitrum optimism then perhaps like only five million can go to arbitrum five million can go to like um whatever token espresso chooses to use how do you guys think about this sort of like value AC cre problem with igen ler or in like a generic like dual token

taking model so when you have a dual token usually the native token is what is performing the governance on these systems right and you know you can ask what are the benefits of uh you know the first point is because you know that is the token of governance you could you can you can see that uh over time if there is you know no benefit in using dual staking one can steal towards more and more towards a single staking Market model right so one of the nice things about item layer is it's not a binary switch it's not like do do I need like single staking or dual seeking it's about like how much dual seeking do I need and you know you can say yeah I'm

only giving 5% of my fees to like each sing you know it depends on what you're getting from it okay so that's the first point so because the governance token is the one that's going to have control over like how much fees is going to flow between these two quorums really the value acral is you know if you look at the long-term Horizon the value acral is fundamentally towards the governance toen it is not towards any other system so other systems are there only to the extent that it adds value to them right to whether I layer or something else there only to the extent that it adds value and that's I think like the correct principle for building like open marketplaces is that to the extent that they provide value they should be using

it to the extent that they don't provide value they should not be using it and the thing with the dual quum model is and then you can come to the question of why should somebody continue to use dual token model you know in the long run versus just use it as a bootstrapping mechanism the firstly there's value in it as a bootstrapping mechanism right so you know you want to get you know more Economic Security than you can offer by yourself but there are also other reasons to keep the Dual sing model the uh eventually for example imagine that there's a whole bunch of services that you know a given application uses and all of them have one one one site of the of the security coming from each

staking and the same large like token commun token um stakers right we taking into all of these things then you know that to compromise that service to compromise that particular application you have to compromise that Corum that's needed so you are actually this becomes a crucial trust assumption that you know to compromise any of these um to to compromise this application I have to compromise this large eat token Quorum not any of these separate token quorums because just doing that is not sufficient to compromise the system so you have this chat security benefit so over time if this shat security benefit is not adding up to what it is they yeah you know I don't need this shat security

I just my own way totally that's that's acceptable and so fundamentally I think if you look at the value of the the future kind of value of the token which is due to the Future cash flows or profits or whatever you know the mechanism I think it doesn't fundamentally change I is just optionality to increase it and to also protect it against potential death spirals so one of the death spiral of like you know endogeneous staking is as your Stak goes down in value and your you know the economic security goes down people consuming your service may want to disconnect because you know if Economic Security is what they were

relying on the first place now it has less economic security so less ability to support new applications and as applications disintegrate out of the kind of service you actually have a kind of positive feedback cycle in the negative direction right so you basically kind of can collapse whereas with reaking you have like one side which is provided by each staking and you know even if your go token goes down value you still have each value and as long as you're able to pay enough to compensate for the opportunity cost of that Capital you know people are not whereas in the in the native token Corum people are holding your token and getting exposal to the downward price

moments whereas in this one they just need to be paid for the opportunity cost of like reaking into your service and so I think it is a protection against de files that may occur in protocols it's a protection it you get the benefit of shat security if one quum is common across all the different Services you get all these benefits over time if you know if those benefits are not significant uh you know you may start getting composability benefits right so because all of these are in common trust trust Zone you know an application will be much more interest you know interested to integrate because it's a common trust assumption to integrate with all these services so you know these are the benefits that you get by sticking to I layer in the long

if it takes off right so that's that's the answer yeah I think it's really interesting how like I ler can also act as like a backs stop there and I think it's important to remember that abss like do have the power to sort of choose how much like if is contributed as sh security to the value of the network right and there's another interesting point you mentioned there that if it for example doesn't make sense for a middleware protocol anymore then they can stop using I ler at any point of time I wanted to ask the opposite question of that is there like a service level agreement for how long a validator has to secure a protocol leveraging ion layer or can the validator stop securing it whenever they want um thinking of

like a slightly ridiculous example like there's a large Oracle protocol they do something that's like really malicious um or they did something that's like really bad socially in like the real world um and then all of adors say like that's not good behavior I don't want to align with that anymore I'm going to put out right now how how would something like that play out so right now our mechanisms are instantaneous which means you opt in you have like a service period or an unstaking period you have to stay for that period but beyond that you don't have to stay uh but we are already doing Rd on longer term commitments so I stti and then I make a promise that I will stay not only for like this period but also for a much

longer period why would I do that now you would give me additional APR or whatever because you know you need some amount of s to stick with you through like you know the the the good and bad so you can actually have these you know mechanisms programmed into uh I already like out of protocol but you know will'll start giving more and more functionality in protocol for these things it is somewhat similar to reserve band with in Amazon ec2 so it's like reserving security for a longer Horizon yeah so so the idea there is that you um when you stake you make a commitment that you're not going to withdraw for

some amount of time from these salices and so by making this additional commitment you get additional payment and uh in an out of protocol out of I way to do this is to say that you're giving rewards asymmetrically to sers who have been around in your protocol for longer period so that create faster incentives to to remain but through I you can also if it becomes in protocol you can also enforce that you cannot go anywhere else so that is you know these are uh things under research but over time they can be done uh to answer your other question which is whether validators don't like somebody and then they can uh unravel um I think that this is not the

default uh type of thing that we expect to happen the the thing that we expect to happen is um people and stake because there is a more U lucrative staking opportunity and we think actually reaking lens this kind of like a dynamic because you know you you don't have to choose between sing to service one versus service two you're sticking to all the services that you actually want so you're not choosing between sing to service one versus service two so your capital is being allocated across all of these Services the for the dynamic DC with I is if you're operating cost if you're

not able to make fees more than your operating cost then you absolutely will opt out and I think that's the right thing you should do because like why you wasting your time and energy doing something that doesn't is not worth your while and so but that is a very different uh level of when you will opt out than if if it's just like a race between hey I I have like 3 to each s I have to choose between dance protocol and R protocol and chix protocol like that puts me at like oh shik's the most profitable screw down and R right like that is exactly what we don't want that's what reaking I that's why I think Shad security is the right framework for thinking about it rather than you know reaking the idea is that I have a pool

instead of like securing you know 100 million to Dan 100 million to Ren and 1 billion to sh because Sha's more attractive I'm going to kind of like put the 1.2 billion in all three of you so you know know all of us are better Ross could you just elaborate a little bit more on like exactly how that mechanism Works cuz like I don't think I'm following how you know one deposit equals you know I'm securing as like a pool versus an individual operator or an individual AVS I would say yeah yeah so you know imagine there's like simple numbers like 10 billion and 10 billion is you know and there are 10 protocols there are two different models

one can imagine you know and there are 10 protocols each each protocol gets allocated its own like sub tool says oh you know this 1 billion securing that protocol that 1 billion securing this protocol and so on and you know they they have no like relationship with each other this is one model of the world right in this world to attack any one protocol I need to acquire a capital of you know half a you know half of 1 billion and then go and make a profit more than the half of 1 billion because I'm going to get slash for that half for 1 billion right but in the pulled world what happens is I have $10 billion stake then every Stak is reaking into every service that's what it means by Shad security is every Staker is reaking into

every service this is not a default or a requirement in I live but if this happens you know this could happen because people want to you know participate in many services and if computational model neck is not limiting them they will restake into many services so if they restake into many Services what's happening is that every Staker is reaking into every service so if an attack happens on service one or service two or service any one service I'm going to be able to slash that service will be able to slash you know all the violators which could be more than5 billion so that is what I mean by poed Security is to attack any one thing you have to attack The Collector so this is

stronger together right it's exactly like you know how Nations kind of like cities come together into a nation and like the nation is stronger together because there will no attack our army for the entire nation rather than my own like cities are right thing that happens with this kind of pool security is you're missing out on attributability let's go to the previous example let's say Dan had like a100 million Stak Ren has $100 million stakes and shaik has like $1 billion Stak now I create a pool and this pool has 1.2 billion it seems like it's a good deal for Dan and ren but somewhat worst deal per Shing because he had 1 billion separately now he he has 1.2 billion shared between these these three individuals right the

thing is in version one of our protocol we have no attributability like ethereum has no attribut no blockchain today has any attributability that for example in ethereum if you know there's a double spend in ethereum and you know Unis swap got asymmetrically attacked or only Unis swap got attacked or only coinbase got attacked doesn't matter some in some party got attacked now the the slashed funds are just like burnt indiscriminately it's not compensated to coinbase or to Unis swap specifically so there is no attributability and that's how I V1 is I V1 is it doesn't matter who triggers slashing the condition is verified and if the condition is correct all the

stakers forms are B right so it's basically like indiscriminant it doesn't it's just like you know it it harms the attacker but there's no specific like um redistribution to like dance protocol or rents protocol or shx protocol so it is kind of smudged into a pool uh in V2 we have like upcoming mechanisms which actually solve this problem where what you say is you have a pool of $1.2 billion St but Dan buyon so the slash funds are not like just birn maybe some portion is burn but some portion is redistributed and how it's redistributed is based on slashing prior priority that Dan bought and ren bought and sha bought maybe Dan bought a slashing priority for

100 million Ren bought it for 100 million and then cha brought it for 1 billion and the total sashing prty sold is less than the total amount at stake so essentially you can get the benefit of attributable security you have your one you know 100 million to slash and ren has his 100 million to slash and CH has is one million but to attack any one protocol I still need to attack the collective so you can get the benefit of pool security and the benefit of attributable security Allin one so we we think this is simply like Superior to anything out there so this is actually the right way to build Shar Securities I appreciate that that makes a ton of sense now and is it is it opt in for the Depo uh the stakers saying okay you know shavei has this great protocol this

great abs I'm I'm interested in actually helping secure that with my stake but Ren it's a shitty protocol I want nothing to do with it so like can I opt in as a Staker so Staker has opt into to like which subset of Services they are securing but they how to allocate the insurance or slashing priority is not up to them because you know your commitment is you're a Staker you're opting in if you you do something bad your money will be taken away that's your commitment to the protocol it's up to the protocol to figure out like how to take all this money allocate Insurance to different parties so that's that's up to the protocol not to you so on the topic of slashing I had one question you know there's a few ideas of avss out there one of them is

the fast finality chain where you get like it settles uh settles optimistically on ethereum and it verifies it through ZK proofs and arrives at consensus within seconds um and another thing that's enabled sort of through ethereum inclusion trust is is single slot finality and both of these things are sort of like within a block things um for lack of a better term on ethereum Main net um so my question was when someone is slash an igen layer how long does it take for for ethereum to know that that has happened and for these like abss which require what I would kind of Define as like low latency within a block uh confirmation of like security so to say is that like a potential problem if the slashing like

um if the slashing event takes a long time to like be transmitted to the base layer the slashing the the SL all the abss in ag they're kind of optimistic in nature so whenever a task is uh done so there is a challenge window and uh within that challenge window anyone can raise a uh put a put a submit prove that something malicious has happened and if that is passed then the person then the particular operator gets uh like slashed so this the moment the challenge window like moment someone has done put up a successful challenge that moment itself this slashing like that slashing gets finalized so that's the that's the like

a uh basic model but we have something like VTO committee which Vos the slashing which introduce like certain other delay that VTO committee has to check is this slashing request being like uh is the challenge uh the slashing that is being requested is it correct or not based on like if there a software bug or something maybe that's the reason that's why uh something like slash event has happened so the so this is so this is the additional DeLay So this additional delay that we are introducing like we'll keep it as at a level where which is acceptable to most of the like uh to the stakers because at the end of the day this veto committee is acting as a safety net right for their protection

against some kind of software bugs or something like that so if so in a manner of speaking you can say that it's instantaneous uh the slashing is instantaneous the moment when the successful challenge is done but we introduce a small delay uh which would be acceptable for in return for the safety net I guess I have a quick followup question to that um for these like uh abss which sort of require ethereum inclusion trust right for example single Sal finality would that require a super majority of freake De to achieve that or would it not require and if it does require like a super majority of fre State de to achieve that what

happens if for example like it starts above that super majority threshold and then like it just fluctuates like below and above and below so yeah the ethereum inclusion trust you're right there's a quality of service that's uh that's dependent on like how many uh stakers or the ethereum validators have opted into that particular service so for example if we uh if we look into Event Event driven Activation so suppose this service says that uh if your collateral is underwater they will fill it up they'll trigger a transaction they will fill it up if only let's say 5% of the stakers or validators are opted in you this service is very bad like once in 20 blocks maybe someone will come and F and by within

that 20 block you might be under ceriz you might get liquidated so it's very bad service so you want say 80 90 yes so that's that's that's a that's a caveat with the ethereum inclusion trust but there are certain ways how you can get around that like for example uh uh you can you can have the uh you can have the daps execute transaction s only if they are signed by a by a pool of economic trust takers outside the outside the ethereum so there's a service uh who says that oh if this transaction we have attested to this transaction only then the DBS execute now who are these people

who are running uh running this like attesting to that transaction they could be the a of validators uh and they could have economic trust that they are they that I'm putting my stake to attest to this transaction after doing some du due diligence or some check and so that's a way to get around the problem of economic like the problem of like high optin is required in the econom uh in the eum inclusion trust so that's that's that's one way yeah so you mentioned like it seems like there's a lot of like inherent Network effects um in terms of the number of valid datas for these ethereum inclusion Trust abss right um and then know this other research report

that you guys have written out for example like me Boost Plus which is the whole idea of like partial block auctions and meev boost Plus+ um eliminating the trust Assumption of the Reay and at that point it kind of seems like there's a lot of these avss which benef which would benefit greatly from a large amount of steak th that's currently Stak at the base protocol layer being weake deep um so the obvious question that I think is also a Hot Topic right now is ement of I layer right um at that point how do you guys think about ement layer given all of these like sort of ethereum inclusion trusts and what does ement of aan layer mean to all of these stakeholders right whether that's the current people that

opted in into aen layer the people that stake their LSS into I layer or for example even investors to understand what the implications are to the VAR stakeholders you know first we need to know like what is what are we talking about by when we mean enshrinement what does it mean what things can theoretically get ens shined what things cannot get igned there's a lot of like Nuance to this um the uh starting point of this is to understand that uh you know the uh so first like why would we want to do you know ignment I think

maybe that's a good starting point is if there are certain things that you can do if things are in protocol that you simply cannot do if they're out of protocol that's a kind of a good reason to think about Ur um one example of where that is the case is this notion of you know ethereum inclusion is if you want to like include transactions into a block and you want to make a promise with I you can only get that to be cryptoeconomic right the the idea being that you know you will lose your eat if that that inclusion condition is not valid whereas if you make it in protocol

into ethereum itself then it's possible that other people can verify that you've satisfied the conditions that you need to satisfy so that gives a higher bar of trust than what you know a of protocol can give so yes so there are some situations where and so enshrining makes sense when you get more Powers by being like default part of the protocol versus not a default part of the protocol um and this is an example eum like inclusion guarantees if if you just take this and this would be like amazing over time to figure out you know how to inine some of those into ethereum uh there's a lot of complexity in in this for example how do

you specify an inclusion condition how much like how do you meter whether the gas for whether all the inclusion conditions have been satisfied because you know you move from one of the reasons ethereum has uh like this user driven model is you only pay fees for what is actually being used but if you're had inclusion conditions which are theoretical which means you have to check them like whether a limit order has been paid every time who pays gas for it what is the gas model how do we make sure that ethereum has enough like bandwidth and cost to actually you know to to to check all these traditions on a persistent basis there is a lot of complexity in these things and what's

the language and semantics in which it's done so one it's theoretically possible to Envision in like a five to 10 year period that you know you can actually create like complex CK proves that you've satisfied all the conditions that need to exist but ZK proofs cannot be done for things that are not provable like for example I I have to include any transaction sent by Dan in the M pole let's say that's the inclusion condition like you know how do you know the transaction is in the me Pole or not so I think this is such a complex thing even we're just focusing on ethereum inclusion trust like there is a complex Topography of ideas and and and conditions under which these things work

so I think the right mechanism is these things get tested in out of protocol and then you know when things are working well but you can add a lot of power by making it in protocol that's in fact what we do Envision is even for ethereum inclusion trust find narrow use cases which are really powerful but are not like complex and those get ins shined rather than some very general purpose thing which we don't know how to meter how to cost how to make sure the complexity is contained how to make sure that programming errors like you know here is one thing we can do out of protocol which is very difficult to do in protocol is like a slashing VTO committee right it's a subjective committee like we starting with small

group but you know we could expand the group size of men but the problem is if you want to make it in protocol and anybody can opt into all these like complex inclusion conditions and imagine they opted into something crazy and then like every everybody's stick blows up that's going to be a disaster right or ethereum Star you know livess is affected because you know you're checking for some inclusion condition but the inclusion condition creates a deadlock you know there are all kinds of crazy things that can happen so you know so this is just for one damage in clear communication class we are building a platform like I said which is very very generic you want people to use this for AI you want people to use this for secular multiparty computation you want

people to use this for all these three things so you know we have to take steps you know one step at a time to understand you know what what are possible but we have that's we have very very open to things that are more powerful being native in etherum and this is you know it fits with like withal and he have thesis of minimum viable ignment right you don't want to go so complex that you're going toine everything you're going to go choose a point where it is you uring the things that if they're in protocol there's so much more powerful than if they're out of protocol you know so that is really the kind of thing that we envision will happen but you know the reason we are not so

worried you know you mentioned various categories of people that you know what happens to take what happens to operators what happens to Services what happens to us you know our investors or something we're not so much worried about the last thing because you know we invented this category we building lots of use cases that will consume you know Shad Security in a way that simply you know other people may not understand how to use so we'll have things to do and I our vision is f to thumb games you know the better the coordination mechanisms the better the Innovation that we can build the sum total of what we building as a society as a community expands and that's what we care about the most so you know happy to see like any aspects of

our protocol become part of it okay that makes sense yeah on the idea of enshrinement it was always like tough for me to understand just because of that exact point of like you know this at the end of the day it has to be a business and it has investors uh but if you're focusing on kind of like the AVS landscape and saying hey like nobody's better qualified to build the best things than the people who have built the the underlying l that makes a ton of sense to me as well um and so maybe just as a working towards the close here is kind of we're coming up on an hour um when can we expect to see some of these abs uh protocols go live I I know you have an interesting relationship with mantle the L2 launching on ethereum uh that's going to use Ian da so just

wanted to get your take around that of like when we can expect to see some of these really exciting abss hit the market yeah um right now uh we are uh you know on the main net there's only s right there's no delegation node operation or services so we are working towards a test net sometimes this quarter which would be which would basically be a public test net where we have you know all the sides of I and IG da running so that would be a major Milestone on the public test net and then you know a few months after that'll hit the main net so that brings in like evss node operators I canba everything together this whole complex Marketplace

uh so we're working uh hard on on getting that ready but that's that's the road map is you know this quarter will get the public D net and q1 you know q1 to Q2 we should mean it fantastic I know that's everyone's least favorite question so I appreciate you tolerating it but uh sham and shave this was an awesome conversation I really appreciate the both of you coming on to to jam on all things Ian layer and like you mentioned earlier we'd love to have you back on when uh governance gets a bit more closer to being live and and we can really jam on the inner workings of of the design so maybe Shri Ram I can leave it to you to kind of leave the audience of where they can find more about igen layer or point to some of the exciting research your team is

building yeah um you know the best place to um discover what's going on please follow our Twitter handle at ianl um the the what we're doing over the last couple of months is starting to write a bunch of like blog posts so you know that's on blog. XYZ which explains for example the most recent one is on the inner workings on I layer Shon team brought up another post on you know 15 AVS ideas for what things you can build on a lay so all these posts are super helpful if you uh have any thoughts or ideas you know you can comment on some of them on the forum. XYZ which is our research discussion forum and we'll

we'll look at it as a team I'll put a f here for Igan ler their research is some of the best out there and they just announced the Igan layer research Fellowship one week ago so if you're really interested in contributing to research at aan may definitely go apply for that it sounds like a really fun job I love it thanks a lot Randon Sham shic this has been great cheers [Music] guys