Inside Eigenlayer's Vision for Decentralized Security | Sreeram Kannan

Transcript

igen layer is designed on the principle of open Innovation which means let the people decide what what Let each AVS decide what they really want to do each staking provides objective slash security igen provides inter subjective social slash security and your own token provides alignment which is a different kind of cryptoeconomic security your your token stakers are the most aligned they are the ones who will feel it the most so these are different kinds of cryptoeconomics this this is like a Swiss army knife of cryptoeconomic toolkit so you decide what aspects you want to load on like your own token what aspects you want to

Outsource so it's designed to be this flexible system whing up with the Wormhole Foundation Steward of the Wormhole protocol the leading interoperability platform powering Crossing transfers for 200 plus multi-chain teams came your unique worm NFD just for zero X research listeners today pair is an Innovative pair trading exchange built on top of Simo users can trade trending narratives with one click from Blue Chip narratives such as long BTC short e right through the whiff versus Bonk the Hat stays on Sim's intense Centric architecture enables deep liquidity source offchain and Broad onchain and combined with pair has democratized access to complex trading strategies typically reserved for institutional Traders Avalanche is a

smart contract platform that has a novel consensus Pro protocol subnet infrastructure and the hyper SDK toolkit which allows web free developers to easily launch custom powerful blockchain Solutions you can start building today by checking out the link Bel before we get into today's episode just a quick disclaimer the views expressed on this podcast by either myself or any guests are their personal views and do not represent the views of any Associated organization nothing in this episode should be construed or relied upon as Financial technical tax legal or other advice all right let's jump into the episode what's up everyone and welcome back to another episode of ZX research today I have a fantastic episode lined up uh we're joined by shiram the founder of IG layer a reaking protocol built on a fury

sham thank you so much for joining us how are you doing very good thank you so much uh really excited to be here and chat with you all today fantastic um today is Thursday May 9th but this episode will go live in two weeks uh on Thursday May 21d as a reminder you can use code0x research 10 for 10% off your blockworks research subscription all right let's get into it um sham do you want to give a short introduction on IG layer and what you guys have been up to over the past month with the main not launch the ABS main not launch as well as the igen token yeah um the uh igen layer is a protocol that allows new decentralized protocols to borrow trust from the ethereum uh

staking so the root of trust in ethereum why do we trust the ethereum blockchain chain is because ethereum is built on you know proof of stake which means you stake your e and then make a promise that you're going to run the ethereum network correctly and that's why we trust that the ethereum blockchain is running correctly and what igen layer does is to extend this like people can stake their e and make a promise not only that they're running the ethereum protocol correctly but also that they're running other decentralized protocols correctly so this creates a new economy where like you know anybody can now just build new software we call them abss actively validated Services just think of them as like any decentralized

protocol now anybody who has an idea for how to build a decentralized protocol they can write it as an ABS and write arbitrary code you know there's no constraints on what kind of code you can write and then deploy it to igen layer and igen layer basically brings the you know stakers which is the economic security the node operators which is the computational resources to actually execute these transactions and then the uh settlement which is you know on ethereum there is you know slashing and and and you know uh proofs that you can actually write to make sure that the nodes are doing their job correctly so that's the kind of IG layer economy so you can think of it as it's bringing together stakers operators and uh

avss to actually create useful productive services for the crypto economy another way of thinking about it is what is the kind of dividing line between what we call crypto or blockchain and everything else is decentralized trust and so decentralized trust is the raw material out of which the crypto economy is manufactured you can think of I layer as a Marketplace for decentralized trust so you can actually kind of decentralized trust as a service and so now like once you have decentralized trust as a service you can start building newer and newer interesting applications on top and so that's the that's the goal of I ler so in terms of the road map and where we are at uh you know last year I think

June July we launched the main net where we uh had one side of the marketplace on Main net which is staking so essentially you can stake via two mechanisms one is you can stake in ethereum and set the withdrawal address to go through it's like a stop through the withdrawal flow go through the Aon contracts so that's a first uh mechanism we call native restate the other one is called liquid reaking you can take an LST and then deposit it this these two features were live on mainnet you know now for nine months and you know we took a gradual roll out approach this is our consistent approach we adopt across you know all our deployments the gradual roll out is

you start with a capped amount of deposits and then like slowly raise it so that we can stress test the system all through you know smart contract systems are always risky and because it's immutable uh and you know that's true for I layer also but the way to slowly stress test it and increase the kind of range was is how we adopted it that's the same philosophy we take to deploying the igen layer system not only the T the the deposits on aanl so the next phase of the roll out was uh you know in April where we brought not only stakers now node operators on mainnet avss anybody can build and Deploy avss

on mainnet our own particular AVS called IG da event on mainnet so igda is a data availability system which is a data publishing system for rollups so if you're an ethereum rollup it costs a lot to publish data on ethereum instead you publish data on end da so I da went live and the whole ecosystem kind of came together two major features Miss Missing in the main net is payments you know abss cannot do payments yet to operators in protocol of course you know there can be out of protocol payments which we see some emerging the next one is slashing the co core contracts for slashing we're

already there from last June because you know you can the core of slashing is very simple burn your funds right so it's not like it's difficult to write a contract that just burns your fund the hard part part is actually making sure that the system and like when when an AVS deploys on top of it how to make sure that they are actually the the slashing conditions are solid how do you make sure that there is a you know process for vetoing you know does every slashing needs to go through the veto there's a lot of complex details that matter a lot and so we just taking again a cautious you know roll out approach so that that's what's on Main net one of the reasons payments not

yet on mean net is we want to understand the market structure better as to what is the balance of powers because initialization on some of these marketplaces matter a lot you know if one side has too much Market power they'll basically command so we're observing before we deploy so that's that's the approach that we're taking oh I and you know of all the things I forgot to mention as we were building a several of these the this kind of a mechanism for you know avss to B b decentralized trust one aspect that we found was missing is the ability to slash for things that cannot be proven onchain and so that led us you know particularly uh to to develop a

mechanism to do it and that's give rise to this you know concept of the igen token which we announced in a major white paper uh you know a week or so back so that's the the other dimension to the launch and the going forward you know several of these features are coming live payments coming live slashings coming live again you know gradual roll out of all of these features perfect uh a lot to unpack there but I I kind of want to start off with the igen token I feel like that's the most recent and um I guess like the most like in my opinion one of the most exciting parts of it over the past month I kind of want to like diving into the white paper I think the first thing that

most people notice is inter subjectivity and I kind of want to get your take on like how you would describe this to somebody who is like just reading it for the first time what do you want them to take away yeah um I know it's uh it we we kind the term like pretty seriously but it's amazing to see it's a meme you know it's in crypto we memes matter in fact inter subjectivity itself is basically just another way of saying what we all believe in so let me start from the Basics I think you know back in the day like maybe five six seven years back if you think about like

what people were talking about what is a blockchain and some people would say a blockchain is a truth machine I just heard this thing and I'm like how does blockchain know anything about truth even even I don't know what is true like how does a blockchain know what is true and that was a weird kind of a phrasing and so that caught our attention quite a bit bit so I'm like what does what is really what does a blockchain really do and you know over time I've kind of created our own mental models for what a blockchain is and one one way to think about it is it's a device for commitments right I I make a commitment that I'm sending this money to you which means that commitment is automatically enforced so you could you

know this the reason I mentioned the truth machine thing is it's if something like that exist it's really amazing that you know the there is a truth machine that can serve as a kind of like a center point for society but then you start thinking about what kinds of Truth exist and you you know I would say roughly you can say that there is objective truth which is mathematical truth right like so 1 plus one is two or whatever there's mathematical truth scientific laws are a form of objective truth okay the law of gravity relativity whatever but so what blockchains have access to today is fundamentally just mathematical

truths right so I you know I have a private key relative to a public key and therefore I can send my Bitcoin that's a mathematical kind of a statement that I have the public private key related to the public key so I have the rights to move my Bitcoin around um but if you look at uh a lot of you know Society it's built on social truths you know a social truth might be uh is gold worth anything you know who is Rich a social truth might be who's the president of the United States um who's the founder of ananl these are all social truths and they are truths because we all agree that it's true and so if you look at

kind of like the structure of coordination in society you would see that a lot of coordination is built out of social truths and a a for describing what is social truth is inter subjective you know what is objective is facts right subjective is beliefs that a person holds inter subjective is things we all believe in we all believe that ex you know Biden's the president of the United States and you know somebody's the CEO of some company all of these kinds of statements we all believe in and that's the right kind of philosophy of what how do we bring if you want to BU build a truth machine then you need

to know what is the the basis for what you're calling truth and Inter subjectivity or social truth is what a lot of you know when you think of an article problem like how do you know the bring the BTC to USD price on chain it's basically into subjective truth that the price was at that level at that time you want to settle a prediction Market it's inter subjective truth that um you know the Biden won the election or not not won the election the uh you want to know uh you know get taking closer to home example like we were building igen da a data availability system one of the big faults of a data availability system is there's two functions of a data avity

system is to store the data and then to serve the data okay it's possible to create a zero knowledge proof that you're storing the data is because you know you know the data and you know your private key and you create a zero knowledge proof for it that basically means you can actually prove that you're storing the data but did you serve the data to all honest people who are interested in it that's not observable on chain it's not provable it's not a mathematical statement because you know I may say I served it my network was bad or whatever like you know it's not observable it's not rigid it's not solid but it is inter subjectively observable we all sit outside and we see the system and say hey I tried querying this RPC and never

found this roll of data I went and ran a full node and tried to connect to the peer-to-peer Network for a week and I still don't have the data I tried to run a light node and I tried to sample portions of the data and I tried to download it I still don't have the data and all of this tells not only me but everybody who's outside can observe and say oh the the I nodes all colluded and withheld data so there is a data availability fault so this is an example of an inter subjective truth so igen token was created as a framework for bringing inter subjective truth on chain or to punish validators for inter subjective

false false that we will all agree from outside the reference frame of the blockchain that such a fault happened but inside the reference frame of the ethereum blockchain you you the blockchain doesn't know so that's the high level idea of the starting from if block chains were to be truth machines what do you need you need basically a Locus or first a definition of Truth objective and social truth and then once you have social truth how do you bring social truth on chain and so that basically sets up the problem of we want to either punish validators for things that are socially attributable but not objectively attributable so that's the scope of what what we want to do it is

you know super fun that inter subjective is a kind of like a memeable term and a lot of people meme about it I'm super happy about it wormo is the leading interoperability platform for multi-chain Bridges and applications with over 1 billion messages sent and 40 billion in value transferred woro leads the industry in cross-chain message volume and total value transferred if you're a builder wormo has an extensive sued of tools and infrastructure to securely build multi-chain applications you can use Wormhole to power D5 nfts governance and so much more and wormhost supports over 30 blockchains and six different run times unique Wormhole NF just for zerx research listeners today fill out the form prove that you've listened to today's episode to receive a PO app and

you get an email over the coming weeks to claim your limited edition nft pair is an Innovative pair trading exchange built on top of Simo users can trade trending narratives with one click from Blue Chip narratives such as long BTC short e right through the whiff versus bong the Hat stays on Sim's intense Centric architecture enables deep liquidity source offchain and Broad on chain and combin rep pair has democratized access to complex trading strategies typically reserved for institutional Traders the age of interoperability is here with the launch of teleporter Avalanche becomes a seamlessly linked network of many sovereign blockchains that communicate so smoothly they feel like a single chain teleporter is an evm compatible

cross-chain communication protocol built on top of avalanche for me it allows any evm chain on Avalanche to send messages assets and data to one another in a simple fast afford able and secure way teleporter and Avalanche rep messaging provide a new foundation for how Avalanche blockchains can communicate and how developers can easily build cross chain applications out of the box build anything you want any way you want on the eco-friendly blockchain designed for web freep check it out by clicking the link below I definitely love to dive in specifically how you like use inter subjectivity within the igen system so I'm curious I guess to start off what is the actual like mechanism to

decide or to actually like slash within this system um the the next point so now that we kind of like set up the stage that okay I want to bring social truth on chain and how how do I do that um particularly in the IG L ecosystem we have a narrower scope the narrower scope is we want to be enable arbitrary decentralized protocols to be built like that's our that's our goal and to supply decentralized trust for protocols that are um validated by uh that don't need to bootstrap their own trust Network so

that's the kind of like a core core requirement so we start from the broader concept of inter subject inter subjectivity or social truth but specifically what we want is hey some group of nodes are are promising to run a certain kind of validation and they fail in that particular task of doing the validation correctly and then you want to hel hold them to account that is the core narrow uh problem that igen layer wants to solve why does it want to solve it because we have this economy of actively validated services that are being built and you cannot slash e for things that are not provable on chain you can by

putting a committee and saying the committee is going to decide then you don't really have like higher trust than the committee itself and of course some people are going to do it and but we don't it's not a rigid enough principle to build a long-term sustainable economy on top of so how do we slash validators for FS that are attributable outside but not socially but not on chain that's the specific scope that's why we call it an inter subjective work token or you can think of it as a inter subjective staking system right so we want to just so this is in indifference from things like you

know claros or like an Aragon Court where you have jurors and people are actually sitting down and debating and deliberating and coming up with what is the right answer that's not the scope of I igen token at all scope of igen token is nodes are performing automated validation in the normal mode no humans exist right it's just nodes are performing automated validation you may be running an oracle you may be running a DA you may be running an AI you may be running a gaming system something it's all automated and just like running on its own but they all collude and like try to actually attack the system and then when the attack happens they should be slashed that's the like you know narrower scope so that's why we call it

an IND subjective for token you have have to stake the token for doing automated work like digital validation and when when you're doing that you we want to make sure that if you're not doing it right you you can be slash okay how do we how do we slash people for this and uh the the principle goes back to how does ethereum slash when a majority of nodes are malicious majority of stakers are malicious okay in so one of the things I I was really like surprised to learn as an academic coming in in in the academic world a lot of the models for how blockchains work are all based on if a

majority of nodes are honest then the protocol works and the remaining nodes can do whatever they want but the majority are honest then it works that's the kind of like you know models that academics you know I used to study you know three years back but as I entered into the ethereum ecosystem one of the things I realized is actually the the the people building these these blockchains are specifically in the ethereum community worry about even if the majority of the nodes are malicious the system should still survive and go forward which is a crazy kind of like an aspirational goal and that's the same goal we have so you ask you know how does ethereum survive a majority fault

and the answer is Fork the chain like that's the answer if the majority of the it is go Rogue as the community as the society that is observing from the outside from outside the ethereum blockchain we can all see that these guys have behaved maliciously there should be a observability for example the majority of nodes sign an invalid block then it is observable that hey the block is invalid or the majority of nodes sign a block header but the data is not available for ethereum it's observable from the outside so any observable fault automatically triggers all the outside notes to come in and say hey this is actually wrong I'm going to Fork the chain okay so forking is a mechanism to

create not only a penalty for the majority nodes but also to create this you know bring social truth on chain because you f the chain okay so that was the basic starting point that's how you become that's a principle right like so you start with the social truth but then how do you bring the social truth to bear and on on a chain and the principle is forking right you Fork the chain that's how you do it now you come and you know vitalic wrote an article back in June last year uh and he said do not overload ethereum consensus and you know it's a very important article and you know it's sets the shelling point for

what we as the community actually you know agree or disagree so it's a very difficult thing when you have social consensus to agree on what we will Fork the chain for and what we will not Fork the chain for and so vitalic is kind of setting a shelling point for hey ethereum is not going to Fork if Ian l or somebody else screws up that's what he's saying and it perfectly makes sense because a system with undefined social consensus basically leads to chaos a system with predefined social consensus and self verifiable self- observe for example in ethereum everybody knows if the ethereum block is invalid we will Fork the chain you don't need to tell anybody it is a pre-agreed rule that we self- execute so this is another

important thing that I want to point out people think oh igen token is used for social consensus we don't mean social consensus as in we all meet in a room or meet in Discord and then we make a decision that's not what what we mean we mean that there is a pre-agreed rule for how forking will happen that you can self observe and self- execute now you know you have to Fork and so this is how ethereum works now coming back to how do we bring this the problem is if we're building ion layer on top of ethereum and you know something somebody's slashing based on a committee and the committee goes Rogue or they lose the keys or something happens and they just do this Mass slashing then now ethereum has to for that's a disaster so we don't

want that we want to containerize a certain portion of slashing in the ion uh ecosystem and so what we did is we figured out how to create a system where forking does doesn't happen at the level of the blockchain but forking happens at the level of the token it's a brand new kind of idea it the roots of this idea exist from before and I can talk a bit about it but the idea is you know I don't need to drag the ethereum chain into my like it's like my family dispute and I'm dragging like the United States to it you know it's not fair not reasonable so can we actually build a containerized uh system the token can Fork like it's a forking at the

application layer rather than forking at the chain layer so how does this work um the igen token is what we call a forkable token just think of like okay it's why is this particular erc20 contract the igen token contract because we all agree it's inter subjective truth it is social truth that that contract is the IG in fact you can go to like you know uh ether scan and you'll find like scam tokens called IG already deployed right somebody's deployed them and they think that's but nobody believes that that's the igen token that's why that is not the igen token which is the igen erc20 token is a social truth it is not a objective truth so okay so that's the starting point so

now you can say oh maybe all you need to do is yeah there's the erc20 token and if some people go crazy we just like Fork the Aon erc20 contract that would be the simplest version of this and and really 's just a system built on top of this core idea that all you need is all value is the igen token is that the erc20 contract of wagon token is a social truth can we if the majority of the nodes which is the problem that we're worried about is majority of igen stakers go and attack an oracle because you need a majority to misreport in order to get a wrong answer need a majority to withhold the data in order to get a wrong answer you need a

majority to do some of so each of these avss you need like a lot of lot of the I stakers to actually behave badly in order to create a malicious outcome and maybe then it's worth it to actually Fork the token okay so that's the kind of like the next level of the idea is you have all values into subjective let's just Fork the token the problem with why you can't do this for any erc20 token today is that token has a non-containerized life which means somebody is Trading that token in an exchange somebody is using that token in a DEX and now you say for the token then like it's chaos utter unrestricted chaos because you know somebody's selling the

token but you say no that's not the token and this is the token what happens right so that is a major problem so you can't just go vly Fork a random erc20 token it doesn't work because that token is the surface area of the token is very wide and so now like you can't go and Fork the token if you Fork the token then you have you know uncontrolled Downstream impact okay so now and that's really what the igen systems designed to solve is we call this property isolation isolation is staking and forking should be isolated from holding a token from you know having a token in defi having a token in all these other places if I can

create a system where you know a version of the token is forkable but another version is not forkable then you kind of solve this problem so that's the high level idea then there are other problems which is forking creates a social cost okay so suppose I have a token and then like WTI comes and says oh this is the real token and then now like all of us have to be create chaos and confusion and everybody has to sit down and debate whether that's the right thing or this is the right thing so soci forking has a social cost so we need to meter the social cost and make sure that you know there is a there is a system of reward and penalties for creating a fork if you create a Right

Fork you should get rewards if you get a wrong create a wrong Fork you should get penalized otherwise we are wasting everybody's time here so metering is a second problem the third problem is I call universality universality is you need this to work across a range of applications because igen layer is a open Innovation System we're not building the igen token for IG da building the token so anybody can come and tomorrow you have a better Oracle you have a better AI in front somebody else has some other thing they should be able to leverage this to stake and support their security so universality and universality has multiple Dimensions one is um one dimension is the pre-agreement on the token should be

that it is universal like that it will fork for all these avss so I'll give you an example there is a this kind of a forking idea was already there in a prediction Market get called augur and so that's why you know a lot of the idea like flow is you know this was imp this was there in AER and prior to augur it was already there in some of vital Expos and I'm told vital Expos has some precedent from this thing called Truth coin we haven't traced that lineage down carefully Beyond vitalic yet but we'll do it the uh the the idea is that augur was a token which was used for it's a forcable token but used in the context of a prediction market and because it's

used in the context of a prediction Market uh augur will not fork for data availability or some of go and stake the augur token and something bad happen somewhere else it is not going to Fork the pre-agreement is not there that it will fork for all these other things that is the igen tokens like designed to be Universal so that's one part there's another part to universality which is for example when augur was built augur had an internal cryptoeconomic calculation what it would do is it would maintain how many bets are open on all the prediction markets and say oh you know the total amount of bets open on all the prediction Market is you know 30 million so I need to make sure that more than 30 million is slash in the auger

token so it has an internal metering of the the profit from corruption an adversary how much they can extract profit if they go Rogue and what is theost cost of corruption how much can they slash the adversary this calculation is intrinsic and maintained inside the auger system and because it's application specific it's easy to do it because you know oh that's the total amount of bets it's so the in in terms of inter subjective tokens you can think of uh augur was a application specific inter subjective token and Ian is a universal inter subjective just like Bitcoin was an application specific

you know uh blockchain and ethereum was a universal blockchain this is the same kind of mapping um but the problem to solve is how do you know like if igen is uh used across 300 abss how do you know the profit from corruption how do you know the cost of corruption how do you maintain the system in cryptoeconomic equilibrium make sure that it's always secure so that's the hard problem right it's not to just say here's a can fork for anything that's easy to say it's just what you say but the maintenance of that condition is hard and so how we do that is a whole thing that's a third problem so I talked about you know the

cost of social consensus the uh uh isolation between D and and uh you know forking and then the um universality and then the last one is what I called compensation so compensation is basically the idea that you know if I'm using Ian da and I got wronged because you know validators basically colluded and made data unavailable my rollup gets rugged so I'm I'm suffering the consequence of the bad action of these igen stickers now I should be given some compensation for this like bad action of

these sckers till now before igon layer no cryptoeconomic system had a system of slashing and redistribution this was not there and we introduced this in a kind of a major paper called stake sure for just like objective Falls what you do is you you stake and instead of just slashing and burning the funds you slash and redistribute the funds according to certain you know calculations and that is the fourth property that we built into Aon so I've just given a property tour of what what set of properties we have there a whole thing to go into like how we get these properties no that was a really comprehensive answer I actually had a

question later on about like Stak sure and with respect to like attributable security because I did read that paper that you released back in January it was like really thorough and I think identifying the the cost of corruption versus the C the cost of uh I'm sorry the profit of corruption and I think it was my understanding my conclusion and please correct me if I'm wrong before I I guess dive into next question question was it sounds like the cost of of security should be related to effectively like Bridge volume of a given abs and not necessarily like the tvl because you looked at I think you used like ethereum as an example there's X dollars of tvl in ethereum but there's only I think like 33 billion dollars of Stak uh stake to the network which at that time in January I don't know

exactly what the number is today but um it was it was a really interesting piece overall so and I know a couple other shared Security Solutions in the market are actually like using that as a point of reference for how they plan to to I think create like a similar reaking uh mechanism to igen layer um one question I do have like around the universality of igen um the igen token for in inner subjective slashing and then I think drawing parallels to Auger where the being application specific in this world of a hundred of a thousand abss leveraging I layer there's obviously a cost of a forking there a social cost of a forking how how does the system how is the system be able to challenge

every inner subjective Fault in in a world of thousands thousands of ABS obviously which is the I think hopefully the end goal of I layers is where we see uh you know this level of adoption because I I understand having uh taking off the uh shouldering the responsibility of the of not overloading social consens on ethereum but I'm concerned is that now itself is overloaded it's great question okay so the scope of the current white paper is when what what I would call perfect reaking or you could say everybody staked into everything okay so imagine

all the IG stakers are staked into all of these different applications okay which which is in some sense the most overloading and the thing that people are most should be most worried about but actually I think it's the best case um okay so what what happens in this scenario the so firstly when we talking about slashing we are we talking about a majority of so we only talking about abss where a majority have to do something bad in order to get to a fault not like one guy did something bad and then like you to get to a fault so that's not the scope of the type of Slash ability we are talking about and the way you would handle one guy going

bad is by saying that if one guy goes bad the other guys can then like come in and double check so you can have a two-tier escalation model and know the other nodes which are opted into that particular ABS can double check it and like only if they're all vent bad then something bad should happen so that's the starting point first everybody's taken to everys that's the scope of the current white paper we have more coming up to expand it but already this is a mouthful so let's let's keep this are a brain full uh the next one is suppose everybody's taken into every AVS and the avss are such that if a fault happens a majority of nodes or like you know one3 or some significant fraction

are attributable that they are wrong so that's the starting point and that already like makes this a very extreme very narrow like rare event just to take the history of all blockchains this has happened zero times right and the so then you be like why why build things that happen zero times is because these things exist it will not happen right because the system is designed carefully that if you attack it you're not going to profit you're going to get screwed you will never attack it it's kind of the you know mutually assured destruction not mutual but assured destruction to the attacker basically means that you know you never do it so

that's the principle so firstly slashing only when um so so everybody sticked into everything second a majority of to be bad to create a fault and therefore a pro even a challenge that this should be forked should be very very rare it's not a normal thing so it's not a you know we say it's a for cable token but it's not a for King token in the sense that you don't expect it to be forking all the time it's not like every dispute in a fork nothing it's a very very rare to fork and how do we impose it okay so you can say yeah you know this is what we expect to happen how do we ensure that forking is rare and so here is a couple of parameters

that are built into the protocol called cpf and DPF cpf is called commitment per Fork commitment per Fork is think of it as if I want to create a fork I need to burn a bunch of ien tokens in the current Fork why am I okay with burning a bunch of igon in the current Fork because I I will of course issue myself those igon in the in the new fork and I believe the new Fork so I can kind of burn the I in this Fork it's really serious right so for you to actually burn a bunch of like tokens to say that the other one is the right one and you know what is this parameter is it like 2% is it 5% is it 1% this is needs social like you know

setting up of this parameter it has not been set up but some par meter exists like that cpf commitment per Fork which is basically saying that if you don't B burn enough amount of igen you will not be allowed to create a fork so forking automatically becomes serious it's not like uh oh I just you know callously say oh here is a fork here is a meme and therefore I'm going to Fork you can't do it okay so that's number one number two on the other Fork like cpf says that I'm going to burn on the original Fork DPF is called deflation per Fork which basically says in the new Fork the total Supply should be lesser by

DPF and why should it be lesser because you know you found an attacker you're going to burn their token at least some percent of their token that means the total Supply should be lesser what this does is if forking worked correctly which means either this Fork took off or that fork took off then in both the forks the total supply of I reduced so and you have compensated for the social cost of all the people who had to worry about is this Fork right or is that fork right you know there's the social cost incurred and you're paying off the social debt by extracting the value either by the wrong from the wrong attacker or or the wrong Challenger or from the malicious takers you're extracting the value and distributing it

to the to the community so that's how metering basically this is our solution for metering ing the cost of social consensus that's how it works this is why we think if the systems well designed and like properly set up focing should never happen just like 0% of the in the history of all blockchains that has been a majority malicious attack which which lead to majority slashing there have been isolated incidents where you slash one guy we not talking about that that could be handled inside the AVS the AVS has to have a majority fault before it gets triggered to the igen social consensus so that's how it's handled so you b so this is you know a principle in

computer science called amplification is basically you say a fault occurs in a small group you amplify that fault to a majority of nodes and then now it's enough and it's serious enough that you know everybody can consider that whether that is a fault or not so that that's how you know the igen system works so it's really rare to to have to be in a scenario where you need social consensus when it is there it's you know you can exert it and actually create the for imagine a majority of Igan da stakers actually go Rogue and do something crazy then you need a last resort to actually slash that and this creates the last result the SL um sure I'm kind of curious about ABS

is how do you see the um supply and demand match over there like it's $15 billion too much insecurity for what's needed how do you see the market playing out yeah um the um the core bet of igen layer is that once you make decentralized trust easily accessible lot of people will want to use it that's a COR bet it's kind of like when you try to size the uh Market in many times when new technology comes up the size of the

market expands because of the ease by which something is possible you know so that's the core bet how long will it take for that to manifest you know it's not this year right it's not this year it's not next year it is going to take time for actually you know thousands of avss to be built and those AVS to have productive yield absolutely um the but I think people are thinking about the yield question a little bit uh

erroneously the idea is if you're not so imagine this $15 billion State and there's really only demand for $5 billion of like State you know um of security as long as there is uh you know liquid staking or liquid reaking tokens which basically anywhere other others otherwise exist right imagine Lio exist and there is an LSD and what is the opportunity cost of taking on Lio you you get St versus you know you you can hold e as long as powerful liquid reaking

protocols exist on top of ion layer the opportunity cost of actually staking through ion layer can be very small so what's happening here is basically there is a trust flow under trust assumption so that's that's how like you think about ion lay is staking and Val you know having proper node operators run the software is the trust assumption that ABS is taking but and and the way you're getting the trust Assurance from Ian ler is that hey these guys can be slashed if that goes wrong but for somebody who can actually assume that stakers and operators are going or or operators are going to do the right thing there is no additional you know

real risk of course there's smart contract risk but that also exists in a liquid staking protocol right so the there is a world in which even if the relative yield off like I layer is small still worth it because it's a little bit more for and you still have liquidity and you still have like you can go and do all the other things so okay so that's just to set some you know comparative expectation because you know you have something like Li already there you get an you stake and then you get an st and what is the fundamental risk of holding s is that the operators don't do something malicious that is the fun Al

risk of holding s and there's something similar that you can build on on IG layer of course you're adding the fact that the AVS slashing contracts are correct and some few other things but the core I layer contracts is similar to the core let's say l of contracts and so you could do that okay so that's the first thing to normalize you know even if a lot more stake is sitting on igen layer as long as there is the the liquidity premium is you don't have to pay because you have liquid state tokens that actually have you know uh can be used elsewhere that's number one number two what is the rewards that people are getting you know we think about real

real fees actually being paid to the abss and then come back I think that's still going to be a bit further in the future what we do see is avss have token emissions that they us using for already for security they'll be paying a portion of it to you know whoever is securing it usually you know you look at the total Market of you know in in Cosmos in like every chain how much is given for inflation a portion of it can be allocated to you know each stakers in I so that's a second like you know thing that we starting to see will happen and when we're designing you know for example I mentioned the payment router

and the payment system we're observing and understanding what is the market Dynamic here that actually will play out so that we build a payment system that's actually compatible with it so that's the second one so the because a lot of the the real yield which is actually people paying in E and that satisfying the E stakers is further into the future it is not this year right so that's the next thing and the this is different from oh you know because people are using this system they're going to addir drop or some other thing like that's not what I'm talking about I'm talking about

paying a fixed fraction of their inflation to the stakers that are actually securing their system there is a clear Synergy and a win-win deal there it's not a out of Goodwill you have to do something okay so that's the other other part and the third one is I think when people people think about this you know I saw people riding like thread and so on about uh how much security is enough for a given AVS and in a pool security model for everybody like oh 1 billion is good enough I think this is I think maybe other people built reaking this is how it would have looked we do not you know we are not building a

system where everybody thinks they have the same security and that security is completely leveraged that's not the system system we building we have a in igen we have a system called attributable security which you know we we just talked about uh the idea of attributable security is that you can slash and redistribute a portion of the stti to the you know to the different abss and when you have attributable security this is non-leveraged so the total amount of attributable security so IG layer maintains a calculation called proof of solvency and Pro Pro of solvency basically means even if you know each each AVS has a certain amount of attributable security from the Aon L

protocol even if all of them trigger slashing simultaneously the the total all the attributable security can be honored simultaneously that is the proof of solvency so the way we don't do this by decoupling the pool we don't say oh you know uh in in avs1 I have $1 million state in AVS 2 I have $100 million St no that's not what we're saying saying the stakers are staking into potentially everything to into whatever they want and potentially everything but the protocol the igen layer protocol takes up the role of how to allocate and these attributions across these things collect the fee of all of them and then

redistribute to all the stakers this kind of like a central cryptoeconomic engine is running through a so once you go to the attributable security world it's clear that avss a base built as an AVS would want much more attributable security than a kind of like a random AVS and so you want to make sure that there is this expressivity in the protocol so that people are paying for actually something that they're getting from the protocol this kind of like attribution and attributable security is built is coming up not only for each staking in igen L but also for IG staking so I staking when you know you go back to the

question of whe there are hundreds of avss trying to use the same stake how do we know that the system is in cryptoeconomic equilibrium because each AVS can specify how much attributable portion that they have that they can slash and redistribute so that's how the kind of igen layer system works so the three points are number one is minimize opportunity cost number two is you know before real yield comes in the the avss use a portion of their token to actually fund security for themselves the third one is actually create an incentive for people to to have a portion of attribution to them instead of being

smudged into a big pool where they don't know what they're getting and therefore not going to pay for it that makes all that makes a lot of sense I've I've personally been pretty involved in like the cosmos ecosystem for a while and like saw a launch of replicated security so all the issues with replicated security obviously there's like a lot of lessons learned and best practices from I think like polka dots like crowd loan and and Par slot par auction slot mechanism um I think one of the questions I have is like around the payment structure specifically one of the concerns that I've had and this is not IG layer specific this is shared security as a whole because you kind of hit the nail on the head like a lot of these abss or consumer chains whatever you want to call them it takes years for them to hit

sustainability where they can really function uh on their own stand up their own their own validator set um and I think that's been a lot of the points of contention I think in the cosmos ecosystem around the cosmos hubs like consumer chains um is there like a a general one of the concerns I think I have is like this idea of security mining so in defi you saw liquidity mining you saw people farming you know uh emissions for defi protocols I think one of the concerns I have and you're going to see I think in in other shared security models as well is like depositors are going to farm emissions to secure the chain and then the moment that those security those emissions dry up they you know pull out their their assets and they maybe move on and and attribute their their security to

another ABS um is there like a a mental framework or is is there a framework that either like the liquid reaking protocols themselves are taking on or maybe I layer igen Foundation is taking on to try to like create a model for like what does good a good emissions budget or security budget look like as an ABS maybe this is application specific to ABS but we love to for you to like kind of opine or further yeah I so the way we saw okay so there's two parts to this how does how do you set the security uh budget from an AVS point of you what's the right inflation how much do you pay very very hard kind of question because you know when you set

this parameter we could for example set a parameter and say you have to pay x% right of your token that's what is you have to do if this is misset then you either miss a large portion of the market they just go away or it's it's it's too low and you know like you know the real fees model that the replicated security was running on where it's too low it's not enough to keep the system running so how do you have enough expressivity in the market that basically you are making sure that all the par Market participants are actually you know meeting their demand otherwise you know this Market will unravel and it is a hard hard hard question okay so know not we we think

about it a lot it is you know this is not the kind of thing where uh you you can simply like many many like systems which look like they're taking off will unravel if the market structure is not correct so it's it's very important to get it right but going to the the the thing of what we are doing despite like how to make it robust right one one thing is draw a lot of pools always have a lot of pool security but have a portion of attributable security so this is the kind of first principle and the principle is that hey if my rewards dry up I I should still get a lot of pool

security what do I mean by pool security to the cost to attack my chain is still high I want to maintain that but I will not be able to slash and redistribute a lot it's okay I mean like that's all that you can offer that's all that you are paying for that's all that you'll get so this differential quality of service which is that if you pay more you get more attribution but a lot of people get a lot of pooled security so that is one very important concept but one thing when we're talking about pool security is right now we are only calculating the cost of capital there also a cost of operations so we have this framework you know again like everything's gradual and you know in faces in Aon ler but we have

a framework for computation payments as separate from Security payments in fact this I think is the biggest error in all of crypto today is these two are smudged into one thing like in ethereum you're paying like you know a transaction fee which is gas which seems like it's for computation but really computation is much cheaper in ablers why are you paying for gas on on ethereum because you're getting security security and like computation are smudged in to like one cost they should not be they should be different so I ler will have computation payments which basically keeps the computation and The Operators are float because that's the marginal cost of pool

security is I have a pool I want to just add on more things I need to keep make sure that the operators can run at a healthy profit margin that's the computation payments and things that have more computation should have more computation payments things that have less computation should have less computation payments so computation payments as a separate phenomenon from you know uh Security payments and Security payments for pool security has a different phenomenon from uh attributable Security payments so basically like create this graded structure of what are you paying for what are you getting how do you keep that market participant healthy and active so I think yeah no that that makes sense

it's definitely not an easy qu easy problem to solve no nobody's really come up with it but I I agree generally like there's two payment structures it's paying the note operators and then paying like the security which is paying the um the delegation the the the assets that are backing the the node operations uh no node operators one question one followup question to this and obviously wesy and Bachi would love for your time and after um what are your thoughts on like validator bonds specifically uh because one thing that I think I've seen like I think eth research write a little piece about as well is at like in dpos systems an notor operators are obviously delegated their

their stake uh overwheling majority but uh at what point is like they don't necessarily have skin into the game themselves because they maybe like I I know some of the top operators in Cosmos ecosystem for example 99.9% of their stake is delegated to them and they have no they don't own their the own native asset of the proof of stake Network do you do you envision and I know this is very hard to have like onchain rules for like offchain actors because you can especially like with centralized validator centralized exchange validators that control their delegated stake they could always put up their own customers deposits as like a validator bond but do do Envision a system around Igan L where there is like the more stake you get through through Igan lay or deposits um the more Bond

you have to put up up as your own on your balance sheet no we don't have that at all uh somebody could could build an LST or LRT on top of I ler which actually enforces that kind of like rocket pool right so rocket pool has that kind of a mechanism but we don't have it we don't enforce it because why is this uh imagine you're black rock and you want to make a deal with base to actually run you know your node operations correctly there is a massive amount of not only trust but also legal compliance and other things which makes the operator behave correctly so by enforcing anything in protocol that now

coinbase has to put up like you know $1 billion for every $10 billion from Black Rock we are enforcing a market structure that is not necessary for that participant so we want to create like a universal mechanism so we don't want to enshrine a particular opinion that this is how trust is formed between the Staker and the operator but we want to encourage many many mechanisms of trust to be formed between the Staker and the operator so some LRT would be like rocket pool and they'll do this we also have other mechanisms for enforcing trust between Staker and operator some operators already we we have like uh lrts like this which are run your uh

run your node operations inside a trusted execution environment which is a hardware protection so unless Intel's crooked or whatever like I would basically be sure this is the code that the uh the node operator is running and you don't have to run the entire code inside that like Hardware module you only run the critical code like signature issu on and you don't double sign like that is happening inside that module for example so so we call these anti slashers right so you're basically like making sure that you can't get slashed because the the thing that trigger slashing is double signing and the double signing code is not run like you know it's run inside the trusted execution environment so even if the

operator goes through unless they break the te they cannot actually you know break the get your slash so this is a different trust model right between Staker and operator and solving the principal agent problem and in general putting this pressure onto the Staker that hey if you don't delegate correctly you will lose your e is actually a really interesting and important thing it's one of the things we're thinking a lot about is how to make sure they feel this pain so that they actually do this delegation correctly if you make slashing really really really rare then they'll never feel this pain and there'll be just like one major you

know catastrophic event where everything gets slashed so you have to make them feel the pain in in the shortterm so I don't want to give up too much of our new slashing design but that is the the these are important issues that we have to consider to actually build the system correctly awesome I I definitely want to get your thoughts on so we talked about like basically this two token system of eth and igen what we're going to start to see is like almost third tokens in the system which is the abss themselves like obviously as an ABS you're going to want to provide the most value to your individual token so I'm wondering your thoughts specifically on how like a

third token an ABS token potentially fits into this road map what are some things abss need to be considerate of of like yes you want to provide value to your token but you also need to provide the most security to your users Etc so like what are the things you're thinking about of like introducing a third token yeah uh IG layer is designed on the principle of open Innovation which means let the people decide what what Let each ABS decide what they really want to do how much of Shar security do you want how much of personal security do you want but I want to just say that these three different quorums provide very different kind of security each staking

provides objective slash security igen provides inter subjective social slash security and you token provides alignment which is a different kind of cryptoeconomic security your your token stakers are the most aligned you know they they take exposure in your ups and downs so if there was something which was neither like inter subjectively attributable nor objectively attributable but it's going to lead to downfall or degradation in your own value they're the ones who will feel it the most so these are different kinds of cryptoeconomics this is like a Swiss army knife of cryptoeconomic tool kit so you decide what aspects you want to load on like your own token what

aspects you want to Outsource so it's designed to be this flexible system um so a lot of people say to to me that oh you know over time maybe all the avss they'll rely only on their own token all of this it's possible if that's that happens and that's what leads to more Innovation it's amazing that we helped bootstrap lots of new avss but we do think there is a lot of value of shat security that people don't fully understand for example when you have five Services da Oracle t storage ji whatever that you're integrating to actually create an end user application if the same stake is taking to all these

applications I only need to take 1X the attributable security instead of 5x the attrib imagine there was a separate insurance for an oracle separate insurance for a DA separate insurance I have to pay 5x the premium to actually get you know I need to make sure the Oracle is secure 400 million I have to make sure the da is secure 400 million I have to make sure the each of these things are secure instead what I ler does is offer an economy of scale so that the same stake secures all these five things you only take need to take one ex insurance and that protects you against all falls of all these different people that's a really really really crazy benefit so over time the economy

of scale out ways so just kind of sense of numbers and how we think about it a typical Cloud application uses 15 SAS services in the back end this is our vision for where crypto will will head a typical crypto application will use 15 abss in the background and like you just bundle all of these so igen layer allows you to create postmarket bundling post Market bundling so you I build I'm the best guy to build an AI inference you are the best guy to build an somebody else is the best guy to build a threshold encrypted meole they build all these things and I just go and subscribe to these services and I I have an economy of scale as though they were built pre-market somebody decided like

Tara bundled the Oracle and the chain right and you know they bundled the Oracle into the chain and this needed like somebody to think that these two had to be bundled igon gives you postmarket bundling open Innovation and shat security and and there's a superpower of economy of skill there and if if it's worth it people will stay if it's not worth it they'll go to their own token and it's exactly how it should be perfect um sure I actually have a question on how you guys approach avss or rather like how avss approach you this is something that I've wondered for some time um because obviously the end goal is like hundreds and thousands of

avss what level of support do you guys provide these abss and like how does do you guys actively seek services do they come to you what's the relationship like yeah I I think right now I would say there are three categories of how you know we interact with abss um there are um one is I would call them bundles so pre-existing bundles like alt layer has a bundle called restate rollup it's a bundle for actually running like a uh a rollup with sequencing and validation and finality on top of igen and so they've already built it and so

what they do is they go and get like lots of people to come and use it so this is a bundle ethos is basically built a cosmos SDK on top of igen and so they can basically get consumers to actually come and build it so they they have inside their own framework some kind of atomic composition whatever so they have their own economy but it is so this is this is one one aspect bundle so the bundles can scale a lot because it's just the same box being run by like hundreds or thousands of these different people the second one is you know there are outside teams which are actually building things like sdks to help avss build new things authentic is one of

them we have others coming up but that basically is a market where they're saying hey you know you want to build an AVS I'm going to help you build an AVS so that's that's a second one they have a set of tools they have you know personal interaction and so on and of course we also do it but our bandwidth has been somewhat limited and we're scaling that effort quite a bit the third one uh is what I the third category of AVS is is you know we call these actively validated services not chains and it's for a good reason it's because we think that these services are the minimal Atomic unit of what needs to

be built and then you should be able to concatenate many many services and create like your own you know uh end application so this is a different view than both the what I would call uh um a single chain view of the world or a kind of like a rollup view of the world this is basically saying each application basically can G it's a bunch of like you know decentralized services and that's what an application is and the um in in this world so what we do is we have put up a lot of ideas on how what is the structure of abss we work closely with people who are inventing categories

of abss so whether that is a ritual which is doing an AI interest whether that's a phoenix that's doing fully homomorphic encryption or inco or you know people who are basically like inventing new categories of like what could be structured as avss that's like really hard because you have to come up with a brand new protocol you had to think about it from first principles you had to figure out what is the trust model all of that so because that's the place we our own inputs are most helpful is kind of navigating the landscape of frust Ms so that's the three categories that we have with bundles you know uh custom help and and tooling and then the third one is actually you know working

together to design the protocol perfect and kind of going off that um IG da is obviously um closely related to ig layer are you guys planning on launching any more um yeah like abss or inter subjectivity Services no we are expecting or working with other people to launch more services we we already have way too much to build just to get I can layer going so it's not our intention to necessarily build all the avss on top Our intention is to encourage people to build abss on top and the thesis is open Innovation right and the thesis of open Innovations

basically says that the the amount of insight and Innovation that exists out there is far higher than the amount of insight and Innovation that exist inside any any entity or company and or a core development team and that's our core principle but there are some things that unblock a lot of other things Downstream right like those are the only things that we will try to build is things that require coordination across all these avss things that require like you know a place where all of them will come and coordinate so we want to be the coordination layer for Innovation so other people innovate we just help coordinate perfect um Westy eort you

guys have any other questions for Shir yeah I can uh so in the paper um it mentioned the security Council and it mentioned like this progression from like V1 Ian ler to V2 and then V3 and how eventually in I guess like the end State um Ian ler doesn't have to rely on the trust assumptions of of the security Council um it is my is my understanding correct that the Security Council is likely going to be something as part of the I layer system for forever but V3 but in V3 like it doesn't necessarily require the security Council to be around or okay so there is a difference between the igon staking and

E staking okay so in each staking because all falls are subject you know or are objective and onchain what's going to happen is each AVS will initially be on the security console and once it's aifi it can go you know without a security console but that has to be something like that for each staking because each staking has onchain slash Falls and only some set of them can be so like well tested that you can just go without like a you know slashing VTO on the other side with the um with igen because the inter subjective like forking is much more

powerful than a committee right because you you don't you can actually Fork on things with inter subjective understanding you don't need a committee so the goal of IG taking is to be free of any committee and that's the superpower of Igan taking that we unfortunately haven't found a way for how to get rid of that for each staking but each AVS can have a journey where they say initially they go on the security conso and then they aify and then they go out of the also because now they are well tested um perfect I guess like just to wrap it up shiram do you have any um thing you want to we talked about the

road map early in the episode do you want to share a bit more about what's coming up yeah um the um the the next Milestone that we we're looking for to hit on the main net is payments because this is clearly the most burning problem people think slashing is the most important thing there are many many existing systems without slashing right like many chains maybe most chains other than Cosmos and ethereum don't have slashing and you know avalan subnets don't have slashing like slashing is a kind of like an optional feature and we have to be careful when going with slashing so the next feature that's

rolling out on Main net is payments and that I think you know sets up the market structure a lot so or at least gets the market structure starting so we have real feedback to adapt it the payment is not going to be in the final form it's going to be a starting form the thing with payments and slashing is these two are coupled in one way because you're paying for node operations and like stti but you're also paying for like are you paying for attributable security are you paying for pool security these are different like economic models so that'll be an avolution of payments along with slashing but the initial form of payments is going that's going to be the next one you know within let's say 3 months coming up on mainnet and then slashing so those are

the two two main things the other thing that we're going to be doing in parallel is conversations like this where we actually talk about like how forking Works what are these parameters and kind of get to a common understanding of okay this is the set of parameters this is the how much you have to burn in AA and like all of this stuff and it's set and now we can because right now even though I uh staking is going to launch soon there's not going to be the fork ability so Fork ability is going to be a feature that gets out the contracts are there to do it we want to be careful in rolling out again like the face deployment approach that we always have so that's the next uh next set of things that

we're looking forward to go live with perfect shiram thank you so much for joining us um really appreciate you taking the time to dive into the white paper and talk about everything that's coming up thank you so much this is so awesome to do this with the zerox research team and block works thank you all thanks off we have teed up with the Wormhole Foundation stewards of the Wormhole protocol the leading interoperability platform powering Crossing transfers for over 200 plus multi-chain dabs Wormhole isn't just a bridge it's the leading interoperability platform for multi-chain applications so go and claim your unique wormo nfts just for zerox research listeners today hit the link in the description deson to fill out a form and to prove that you've

listened to today's episode and receive a PO app you get an email over the coming weeks to claim your limited edition nft [Music]