Slashing Live on Mainnet with Eigenlayer Founder Sreeram Kannan & Risc Zero Founder Jeremy Bruestle

Transcript

all right we are live [Music] [Music] GM gagen gzk ZK is the endgame and uh we are here to celebrate uh the partnership between uh boundless by risk zero and igen layer uh

and talk to you guys about slashing the long awaited slashing update uh Jeremy siram guys thank you so much for being here welcome welcome back to the rollup absolutely good to see you for doing this super excited from our end yeah absolutely yeah I mean guys like uh this has been long awaited you know I think you know so many there's so much anticipation people have been waiting for so long for the slashing update boundless has been making waves on the Z ZK scene as well and so we're going to keep like pretty casual open floor here um I I first just want to talk about like the partnership between you two to ultimately bring slashing to

a head um yeah you know what serir maybe you could start talk and just you know give us a little bit of like what has been going on like aan lay the token unlock etc etc and the road map has been progressing But ultimately slashing has been long awaited kind of catch us up and then and then you know we'll get into what this partnership means uh for both boundless as well as Ian L yeah absolutely um this year has been quite eventful for I can L starting from uh you know all the different things we have had to do this year I just you know things move so fast in crypto that we forget like how how far things have come maybe last year this time you know I layer was just the only

thing you could do on mainnet was stake and you know we had less than I don't know 250 million of take today last year something like that so you know and uh what has happened over the last one year is um the protocol you know went with various other various features on mainnet which is you can now to staking you can delegate to operators you can uh run avss avss can pay the stakers and operators there is programmatic incentives which automatically gives reward wants to St as an operators all of this is now like up and live uh what is remaining and on the horizon is uh when you stake into an AVS uh can the AVS actually slash the

stakers and operators and that is not yet live right and the the principle I think like some people think this is uh um a big issue of course it's a key protocol feature but the way we think about it is when when you when you're trying to build an ABS you know which is an actively validated service a service built on top of wagon lay where you can run various protocols and decentralized networks and when you're building a protocol on top of i l the first thing you really want to do is to build something people want you want to get like actually you want to build a useful functionality so that should be top of mind for if you're building an ABS right so that's the first s so you launch not

thinking about how do you do payments how do you do slashing first let's start with hey can we actually build something useful so that's the first step then you go up and then you say okay I've built something maybe it's useful now can I how do I actually make sure that I'm solvent and I can pay my stakers and operator so that's the next step build something useful then make it economic right so that's where payments come in and then the next step is make it high trust right and that's where slashing comes in where I know that you know these stakers are downloading and running in a decentralized manner that is already some amount of trust right because you have a decentralized network running and validating your stuff in the open it's fundamentally very different from a single note single party you know can only participate in the system so

already there's some amount of trust just by like you know availing of this open permission L Network to actually run your service but High trust comes when you know you say oh it's not just you know these guys have staked and you know they're running the network but if they do something wrong they'll get slashed and so slashing is a key core feature that we thought a lot about but just you know to calibrate the relative thing going on in all of crypto there is no slashing in salana there is no slashing in pretty much any other chain other than in ethereum and in Cosmos even in Cosmos slashing is 5% maximum just get 5% of the Stak slash so I think people think that you know oh I doesn't have slashes no actually like you know pretty much crypto slash we

want to make sure we think carefully about all these things we want to make sure we build the highest grade of trust in anything but there is steps to how you actually execute on execute on this build something people want make it economically useful sustainable then think about adding High trust into it so it's in fact a fun anote is when we launched our first main net I think it's June 2020 July 2023 already our code had slash I mean in some sense slashing is really simple right it just says if if the AV sends a signal burn the money right it's not a hard thing it's not a hard thing to code it's just a really hard thing to get right right because you know you can just say yeah you know

that the guy says burn you just Burn Right like it's not that difficult how much are you burning you know how is it uniquely attributed how do you make sure that you set the right guard rails these are the hard things and so we wanted to make sure that we take a careful measured approach to it so that's okay so that's the road map and why we prioritize things in a certain order in this particular order in fact we did programmatic incentives even before you know abss pay any fees to stakers and operators programmatic incentives is I emissions are going towards you know uh you the the stakers and operators why is this this is because this is like you know bootstrap credits right so you go to Amazon web services first time you know uh risk zero and us we we were

early State startups sometime back they get a $100,000 credit right like you know hey go knock yourself out use the these Services then once you get started and you're already like deep in we continue paying more and more fees to these guys right so that's the same stage we we started with programmatic incentives it's free everything just just use it and as you kind of graduate and go deeper you'll start figuring out how to make it sustainable okay on slashing there are a few things which are really hard and why I'll I'll convey why we're very excited about the partnership with boundas Ando in in Broad terms when we think about slashing what is slashing slashing is an AVS telling that hey you know this portion of funds where actually you know uh the

stakers of this uh stakers and operators need to get slashed okay so that's the kind of high level of slashing and when we so so one other thing I think this that also happened this year is we redesigned our slashing protocol so the original slashing protocol was what you would call Pool security pool security would mean I put a bunch of money and I stake and into like 20 avss let's say I had 10 million that I'm staking into 20 avss any one AVS goes wrong I'll lose all my 20 million so that was the original pool security model and we had already return return some kind of high level white paper about this stuff called unique security so the unique security is I have 10 million and then I can allocate you know

1% of that to this AVS 20% to that AVS and 7% to the other AVS I can be very precise in controlling how much I I allocate my stake to these different avss what that does is for each AVS now when they know that they have you know 1 million slbl stake that that is uniquely attributable to them so they're not like competing with other people to actually like you know slash that stake so that makes the accounting you know people are worried ion ler is overleveraged there no over leverage you can actually precisely track how much each ABS gets so this was actually like you know in some sense it's an obvious system in retrospect the reason this was difficult was there is another thing that we want to do which is hey imagine I'm spinning up a new AVS okay I come in and I say

hey here's the new AVS it can do some really cool stuff and I want you know stakers and operators to support it the amount of unique security I can get initially when it's a new service will obviously be small people be like okay you know maybe I put in 500k not I'm not going to put 500 million on some random new service so it's it's small it's understandable it's small but if I'm I'm only getting that security of 500k why not just go get that somewhere else like why go to like a big platform like ion which has I don't know 15 billion sitting there and we kind of like balanced these two like totally different kind of concerns one is about unique attributability of stake on the other is we want um some kind of bootstrapping feature that anybody should be able to come in and get bootstrapped with you know not high risk

and so we have this other feature we call Total stake total stake is I have 10 million and I'm staking that to this AVS but I'm only can be slashed for 500 okay and so now I mean what is this this is interesting so why is my total stake matter total stake matters because imagine totally for this AVS there is 10 billion which is Stak in total stake right but only attribut or uniquely only like 1 million 2 million is being sted okay so now to attack the AVS I go and just like I'm an attacker and I'm just trying to get new Stak in and try to attack the system putting up 1 million is not enough now because I I there is 20 billion St for all the other other things together and

they're all like supporting in their weight in this so you can think of like one is the Cil resistant weight that you're allocating to the AVS which is the total State the other one is specifying how much are uniquely slashing and so what this does is if I'm a new like avas I can come in and gain a lot of stake which protects me from randomly somebody coming in with new stake and attacking unless you have a lot of stake you can't do that so it gives a benefit to the tail end of like new avss solving their bootstrapping problem on the other side you want to build a very high quality Ava saying that I have $100 million weekly volume I have 100 million dollar of stake I'm perfectly secure right like that kind of a statement is also possible to make so

it serves all the category of use cases that so this is a major upgrade we made to our slashing protocol in you know uh mid this year so now that's under development M so okay so now this sets the context for why you know how slash so now now that slashing is clear how it works now the important thing for if you are if you an AVS whenever you're counting whether did you get enough signatures right you know you need to know whether whoever is sted into your abs has submitted enough weight of signatures right so let's say there are three operators each of them have 10 million have I gotten enough like a majority of the operators to sign off right so that calculation needs to be done on chain to do that calculation the first thing I need to know is who are

how much of each Staker and operator have opted into this AVS we call this the stake table how much is Stak how much is this guy staking into that AVS uniquely or totally whichever type of model that I'm using I need to be able to know that and this keeps changing right because somebody will be staking they'll look at the the AP and they'll go somewhere else you know people come in and out like pretty frequently okay so this table is not fixed and static it just continuously changing so we need to have a mechanism by which we bring in this table securely into the blockchain right and you can theoretically compute it all on ethereum but it's really really expensive to do all these calculations because there are so many avas so many stakers different

models of security all of these calculations have to be done and brought in on chain and so you know we were thinking about hey well how do we do this how do we do this calculation should we run it like our own rollup should we what should we do right like this is a kind of technical architecture kind of question and then we started realizing that the whole security of the system is absolutely dependent on this one piece because if I can cheat you know if if nobody has staged into my avas but I can show that oh I've sted 20 million into the avas or 20 billion into the AVS and cheat this then I break everything Downstream of it so the state table is the most important most secure thing that we need in the system so obviously when your demands for security

is high you want to think about what is the highest most secure mechanism by which this can be brought in and so that led us down like oh maybe we should do ZK proofs so and you know of course ZK is great in in theory but in practice you know if I want to actually do this is it going to be easy do I need to write like some custom circuits to make it cheap and you know how hard it is going to be and we were in this scenario till you know we had a chat with our friends at RIS zero and you know they they're an amazing amazing customer friendly team they went all in right Tim kon's from Jeremy team and several others like did a lot of work with us in actually figuring out oh yeah okay this is how you should do it we'll actually help you

know partner with you and actually figure out how to do this together this is because this zero is specifically designed for arbitary general purpose computation right and you don't have to write it in evm you can write it in Rust you can you know the the level of programmability and expressivity is really large and so that was that set the kind of table for this partnership how do we bring the stake table which is how much stake is in each AVS and update it relatively frequently into like a single update lowcost update on chain how do we do this because once we have this on ethereum it doesn't stop that the crazy thing about like this stake table is once I have a secure mechanism for bringing it into ethereum I can also Bridge it to every other chain now on

every L2 on salana on near on like every other chain you can actually get this information about who State on what services on I layer what that means is suddenly now these avss are usable from anywhere so you don't need a new reaking platform for Solana or you know of course they're welcome to build it and we we encourage people to build all these things but you don't need it because you can actually do this stuff so this is the high level you know of the partnership but I think if you think about the intersection of ZK and you know what R zero is doing with boundless and and other stuff and what we're doing at they're kind of perfectly complimentary right we think a lot about

cryptoeconomic security which is how do you bring slash ability how do you bring uh you know the power of Economics into the trust model and ZK is cryptographic trust models right and these two kind of like jigsaw really nicely in in various ways for example an AVS wants to trigger slashing so the AVS needs to write a slashing protocol how does it write a slashing protocol you can actually use R zero to actually write like you know express your slashing protocol into a single ZK proof and if something wrong happens just go trigger this you know submit this ZK proof in fact you know the zero team has created an OP stack version which which operates on a principle like this so there is a lot of things like this that we can do going forward where every AVS when they're

thinking about I need to build a slashing protocol then now they have to worry about how do I write my slashing protocol do I need to write it in in in evm you know but hey I'm writing a movement move chain like I don't want to write an evm like what can I do and that's there's a lot of things that can be done d stream of this but this setss the stage for like what what we've been doing also one cool thing is there here in Seattle you know just a few blocks down the road with us and they come from a similar kind of deep TCH futuristic thinking team so I've been very excited to you know knew Brian Redford the former CEO very well now Jeremy so really excited for this partnership Jeremy feel free to pick up where sir left off I you know I think

that was a great um intro to uh essentially why this partnership happened um and and why you guys are both so excited about it so Jeremy would love to hear your perspective yeah yeah absolutely I mean I think um you know to to shri's points um you know we one of the great things about sort of the the zkm is the capacity to sort of Express arbitrary logic um including existing logic and get a zero knowledge proof made of the correctness of that right and in the case of this protocol right there's a lot of calculations that you know would otherwise need to be done on chain um and you know with uh you know like steel for example we can actually literally take and run these

sort of um contracts that gather all this onchain data make a lot of computations on it put into sort of merized forms for for examination um and we can do that all compl completely offchain in ZK and then verified onchain so you just save you know massive amounts of of gas cost and stuff um you know and this is sort of one of the major we really believe that a lot of computation um is going to effectively move sort of offchain uh because of this right like right now you know um we live in this very scarcity driven system where there's only a finate amount of block space and so if you want to do more Computing you're competing with everyone else for that same block space and as demand for compute goes up you know prices go up right whereas when you

look at the ZK version of this right as the demand goes up you just do more ZK Computing off chain the prices are the cost of the computation itself you can aggregate all of these things and verify them on chain very inexpensively right so you know but but I think in terms of how to actually make that manifest right there's there's there's the question of well yes you can use ZK but there's also this other question which is around sort of well who's going to do the proofs how are you going to make sure they get done how are you going to make sure that uh you know they're done in a decentralized and censorship resistant way especially with core you know protocols like igen you know having uh making needing these to be done as part of the the sort of security process so um you know with

that regard you know we we've worked on the zkfm for a long time but we're now sort of working on sort of the next stage which is boundless which we're we're calling like a universal ZK protocol but the idea is to bring sort of ZK Computing to sort of all chains right so the idea is that it's a way for you to request and receive ZK proofs uh on every chain um you know um and you know the really there's it sort of divided up in sort of these core services like the decentralized proving mechanisms themselves aggregation um settlement mechanisms for how these proofs get paid for on the various chains um and then you know a number of different extension you know that that are that are designed to help uh you know make it easier for

specific use cases like for example in the case we're making heavy use of the steel which is basically a way to run um you know solidity code effectively inside of the zkm right um and you know there's a number of other sort of extensions that we're working on with regard to that um but yeah I think that really you know are we see this kind of offchain computing use case uh as sort of you know something that's going to occur more and more um and you know will help sort of allow some additional scaling of uh blockchain and not only in terms of like do you get this very large cost savings right which we've alluded to but but but also you can start to look at Computing

much larger um sort of things than you might have otherwise been able to do right so and we've seen historically um you know every time someone makes it easier or cheaper to do Computing uh yes certainly costs go down but also new use cases emerge people just do more right I think you know computers have been getting fast at mors law rate for you know the last you know 50 years and uh has anyone ever like just like done and like oh I'm just going to do everything faster and cheaper now no is's always new uh new emerging uh use cases so so we really sort of see uh you know boundless say basically helping to sort of massively expand the kinds of things that you can do in a in a truly decentralized environment and

this sort of onchain environment by getting rid of this sort of computing bottleneck that has sort of been limiting the the space I think um so you know we're very excited about that um so yeah absolutely and yeah to your point I think you know there there's the saying that I take something a thousand times or 10,000 times uh better it starts to look new and you know we can expect a lot of these new use cases to come as a result of this uh performance and bandwidth Improvement now that we have ZK at our disposal we can do a lot of these computations off chain there's a lot more that we can do with it and then we can ultimately bring it onchain um and so you know as we were kind of thinking through this and and thinking

through slashing I I did not expect that uh bringing ZK into the fold was going to take layer make it multi-chain that was one of the one of the uh realizations just from the first few minutes of this of this chat um and then ultimately once it once it is on every chain and it and it's on ethereum and then therefore on every chain there's so much more that we can do with it as a result of this um now I I'm sure you know you guys had plenty of chances to hang out given you're both in Seattle uh and there was like lots of friendly conversations but I have a feeling there was something that you know was said or mentioned or put together as a part of these conversations that you know kind of led the I layer team to choose boundless and risk zero right for this

particular use case and so if I'm kind of you know visualizing this and imagining how it went you know there was so many iterations there was total stake unique stake there was this this drive to bring the stake table on chain and then maybe this light bulb moment suriram where oh we ought to use ZK right and then maybe you started looking at the ZK options that were out there maybe building one inhouse but there's plenty of options that already exist now what was it specifically about boundless what you know what did get mentioned in the course of those conversations that led you guys to say yeah we want to use ZK and we want boundless to to be our ZK provider I mean the the starting point is is obvious uh Brian and the is zero team

Jeremy and the is zero team have been working on this the longest and hardest so it's uh it's clear first time when I heard about like oh we're going to ZK Pro like an arbitrary risk fire instruction set I thought man these guys are crazy it's not happening it's not happening for the next five years and the amount of progress that has happened just like frankly for our whole space and area here is completely mindboggling it's just like way out of like all original predictions so I think hats off to the zero to and in general like our space for actually driving this uh this whole research development program forward um in terms of particularly choosing risk zero I think there were

many many dimensions to it I think the first one is like I said they're the OG they started with this agenda that you just take an arbitrary risk F instruction set and prove it so I've been kind of close with the zero team from the the starting now many more teams are doing it and it's great to see that but you know of course you know there's a we have very high premium for Pioneers okay like that's how that's how we work it's because you know we are in an Uncharted Territory because people who invent one thing are prone to invent other things they'll find new problems they'll find a way to Tunnel through it right and that's that's the amount of trust that we have with the rzo team um and the idea that we can just take you know whatever offchain compute we were

doing and then just like hey here is a ZK proof for it I think that that is crazy that was really really helpful for us uh and then number three is the kind of deep uh uh I mentioned trust earlier but the Deep like customer Centric nature that they have at zero team I think this was you know exceeded far our expectations they're very very very helpful right and when you're we are not experts in ZK right so our you know we we can use you know existing stuff and we don't but we don't have the same level of expertise that a ZK team has and you know how helpful rero was in actually you know helping us steer past like the initial hiccups I think that

all those three Pioneers High trust you know High uh you know customer centricity I think all these three mattered a lot for us and and and I think in terms of that you know sort of customer Centric thing like fundamentally um you know we're really really interested in the Practical solution to making ZK from a theoretical technology into something that it really changes the world right like I think that like even from the beginning our decision to pick sort of you know the ability to execute existing arbitrary risk five code was built by this sort of desire for practicality like when I first started talking to Frank and Brian about the idea of ZK because I was really excited about it because I love the math you know there was a there was

a lot of comments around like well how do you program it like what finite field elements you know arithmetic rints they're like no one's ever going to use that it's like not reasonable from a development perspective and so you know we we've always been very interested in figuring out how to make these things uh you know uh you know pragmatic and usable and actually for us you know we were very excited to work with I layer because they're a great example customer they're technically sophisticated but they they also like by you know they're clear about what sorts of things they want to need and that helps us sort of determine the direction of the product which was really relevant and it was really like you know I like part of the reason you know we with regard to boundless one of like the you know things that we did uh is we're g to

we're going to go from we're going to be decentralized from the start right we're going to be censorship resistant from the start right be these kinds of applications you know you need something that is um you know really has that like level of of of true Independence of the network itself um so I think that you know but I think it's been a great relationship in terms of helping you know we're always really trying with all of the people we work with to make use of that information about what's what does the market really need so that we can really focus on and we're you know very engineering engineering focused team you know um so always like trying to figure out what's what's what's missing what do we need to get from here to there um

so um and and Jeremy I I think you know our audience and the ecosystem at large is is quite familiar with uh with risk zero as this ZK VM framework maybe less familiar with boundless given it's a little bit more recently you know been been out in the public um and so maybe you could you could give a sort of a high level but also like a very practical explanation of how boundless Works um the differences you know between it and risk zero more broadly and then and then you know sort of lead into like how does boundless help to uh orchestrate slashing and bring this slashing table on chain yeah absolutely so I mean I think uh yeah to your point you know bondless is uh you know

somewhat new we've known for quite a while that there's a need for sort of a decentralized improving mechanism in many ways we sort of view boundless as this you know way to make it possible to bring ZK to all the chains and in regard to that there's a number of different components that are required one of the key things is in terms of the design we're not creating a new L2 we're not creating new L1 we're not building a new consensus domain at all in fact where we're actually the the design is um really works via sort of um these settlement uh smart contracts that run on the chain that wants to sort of receive and pay for the proofs um and then sort of a a peer-to-peer Network that does the sort of helping of matchmaking to find who's going to prove what there's also sort of a staking mechanism in inherent in that as well to

make sure that provs perform their their the functionality properly um you know and so in that regard you know we're we're really trying to be very uh you know neutral and provide this ability to get proofs generated uh in as many places as we possibly can for as many protocols as we possibly can using different you know sort of models for for payment Etc um and then you know I think uh to in terms of where we're at you know we uh we've actually been running uh sort of internal uh you know test net um you know Ian participated in this collaborative development program that we ran earlier this year where we actually had a number of thirdparty provs um and a bunch of

people who had proof demand um sort of actually you know kicking the tires on the sort of mechanism the network uses um we're actually uh you know going to be opening up uh you know this process for other developers having providing developer access uh in December this year to on board a larger number of sort of developers to be able to sort of generate uh proofs using boundless um and we we are doing a public uh test net uh q1 next year uh so uh is the current current timing so we're we're going to really you know I think this is something we're super excited about um you know and I think uh in terms of the way that this this process works really the what's going to happen is there's going to be a process by which these requests on a sort of you know every

once in a while basis are to get an updated version of effectively this staking table are going to get built those requests are going to go out to boundless the Prov network is going to sort of um you know generate the proofs make sure it gets posted on chain who you know that there's a there's a settlement process by which they get paid um and then that will keep there being an onchain upto-date uh sort of merized representation of the data that's necessary to run the sort of slashing always there and ready on chain uh and reasonably up to date so that uh you know the the sort of rest of the process can can run without being you know as you know gas expensive so I can add a little bit on on the so we talked a little bit about reso also

about boundless uh I think the idea that you know as a protocol you know if it's a protocol Centric feature like the stti table needs to be brought in and needs to be updated to for you know any AVS to consume to know that what's the recent set of stakes that that is there it's super important that this happens on a regular you know Cadence of updates if an update is censored for some amount of time or just like not responsive or whatever it could be a problem so you know having a you know decentralized per mission network right like boundless to do the proof is obviously like a clear winner in fact it's embarrassing for our kind of you know you the co-founder the rollup right like pretty much all the rollups having like sequencer as a

service to be like deals that you to call and talk to like customer support to actually have the sequencer spin up is actually a little embarrassing for us as a space right and it should have something very similar where you just say hey you know even if it's a single sequencer it's just somebody auctions off the rids and runs it and if they run it well they keep it otherwise they lose it somebody else should be able to stand it should all be automated on on a protocol I think but clearly that that's uh much more if that such a thing were available every roler will plug into it as opposed to saying hey you know and so clearly that's that's what happened with boundless is that once we decided that we're going you know in in this direction it was an obvious choice that

we want to use this kind of a network as opposed to you know doing some you know uh fast deal with somebody trust me we are doing everything we can in our capacity at the rollup to help push the sequencer of the r the rollup space and the sequencers into that in that direction um yeah because it makes so much sense from a liveness perspective from an economical alignment perspective um to to to create this mechanism where we have some fall back and we're not riing on these single sequencers um and calling up the customer service department right to to try to get them back online and so you know when we're when we're talking about bringing the state table because it is it is such a crucial part of every AVS on igen layer and and it's so important

to bring that that staking table data online in a timely fashion and and so that it's complete and correct um you know what what kind of uh like mechanism are we are we looking at we like we realize that it's cryptographic but are we riant on you know this this set of provs or this set of nodes right to bring that data online um like we are like when it comes to the sequencers is it a similar similar uh mechanism I think it it's not right there's there's definitely lots more at play here but Jeremy maybe you could elaborate as to the extent of the set of provs in this Pro Network and and some of the players that are involved yeah so what's interesting is in this case you know the

the network is is entirely open right so um and and unlike uh the case where you know you're doing something like um you know uh sequencing where there must be a distinguished sequencer uh in order for the system to to function normally and to have a little latency in the case of proving um you can you know anyone can generate the proof and land it um and there's there's a mechanism we have for larger proofs for people to to to temporarily uh sort of win a bidding process so that they are have sole rights to produce the proof for some period of time uh but that can controlled by the user along with a stake that's required uh in order for uh people to be able to lock things in uh for you know um being the only the the sole Pro for some short period of time

um and if you don't want to support that you don't need to so if a customer wants to have the provs race against each other that's certainly an option um so but because of the fact that a proof from one party is equally as good as appr proof from another party they're all mathematically validatable completely independently there're sort of objective truth to the you know uh as opposed to sort of inter subjective truth then you get this case where um you know the all that matters is that there's an it there's there's enough parties that it is you know economically there's enough economically rational people that will that you likely get the produce proof produced although you can also additionally make if you're really concerned about you can also additionally make you know um explicit uh M or sorry plays with uh various

provs um so you can say Hey you know like if no one else is going to get this proof in would you know you please do it or you please do it you please do it right so so you also in addition to the Open Marketplace you could also uh provide sort of contractual back stops but fundamentally we you know the goal of the system is to build a uh truly open uh marketplace right from the get-go and then you know you should be able to rely on pricing mechanisms the nice thing is too is that because of the way that the the this this sort of abund sort of if you look at pricing of producing proofs un like execution where there is this sort of finite um amount of available you know compute on a block by Block Level there's there's effectively anyone

can spin up a a approver at any time and so if you're willing to pay a cost that's above the cost of you know what it takes to spin up an AWS on demand instance then there's an Arbitrage where someone's going to spin up an instance and run the proof and you know get the payment um you know and so it's it's it's it's very you have this very very strong availability um and because the settlement mechanisms all happen on existing um smart contracts on the existing chains in qu question you're actually inheriting the sort of censorship resistance capabilities of whatever the chain you're operating on is so there's no additional sort of trust assumptions that need to be made beyond the the chain you're running on so if you're running on eel one you have

all of the security um you know censorship resistance and availability of the eel one uh behind it and you're not relying on some new networks Security in that regard so I think that's a really huge um you know uh factor from a security perspective for a lot of people to I saw you on muted if you want to jump in there and add no yeah I was just nodding along I think it's it's the the security assumption being exactly inherited from L1 is a really strong point here yeah a absolutely um quick question for you siram on on a little bit more the economic side you know uh essentially the there is like this network of provs

that are that are racing or or collaboratively competing with one another to provide the proofs you know for this staking table on the L1 one who is paying for that given that it's kind of like a public good and it benefits all of the avss is is that something that Ian layer you know is paying the fee for is it a particular ABS per request yeah I I is paying initially and then over time what will happen is there's enough like you know from all the AVS fees or payments or whatever there's enough collected so we have to work out the long-term economics but right now it's it's a public good for the agen ecosystem yeah although we did do some analysis in terms of the effective cost of proving and like I said one of the nice things

is the cost predictability will be quite High because the the there is sort of an actual real world cost for these things and we should not expect pricing on the proving Network to go above that by any significant margin or arbitrages occur and it should correct itself uh and you know just looking at the sort of cost of producing these proofs uh on like a sort of a per AVS or a per operator level they're they become very very small when amateurz out over that large number of parties um and so relative to the other costs of running these Services it's a should be very very marginal in terms of any any impact true to the boundless name boundless in terms of performance and

and cost and that amortization is gonna is going to come in clutch uh as we scale this system um got it and guys sorry to put you through this but I have to ask the the question when like this is what people are dying to know like you know this has been long awaited people want this and I think there is kind of this feeling that once we get slashing we're gonna start to see apy and fees turned on these abss and now we're gonna have you know this this very robust Marketplace of economic stake that people are are trading with and and reallocating into different abss depending on the rewards and the slashing conditions and so there is like a lot of hunger for for this particular uh feature to go live you know Jeremy you alluded a bit to the boundless road

map you guys have this developer program that's been ongo Ian ler has participated and now starting to roll out some of these incremental updates but question open to both of you guys because I I there is a lot of you know inter interplay here what like when when can we expect this to go live uh do you want to take the well I can I can answer uh quickly on the the so so I think you know um as I as I mentioned we have a pretty locked down notion for the test net I'm not really quite ready to publicly comment on mainnet but it will likely not be very far after that that said one thing that's worth noting is that um we also have discussed the a number of mechanisms uh in the interim to allow uh the proof generation to occur uh with

with a fairly heavy level of backup uh using sort of you know we have the existing Cloud proving mechanism um uh of bonsai um we've also been talking to a number of other um of the people that are intending to be proof producers uh on boundless when it launches um about providing sort of back stop proving capability uh you know in if to the extent that uh you know slashing is ready to go live uh prior to main net existing um because effectively from the perspective of the uh security you know that you can sort of sort of cover this sort of availability thing in a non- decentralized way uh as a as an interim solution um you know of obviously you know we would like to have the system

fully launched and fully centralized and ready to go um as soon as possible but uh you know there are other ways to make sure that the availability uh exists in advance of that so and then in terms of you know the the the slashing um you know the rest of the project I I'll let shiram speak to sort of that that side of it yeah slashing is right now on devet we have a public test net coming up you know uh within a month right and then uh you know after that it's the you know Audits and make sure that you know contests and everything so that's just a standard timeline after that but that's that's where we are on that um

the like I said uh the hard part for slashing is not just just us building slashing but for avss to not build slashing so it's it's still going to be to calibrate expectation AVS are going to take time to actually sit and integrate things like the full ability to slash but things like the state table are going to be used from day one because you know you have a easy way to calculate now like who's sted into the AVS know this is just useful computation for every every single avas to to do it and you know once it's bridged to other other chains now you immediately have an instant utility of this you know system the the stake table system that we're talking about

here yep and you know you you you're alluding to this uh staking cont or slashing contract that these avss will have to build and and it'll be unique to each AVS and so there isn't you know maybe there's some standardization or templates that that can be done with these slashing contracts but there there will they will go through tweaks and and they will be unique to each particular use case and so you know we we can anticipate that will take some time um and and you know that's one of the things that happens post slashing um and given that it's in devet and soon in testet um you know we can start to work on those slashing contracts for each of these abss but I'm curious what what else happens after that like boundless will go mainnet slashing will go live

I'm manifesting these things will happen now afterwards that's when the real fun starts right like that's when we can really see some very colorful applications via these avss now so Ram like what are what are you excited about generally kind of give us like some some foresight into your mind as far as what happens after slashing goes live yeah um definitely it's going to be like uh new things start amping up once you once you have the full so in some sense once slashing goes live we can say oh the I layer full protocol is now live right like you know every every aspect of it you know is is there so in terms of what happens after that the category of um avss that can be built you know you you

can build any AVS with slashing without slashing but they will start being trusted and transporting more value right so you can say now like when abss are launching they're like they're being used in a lightweight way because people are like okay yeah I know it's there but they don't really have slashing so maybe like we'll we'll wait to actually amp it up which is actually a healthy thing in the crypto space we you know push things off like before we we're ready for it but that's that's a that's something that we expect and for example what kinds of things right so I I gave a talk on on like what categories of abss should be built like a year back and uh you know there were like 20 categories that I talked about and and basically every single category there is multiple

ABS is being built and you know the rollup has done some kind of a series on on this uh ion L unlocked I think that's a really really good one for people to watch um but in terms of these different categories of abss like what are the things we are most excited about I think in the rollup space like what are the various things that can augment rollups right fast finality verifier networks me management uh da right so we you know bridging right like these all for example if I want to do a lot of fast bridging across rollups as long as I have enough economic capital I should be able to like move stuff around so it's clearly as as a so these are some immediate categories of uh abss that come up in the U rollup space then you can think

about co-processors right so I want to run a SQL database I want to run a gaming virtual machine I want to run a AI right off chain I want to run uh you know complex ZK proofs for example like you know align layer is saying hey you know can I post pref facto get pre-commitments for like you know ZK proof verifications off chain so those are some categories of uh co-processors that that we excited about and then we see um cryptography right multiparty secure multiparty computation I have a secret that I don't want to run on one note but I want to spread it across mult multiple nodes you know we have a partner called silence which is building some really cool stuff that actually

like now they're talking to Bank of international settlements and other places whenever you have multiple parties who have information and you don't want to share the information but you want to do a computation on top of that information that's a really hard General problem and you can kind of solve those using these networks you can think about fully homomorphic encryption which is a mechanism of I just compute on top of the encrypted data um we have inco and Phoenix which are building fully homomorphic encryption on top we have uh uh trusted execution environment networks I can run inside cryptographic enclaves inside Intel or AMD or whatever so that's that's another so these are categories the other one which you know has been kind of a little bit surprising for some people but we've been anticipating this quite a bit is the

category of ZK TLS or pro web proofs right so I want to prove like what is happening in my bank account on a blockchain right and you know I don't want to tell everything but I want to prove some aspects of it so these kinds of things are very very powerful because now as we talk about like how do we on board the next one billion everybody talks about this right and I think and and you know we also say oh infra is boring you know I don't need to worry about it you just need the ux no it's not just the ux it's actually making more things verifiable I think Jeremy alluded to this earlier nobody ever said that hey you know I have enough compute power now right I think the same thing will happen here there's no bar where we set oh now we

have made enough things verifiable no if you can make more things verifiable you should make them verifiable because it's just useful to everybody and so this is a I think a key unlock where now now for example I can come into a blockchain and prove that I have a certain rating on Uber or rbnb and now like you can vampire attack cloud from crypto right so you I think with the bull market this is there's a lot of opportunity for new cool crazy consumer apps to be built on top of this so all of them going live I think they can all be built and they are being built even without slashing the slashing just like endows a lot of trust on these services and and then they can just in you know if you go to this imagine somebody going to the bank of international settlements and saying hey

you know we built on built on this we have $10 billion securing it is a very different kind of a statement so that's that's the type of thing that you know slashing unlocks and that that's part of why the the boundless risk zero igen layer partnership is is is such a perfect marriage because it it does combine the best of cryptoeconomic security with cryptographic security so that when we're going to some of these traditional institutions But ultimately have a lot of the adoption uh locked into them you know we can unlock them we can make them verifiable we can bring them on chain and and we can give them the best that we have to offer we have this entire Suite of economic as well as cryptographic security and so why not give them the best of both worlds and

and S you mentioned thank you for the shout out the Igan layer unlock series it's been it's been an absolute pleasure we've I've personally learned so much and one of the things that I'm most excited about is seeing these abss work together and so to to think about how you know the fully homomorphic encryption from inco can can help prove and uh it can you know work together with ZK TLS and prove what's in your bank account we can use that for KY using something like predicate all of these things are infrastructure in their own regard but when we combine them they start to look like the very very uh robust pieces of an application that are important for it to launch Jeremy um you know you guys are absolutely on this path to main net and and such a critical piece uh to this evolving internet of applications

onchain This verifiable Cloud through rum likes to call it um and so you know absolutely you know pleasure having you you guys both on this on this show today and talking about this partnership Jeremy would love to give you a last word and just hear your closing thoughts here as as we come to a close yeah I mean I think um I would just you know recommend people start hopefully anyone who's interested in um you know getting ZK proofs uh get into sort of our developer onboarding process um you know we'd love to help people get you know things going we really see this as the direction of a lot of computing on on uh you know blockchain moving forward and yeah I'm very excited about the the igen you know um partnership I think it's going to be great for everyone um yeah

and uh yeah we're also we're also coming out with a new version of the zkv M uh quite soon uh which will be adding a new uh mechanism for what we're calling programmable pre-compiled which will give you uh additional acceleration on a number of operations uh that I think is going to be uh pretty exciting as well um so yeah um but no I I think that's I think just you know that that's where I've got some awesome everyone listening thank you guys so much for joining this was an absolute pleasure go follow sir ROM go follow Jeremy Jeremy's criminally underf followed on Twitter so like please definitely give that guy a follow um you won't you won't regret it and uh yeah guys thank you so much congrats thank you it's good talking to you thank you Jer you take care you too