Episode 337 - Restaking Research with Naveen & Tarun

Transcript

[Music] welcome to zero knowledge I'm your host Anna Rose in this podcast we will be exploring the latest in zero knowledge research and the decentralized web as well as new paradigms that promise to change the way we interact and transact online this week tun and I chat with navine a graduate student at Columbia University about his recent work on reaking we start with a look into naveen's early work on matching markets and how this led him to work on mechanism design we then discuss how the concepts of reaking were first presented

to the public and how both naven and tun have been working to better model the mechanisms underpinning reaking to understand how they work and to figure out how they could be optimized now before we kick off I just want to point you towards the ZK jobs board there you can find job opportunities working with top ZK teams I also want to encourage teams looking for top talent to post your jobs there as well we've been hearing from more and more teams that used it that they have found excellent Talent through this ZK jobs board so be sure to check it out I've added the links in the show notes also quick reminder the ZK Summit 12 is coming up it's happening in Lisbon on October 8th be sure to grab your spot as space is limited this is our 1day ZK focused

event where you can learn about Cutting Edge research new ZK paradigms and products and the math and cryptographic techniques that underpin ZK systems all info can be found at ZK summit.com and hope to see you there now Tanya will share a little bit about this week's sponsors Alo is a new layer one blockchain that achieves the programmability of ethereum the privacy of zcash and the scalability of a rollup driven by a mission for a truly secure internet Alo has interwoven zero knowledge proofs into every facet of their stack resulting in a vertically integrated layer one blockchain that's unparalleled in its approach Alo is ZK by Design dive into their programming language Leo and see what permissionless

development looks like offering boundless opportunities for developers and innovators to build ZK apps this is an invitation to be part of a transformational ZK Journey dive deeper and discover more about Alo at al.org galot is the first decentralized proving layer with galot users can generate and verify proofs using any proof system for any use case you can use one of the default provs from projects like Aztec starnet and polygon or you can deploy your own galot is on a mission to dramatically decrease the cost of proving by aggregating proving workloads from across the industry to better utilize underlying Hardware while not compromising on performance galot is

offering Priority Access to ZK podcast listeners so if you would like to start using high performance proving infrastructure for free go register on gal.com and write ZK podcast in the note field of the registration form so thanks again galot and now here's our [Music] episode today tun and I are here with naven a graduate student at Columbia University in computer science he works with Tim Ruff Garden he is also working on mechanism design at ritual welcome navine thanks for having me it's great to be here nice and hey tun hey excited to be back for our second recording this week um true it's a it's a it's a record yeah

although to The Listener it will be a week later but we know it's the same week um so today we're going to be talking about reaking and we're going to be revisiting that topic tun this was actually your idea to invite navine on the show so maybe you could share a little bit about what you have planned or what you want to talk about yeah so I think in my mind there's always this sort of like leading lagging cycle in research where there's often times like things that grow a lot in crypto but no one knows why they work or how they work and sometimes hey they're total Ponzi schemes but other times you know it's like Unis swap or you know maybe maybe other l1s are a good example and I think

there's often times this this thing where like a little bit afterwards people start trying to formalize like the the reason why something works either from a distributed systems lens sort of economic lens a cryptography lens whichever direction it may happen to be um and I think I first ran into nen's work before he was doing anything in crypto maybe like a year and a half ago or two years ago um when he he wrote a paper on auction Theory as an undergrad you know it was more from the the lens of sort of online learning and the kind of more the machine learning school of uh the world I'm not sure if that you find that offensive or not uh by

because I machine learning just means so many things now that like some people are like you know could mean everything from chat gbt to like linear aggression so just trying to it's umbrella term but it was it was more in that vein and a lot of times I think people from that world tend to not like crypto or maybe view crypto suspiciously uh especially in the research side but navine then kind of very quickly WR wrote one of his first papers in crypto uh on reaking formalizing a thing that I think people you know even the people who invented it who had spent a lot of time writing research on it hadn't quite gotten correct so I think his ability to kind

of move between fields and and generate new research really quickly I think was quite impressive and I because I think in general systems like reaking are going to I I don't think they're going away like they might have different names and people might do might you know change some of the designs a little bit but I think the reason you know those $15 billion in those systems is people do view those as sort of the new way of having a growing ecosystems and building l1s versus just pure rollups cuz I kind of think like the natural extension of rollups and l1s ends up being things that look like reaking so that's sort of the my my Spiel and Preamble of why I think it's

kind of interesting I mean of course you know naveen's come at this from a totally different angle where he he and I both read the appendix B of the IG lay paper and got two different interpretations of it who uh and and uh then we both worked on different research problems but I think it they kind of uh kind of converge so anyway maybe maybe it'd be great to talk a little bit about what you worked on before how did you get interested in it and before we talk about how you got into crypto research yeah sure so uh I guess research has been a thing I've been thinking about for quite a bit of time initially started working on some stuff uh I guess back in middle and high school at Maryland primarily focusing on these kidney exchange uh matching

markets I I guess I kind of got into research somewhat as a fluke um some science faer judge thought it was maybe worth some professor's time to to work with me I frankly I don't know how that worked out but when you say kidney you mean like kidney kidneys like Health Care kidneys okay that's exactly right maybe maybe we should talk about what matching markets are and kidney exchange like detour yep yep so this is the stuff I I worked on initially and uh the general gist there is that uh lots of people obviously need organ transplants and before kidney exchanges were invented the main way that you would get an organ transplant is through what's known as deceased ster uh donation where if you sign up to be an organ donor you

get put on this list and uh if you know you happen to if now patient needs a transplant um there is some priority ordering that's determined among people that need transplants um and uh based on that ordering they get allocated uh kidneys first of all the supply of transplants is is low and second of all deceased donor donation like deceased deceased organs are not as high quality as living uh donor donations and so uh if you did have a living family member or someone that's willing to donate to you um that's obviously preferable um and uh basically Al Roth came along he's a professor at Stanford and uh he worked he started kind of the the seminal work

on kidney exchanges and the idea there is uh let's say I have a kidney sorry I I'm a patient and I need a transplant um and then I have someone I know let's say a friend or a close family member that's willing to donate one to me but we happen to be biologically incompatible so in the old model of the world nothing could happen here right um but now you kind of had this idea of okay what if you had a bunch of these pairs uh so maybe I need a a transplant and my sister's willing to donate one to me but we're incompatible um and then maybe uh you and I need a transplant and and tun's willing to donate one to you but you guys are incompatible um but now let's say that trunin is actually

compatible with me and my sister is compatible with with uh with Anna then we can actually swap right um and so kidney exchanges are kind of these matching markets actually without money cuz it's lots of policy and law around what you can do around uh organs and organs and money um the one kidney equals one kidney there's exactly exactly you're tring kidneys kidneys for kidneys there's no no money involved and there's all sorts of other constraints around um because you can't place a contract on organ donation actually the donations have to happen simultaneously if it's a kind of a cyclical structure like what what I just described um then people also realize you could add some other types of trading structures where let's say jun's just an altruistic he

wants to donate his kidney uh he can give one to to me and now my sister can kind of choose to pass it on and then donate to someone else and now you can have kind of an asynchronous chain of transplants wow so this was one of the like I think one of the pretty motivating examples for me of an example where kind of math and Market design could be pretty helpful in the real world um it was a pretty practical problem and uh I worked on kind of initially more along the lines of how do you do learning for kidney kidney exchanges uh in terms of if I'm a patient can I get some predictions for what I will how long I'll have to wait uh what what the outcomes will look like stuff like that and uh kind of got gradually more and more into the

theoretical side uh of things uh I started thinking more about learning problems in general and uh I had a brief stint also working with Scott commers who's also now in crypto and uh this was back in high school and I worked on some matching Market problems with him uh so kind of got more into the standard mechanism design literature as well while also learning a bit more about these learning stuff mhm and it there wasn't really like a cohesive thing it was kind of just working out a bunch of these different hodg Podge projects that didn't really uh didn't really touch each other that much maybe a general theme of mechanism design and a general theme of maybe learning um but no kind of coherent narrative and then uh when I went to Berkeley I worked with NAA hugb and she works on online learning and

mechanism design together uh so putting together kind of theol tools of optimization statistics and economics and so that's kind of where I worked on this uh learning and auction stuff that that TR was talking about and there the idea is like let's say I want to run an auction every day is there a way for me to figure out how to run an auction each day so that in the long run I'm doing as well as if I kind of picked the best auction in hindsight at the very beginning so these were all like Theory heavy Pro uh projects um and uh I kind of moved from more practical stuff on the kidney exchange side to more like theoretical econ stuff and the math was super cool it was it was great um but then when it came time to like apply for

PHD programs I had a bit of a crisis of Faith per se in that I so I was I was ring on all these three problems and I I kind of felt like I was working on things in this weird uncanny valley like uh the the problems had when you write the motivation section for the paper you say oh this supplies to the real world like here all these things that people care about but the the results are kind of like too contrived to be clean theoretical results and they're also just like way too theoretical for anyone to use this this stuff in practice right so like the online learning and auctions paper uh credit to tun for actually uh maybe thinking about how to use it practically I I couldn't think of a way someone could use that in practice it

was uh it was it was a nice theoretical result along the lines of maybe possibility for learning but the actual algorithm that was proposed would take eons to run properly um and so I kind of decided okay I want to either go on the very pragmatic side and and do stuff that people actually care about or go purely on the theoretical side and and prove nicer theoretical results than these ones exactly exactly okay which direction did you go so I chose to go so I guess I guess I still work on Theory stuff but I chose to go fully pragmatic in the sense that uh I didn't want to work on a problem unless I really knew that there was at least someone that actually cared about this stuff okay and it was around that time too that I had

seen both Scott who was a previous mentor of mine going to going into blockchains and then also Tim roughgarden who who you know always looked up to him he founded you know one of the founders of the field of algic Game Theory he was he also made the jump and and was talking about uh EIP 1559 um and so I was like okay a lot of really really smart people that I look up look up to have have gone into this area and it seems like a place where kind of all the all the nice ingredients come together you can work on cool mathematics prove theorems and then people actually care about those theorems and yeah that's such a rare thing to find altogether um that's cool so that was uh once once I kind of got sold on the vision it was it was not a

super hard Choice uh for for my PhD process yeah working with Sim has been pretty awesome I mean I will say the one thing maybe one redeeming feature about ML and AI research is at least people care about theoretical guarantees uh which is just not true in a lot of other parts of the world where like the theoretical guarantees are so divorced from what people do in practice that it's like they really are mental masturbation there's really no yeah I mean one of the maybe to go on the the more theoretical Sidetrack one of the main ways of analyzing I guess the performance of a learning algorithm is is something called VC Dimension uh which looks at kind of the worst case

it's a it's a type of worst case analysis for Learning and uh if if you looked at those bounds what those bounds predict current performance would be for for ML systems like it's really bad there's there's a huge divorce between what we're able to prove and what people actually are able to do in practice and so theory has been kind of playing catchup uh on the ml side because it's just so hard to analyze stuff and uh proving theorems that actually help practitioners is really hard uh because practitioners are just doing stuff that really just shouldn't be possible if you're looking at a worst case analysis um and doing like a good average case analysis is is quite tough I guess you know now as we kind of move to uh kind of what you've worked on within crypto you know I think there's sort of two

main areas that you've worked on um the first being reaking and reaking risk and the second being resource pricing so you know for our listeners who are unfamiliar with both of these areas which could actually be a quite substantial portion because a lot of listeners are cryptographers or uh developers who who are maybe not quite so tuned to these areas could you Eli five them and give kind of a a kind of high level description of both of these areas and why they're important yeah sure reaking actually maybe to tie us back into The Narrative of how I got got here in the first place when I started working at Colombia and I roughly the same time as when I started working at ritual too I I wanted to just get some outside engagement to learn about how

stuff worked because I knew nothing about blockchains frankly about a year ago oh wow um and so I was uh I was just trying to learn you know what's what's the infrastructure what do people care about how can I I was just trying to figure out how can I start working on the problems that I think people care about the fastest and uh and so I first heard about reaking actually while I was at ritual um because you know ritual is now uh working with igen layer and igen layer was was coming up reaking was coming up and uh the main idea there is uh so you have ethereum and ethereum is a very secure blockchain in the sense that there's lots of validators right maybe if you have other protocols that you might want to run um and the reason why you might want to run them is that

perhaps they have functionality uh that extends beyond what ethereum can do um or other you know any other blockchain and uh you kind of have this issue of if I want to run a protocol with a proof of stake system I need to recruit people that want to put stake in my protocol right and uh that's that's a tough ask uh because you're essentially asking people to lock up money and uh anytime you ask people to lock up money you kind of need to pay also a cost of capital expense uh because they could have been using that money for many other things right and so you need to compensate them for for for that loss that they would be facing so that's a pretty expensive thing to ask for and it's also just like

how do you you know building a blockchain is almost like building a city like you have to coordinate so many parties yeah um how do you just get these people in the first place uh and so the the idea behind restak I thought was pretty cool it was uh you know you have these these validators that have already committed St to ethereum let's say and so in in principle the cost of capital expense for that stake has already been paid to them by ethereum we can talk about how that plays out in practice obviously there's there's uh different quirks there but like in in principle if if I think it's rational for me to put my stake in ethereum that means I expect that the rewards from ethereum uh outweigh any cost of capital expense and so there's all this stake that's already locked up in ethereum and

now let's say I have some other protocol and I need Stak uh to sec to secure this protocol right kind of the simple idea behind restak is okay there's all this stake that already has this cost of capital expense paid off what if I were to just reuse those validators so those validators that're currently on the hook for ethereum the stakes obviously serves multiple purposes one in civil civil resistance um but also in terms of uh you know it's a slashing penalty that could exist if they try to do a double spend attack or any other malicious behavior and so what if I is another call could say hey there's this existing stake that's already getting its cost of capital expense paid off all I have to do now is pay those validators the additional operating expense of

operating my protocol and now I can recruit them as well and so now that same stake it's on the hook not only for ethereum but it's also on the hook for this other protocol and so when I first heard about this I I kind of had two thoughts one is okay that is a that is a neat way of of recruiting security that's that makes things you know frictionless or maybe maybe not frictionless but with less friction um other thought was immediately okay we we've seen lots of weird structures in in crypto that have led to uh you know Financial demise is this a safe thing to do is there a way this can be done safely or or does this place an excessive burden on validators that perhaps could lead to some type of catastrophe in a worst case situation yeah and so on the reaking front kind of

the first thing I was thinking about was okay is there a way to understand what is the risk that maybe reaking poses both from a global sense but also you know if I'm just an an AVS a service that's that's recruiting security can I understand what the risks uh are there because you know as soon as you start sharing validators the consequences that occur on one protocol can start affecting security for another protocol right and so yeah essentially understanding first bringing out what is a good model to understand all this stuff and then second of all thinking about are there is there a nice crisp characterization of under what circumstances can this be done safely and under what circumstances should you be kind of concerned that things might

go go Ry in this case though so like I mean I think we're getting to this work that you recently released called robust reaking networks but was that more like a description of what already existed or is that like a did you rebuild it from scratch and say like this is an ideal way to do this so when I first started working on uh on on this stuff and and TR mentioned this too essentially the only uh only document technical document about reaking slit security was this initial thing put out by ien layer okay uh and in particular there's appendix B of this thing which was on cryptoeconomic security and so that that appendix did lay out kind of a very basic model of there are some kind of validators on one side there there validators that might want to or I guess

they call node operators that that might want to uh put their stake in some number of protocols uh and on the other side there are these protocols that want recruit security and then they kind of talk about a cost benefit analysis of under what circumstances can you say that the benefits of attacking some set of protocols uh don't exceed the costs it was a it's not a super long section I I think it's a it's a pretty nice initial model but that was essentially all all that I kind of could understand about the system uh as of then and so the first thing that I kind I don't know maybe to to go into what I was actually doing in in that paper like half the battle I think in in doing blockchain research is uh is modeling MH um unlike

Theory work where the problems are extremely well specified they're clean mathematical statements and someone just has to go out and prove a theorem the challenge here you know most of the challenge in some cases is actually coming up with a good model a good way of a formal way of understanding basically what's going on in practice so that when you prove theorems they're relevant to the real world right and so there were kind of two main modeling things to think about one is how do you characterize risk RIS here or what what does risk mean and two is just what is a way of understanding all the entities and their interactions and so from that Igan layer layer document B yeah that's right that's right um I kind of started thinking about things in terms of a bipartite graph one thing to to note is

like I think there was a lot of technical documentation on how the implementation would work and like what the code would look like right but there wasn't any document on like guarantees that you would get for you know and I think the only thing was this kind of tiny little thing and I think when you read what Twitter was writing about reaking and like oh it's going to all blow up dot dot dot it's Luna again whatever and then you read you know you you kind of look at the code you kind of see like a very Divergent version of the world and I think part of this comes from the fact that it's actually quite difficult to describe these guarantees and I think you naine probably can can talk about you know why going from like what was known to like a a rigorous but

simple formulation is actually kind of a a difficult process yeah sure you're kind of working in the fog here right there's there's so many people saying different things right and uh the theorems that you prove are not helpful unless there's kind of a first a general formulation for how to think about everything and there are many aspects to reaking the aspect that I was focused on was first just like global global and systemic risk type of type of stuff MH and so the model I kind of first kind of just dialed out a lot of the stuff on Twitter and mostly just thought about okay at at the very core of this thing there are these validators and they're being reused among some Services right and kind of following

some stuff in in appendix Fe you can think of there being some profit from corrupting any service right if I if I launch a double spend attack on some service there's some amount of money I can make yeah as as to whether or not that's a number you can actually know like that's that's a separate discussion but for now we're just trying to put something together let's say let's say this is a thing you know and uh and so now okay so there's a bunch of these Services each of them have some profit from corrupting them and on the on the other side there are these valers and each valer has some stake right and by the way by service do you mean like what became sort of these AVS AVS that's exactly right yeah service and AVS are interchangeable here um and on the

validator side each validator has some stake right you can think of this as as maybe their e stake and so now in a in a world where there's reaking uh each validator might be using that stake for multiple Services right so you can kind of think about this as a bipartite graph where on one side you have these avss or services and on the other side you have these nodes and you can kind of draw an edge between a validator and a service if that validator is reaking or or using their stake for that service if they're if they're acting as as an operator for that service and so now on the service side each each service has some profit from corrupting it right that's the amount of money you'd make if you were

to launch an attack um and each service also depending on how it operates uh perhaps if it's like a pvft style consensus system there's some fraction of of stake that's required among its Total Security that's required to launch an attack on it right so let's say to make things concrete if there's one service you maybe just ethereum and a bunch of these validators right one validator can't launch an attack you need a third of the a third of the stake right um so the profit from corruption would be the profit from launching a double spend attack on ethereum the fraction required for corruption would be 1/3 right and each validator has some stake corresponding to their ethereum

stake and so now in kind of this General system uh so actually maybe maybe we'll start with that simple system with just ethereum right uh you can ask when things you know safe from a cost benefit analysis point of view and that's basically when if you look at onethird of the total staked uh ethereum if that is bigger than the profit from corrupting ethereum right because if the profit from corrupting ethereum was bigger than onethird of the total stake of ethereum then it would be profitable for oneir of people to to come together and launch an attack right and so the Baseline thing required for ethereum to be sort of secure in some sense is that one third of the total stake exceeds the profit from corrupting ethereum mhm so

this is one way to think about things but even that you might not find super satisfying right because uh there are lots of shocks that just happen in the world especially in crypto where there are definitely shocks that happen in the world um these shocks could be for example some protocol does unintended slashing on some stake uh or some stake just like drops out for whatever reason valers maybe go down whatever it is right you can't stop small shocks from from occurring in the world and so let's say it was kind of the case that uh one3 of the total stake on ethereum was exactly equal to the profit from corrupting ethereum right we might say it's secure or or maybe it's like equal may maybe one third of the stake is like

Epsilon more than the profit from corrupting it right so it's it's kind of just barely secure and this is kind of not that satisfying because if there's just any small shock that happens uh to the stake then it all of a sudden becomes profitable to launch an attack um and so you can extend this notion of security to kind of maybe you might call it robust security which is to say suppose that you allow some small shock to take place in the ecosystem and now you want to understand What's the total amount of loss ultimate loss of stake that can happen after after some attacks uh this is kind of what I was thinking about when I was thinking about systemic risk so to make things maybe more concrete you can imagine in a situation where there's multiple services

that uh some small shock happens in the world and uh it causes some validators to come together and attack some Services right and so now because they attack those Services those validators get slashed they lose their stake but those validators might have been reaking for yet other services and so now those other services are a little bit less secure because people that they that were providing security for them have now gotten slashed and so now as a result other validators might come in and attack those Services uh and said this could you know continue go on and on and and a small shock might result in a loss of stake that's that's uh much greater that than that initial shock how did you like you modeled this I guess but like this hasn't happened in real

life has it like you're not it's not based on anything so like how do you know that that's how it works like that they would great question I mean yeah I mean even right now slashing hasn't been uh at least to my knowledge I don't think iir has started slashing uh some of the but but now there are multiple multiple reaking networks some of them have more more slashing and less like it's actually gotten more complicated than the original model but uh yeah I mean again I think there's lots of facets of of how this setup can uh potentially be you know induced risk this was one particular one in in general just looking at cost benefit analysis and and looking at how

small shocks affect a system this type of small shock analysis like this is this has been done in the traditional Finance literature before uh when looking at systemic risk it is kind of a motivated concept on on that side and it kind of is just the if you're just thinking about costs and benefits it is kind of the first thing you might think of uh as okay how do I how do I characterize risk there's lots of other risks you know you might have you might overburden a validator maybe with computational demand or or you know yet many other things but in terms of what can you kind I concretely understand what made a lot of sense to me was to just look at okay maybe different Bridges have some understanding of what these profits on corruption for their services are right and uh you can look

at this whole graph this graph exists you can see who's reaking for whom uh you can look at the stake yeah this is kind of the first thing uh that came to mind in terms of how do you just look at systemwide risk maybe from ethereum standpoint of if a small shock occurs does the presence of all these additional connections between services lead to a greater possible ethereum loss yeah one one thing to note um and I think like this is kind of one of the reasons I think it's good to formalize these things in an abstract math language rather than sort of like something that's like more engineering or like pure kind of like implementation like is that a lot of

other networks have had things that look very similar to reaking but they haven't like analyzed it formally H so for instance like polka dot par chains and reaking are actually very similar um the only difference is that instead of you know this notion of like the matching Market Happening by node operators choosing Services the equivalent of services here would be a par chain instead of the node operators choosing par chains they there was this auction and then they had to validate that particular service so um the matching was sort of done by an auction versus the matching being done by I get to choose where I place my stake but fundamentally they actually had very

similar properties I think the the main difference is that when you're doing something like reaking you can move your stake around and so you don't have to get locked into one service so like the worst case thing that could happen in sort of the polka dot world and I think this is often why people had trouble developing there is like you as the service had to raise a ton of capital to kind of win this auction mhm and the thing is you had to raise capital in dot so you'd have to sell your token for DOT and then like yeah and then you have to do a crowd loan and and the thing is you could argue that the crowd loan plus auction is kind of as close as possible to what reaking is except reaking has it's just easier for the end user like

operators can just choose to validate one Or Another We're validators zkv is a validator in polka dot and so like you're validator in polka dot but you don't just choose which other pair of chain you're valid validating on you also would become a collator on those actually there's another system like cosmos's IC or interchange security which I I'm now I I only am putting it together now that it sounds even more similar because they're it's it's very new well it's like a year old or something and as a validator on Cosmos you are actually actively choosing if you're going to be active on these interchain security chains I don't know if you've if you've looked at that is that similar or is that so so so one one

main difference between the pokedon cosmos versions of the world and the ethereum version of the world is that in ethereum reaking there's a floor on the amount that I'm earning so I'm always earning ethereum staking yield as at a minimum the problem with all of the pokedon cosmos designs is they're very favorable to Adam and Dot as tokens and they're very unfavorable to the services the services have to pay in Adam or Dot and they don't have Capital they have like a small amount of fees they have their token so what they have to do is sell their token buy Dot or atom and then participate whereas an e staking you have people who are already staking eth and earning 3 to 5% yield and this is viewed as supplemental income versus

you know is the main thing needed from the services so I see I see I I think the economics in ethereum are just strictly better and also I think there's a little bit of greediness in the in both IC and the dot auctions in that they only benefit the validators they don't like the social welfare like the splitting of revenue between the services and the validators is very unfair in a lot of ways in the dot and atom worlds whereas the ethereum one is closer to kind of and this the point of these matching markets right is like they're trying to do some welfare maximization type of thing between the two and and like I guess nine working on that in the past probably inspired this particular model you know like what's

like the Genesis story how you what's super weird about the reaking matching Market in particular is uh it's actually a matching market with the negative externalities normally uh normally in a matching Market as uh as matches take place you can talk about welfare increasing but uh here what's interesting is as there's more matches uh there's kind of more potential risk depending on how it's done and uh you know the reason for that being that more things are kind of tied together and and from a systemic risk point of view you might expect things to get uh to get worse actually maybe I I can say the conclusion since I only said what the risk measure is and and not how to how to how to mitigate it sure um but uh the main upshot is that uh over

collateralization is basically what's both necessary and sufficient to to to mitigate this risk so what that means is uh if you go back to the uh the ethereum example where there's just ethereum and then a bunch of validators there you kind of require that to make things robustly secure here where even if a shock occurs things are still safe you obviously need some buffer right between the profit from corrupting ethereum and and onethird of the total stake in ethereum and it turns out that from a global point of view things are kind of similar you need it to be the case that for any collection of services and any collection of validators that can attack those Services there needs to be some buffer some some buffer between what the uh profit from corrupting those Services

is and how much stake would be lost in that attack and you can actually bound let's say that that buffers some multiplicative Factor you can actually bound the total stake loss after any sequence of attacks uh in terms of that buffer so for example to be to make things very concrete let's say everything was always 10% over collateralized in the sense that uh any attack on some Services always costed 10% more that the total stake required to attack some Services is is always 10% more than the profits from from attacking uh those Services then you can actually say that even in a worst case situation let's say that a 0.1% of total stake was just lost

arbitrarily then the total ultimate loss after any attacks thereafter is uh upper bounded by 1.1% so you can really like concretely bound worst case loss in terms of this buffer which is a cool property of course there's there's there's caveats like I said earlier um knowledge of what these profits from corruption are might be limited to Bridges or or folks who are actually running those Services right um and so even though from a Global Perspective it would be nice if you could always have this overcollateralization you might not know how much you need to over collateralize so so so you as a bridge may may not know what other other folks are doing right you like this was only a global result of uh you know okay if

there's some random shock that happens in ethereum then that's contained but how do I know that everyone is over collateralizing right and that's that's the requirement for this for this mitigation to hold true and if I'm a service in particular how do I make sure that I'm protected right um and so the second part of this paper is really analyzing stuff from that perspective of if I'm a particular service I want to mitigate myself against you know I might not even understand shocks happening like I might not have a good understanding of what shocks might occur outside of my own ecosystem right yeah yeah but how does how did those shocks affect you anyways to right right and so is there a way for me to uh in a similar vein mitigate worst case loss if I you

know just from looking at a local perspective if I only know my own profit from corruption and maybe the profits from Corruption of some Partners is there a way for me to to mitigate this and it turns out there actually is uh you just have to do kind of a different different type of overcollateralization scheme andna to your point about collators I mean there's sort of this thing of like hey if you have a collator that's shared across multiple pair chains and they drop out and messages don't get relayed and there's no and there's no other you know it's sort of a similar type of thing that happens Al although Co leaders tend to be per Network so you have like the main validator of polka dot and then you have a per per parach chain collator set okay okay they they're never shared I mean

may I mean you might have the same companies running multiple in our case we only run moon beam and like we yeah we don't share it with anything else I actually wanted to ask you a little bit more on the like alternatives to staking because I wondered if like liquid staking is in any way in the same category I realized it's it's different I mean liquid staking is like you stake tokens and then you have synthetics of those tokens and you can do things but this is also like you stake tokens and then you can do things with those staked tokens so I just wondered if there's ever like if if the reaking research ever looks into the liquid staking and how that played out J you you wrote a

blog post related to this right yeah yeah a couple things um one is the model that Naveen has uh in his paper doesn't really cover the principal agent problem where the node operators don't own the capital that they're staking right so the idea is that for the node operator The Profit that they realize um in this model is the profit from Corruption of the set of services minus the total stake that could have got slashed now if you add in the principal agent behavior of someone is delegating capital and the node operator if they get slashed it's not their money or maybe only some of it is their money right like you as a

validator right the majority of your capital is delegated it's not your your Capital yeah so the principal agan problem naturally leads to this kind of centralizing effect now the nice thing about this kind of graph model is it's actually not super hard to include principal Asian effects so I'm not sure get as strong of theorems but there there's kind of a natural way to do it so there's sort of two papers from 2020 to now on liquid staking sort of principal agent effects there's sort of a very old one that I wrote in 2020 and then there's a more recent one last year by dianis Z zros of uh common prefix SL I forget where he was before s you know

the question is like what's the minimum amount of stake that you need the the oper the operator like you zkv to have to put in yourself in order for the probability of you being like I don't care about getting slash as long as someone pays me enough for being part of their attack cartel um should be it's like almost our own stake on the line like the the Val some amount of yeah and and a lot of liquid staking protocols actually do require the operator to put up some stake like yeah I think for Rocket pool it's like 2 to four eth if I remember correctly and then then I think for a lot of the liquid resing protocols there's there's a minimum amount also MH um and so the idea is like how do you

choose that minimum like what's the minimum you need to like guarantee secur so one way you could do this is extend the bipartite graph description to a tripartite graph where it's one partition is the capital holders who are delegating the next partition is the node operators and the next partition is the services and then you look at the flow across that graph and you know you write out a kind of natural a principal agent problem thing and show that has equilibrium such that if the uh node operators always have at least you know say 5% of the stake like is their own then they're unlikely to deviate from the strategy as if it was their own Capital by more than you know by a small amount and to be fair this is actually

true in society at large not just in crypto that like the agents as in the people who are acting on behalf of the principal like the capital holder often have to put their own money so like for Venture funds or private funds uh or hedge funds often times the the agents the partners who are investing money have to put in their own money and so I think this these types of things naturally show up in liquid staking uh it's just like more messy to reason about but bringing it back to reaking does reaking also have a principal agent percentage minimum yes it does there's no there's no doubt that that's still an open problem I would say that's not that's not soled a lot of a lot of uh yeah definitely designs under specified

currently oh but but but I think you have to understand that like you're with reaking we're starting from nothing so like as to give a sort of example of why I found this really interesting is like in 2019 um when unisoft I guess unisoft was November 2018 but you know same thing close enough 2019 when unisoft first came out there were like all sorts of you know the code was running people were liquidity providers there were like people definitely doing Arbitrage uh but then you would read documentation about it and like the documentation from like people working on the protocol itself was like didn't make any sense right they didn't really have a good explanation for why it worked it was like here's a formula it seems to work

we don't really understand why and like that is what led me to try to figure out okay maybe you should like think of it as this optimization Problem whatever MH and I think what you know naine did was like Hey look at this thing there's actually secretly this graph problem and matching problem if you squint itent it could be and not if zoom out and think of it as this matching problem then it becomes much more easy to reason about and I think often times in research like this a very important thing to do is more important than the the actual solution to the problem often times is like having the right definitions because like you know there's often times this trade-off in math of like I either start with like no

definition and very simple things and I prove a really complicated result or I start with really long complicated definitions and then the result is sort of trivial because I put all the complexity in the definition of the thing instead of in the answer okay who did which kind who's which on in that description I think I think the ideal case is that you have a simple definition and then a simple a simple proof you want something in the middle in the middle of no but I think even even to your point about uh this ending up being a very like clean graph problem there's uh for for those uh anyone who's more on the uh TCS theoretical computer science side uh actually in the case where uh all everyone's the same stake every Service uh has the same profit from

corruption the problem actually of checking for the security becomes uh checking whether or not this graph is an expander graph which uh happen to be some nice there's just some nice structure in the problem I want to go a little bit back to the connection between the matching work and the paper because ter you're saying that sort of the technique was you was turning it into this matching problem but somehow for me I've lost that connection like I heard your beginning naine like where you were talking about what you were doing before but yeah what what's matching in here so I guess as it was framed before in in appendix B uh of of the IG L white paper you kind of just

had these parties right you had these services or abss I guess I think they call them tasks even and then you have these node operators or validators that had some stake right and you can kind of think about these balance conditions in this kind of cost benefit analysis that I was describing of what is it the case that there aren't any validators and tasks for which or services for which those validators profit from from launching those attacks right the thing that that model uh doesn't kind of quite help you with is uh looking at I guess broader counterfactuals or looking at the structure between what validators are are putting their stake in different in different Services um and so is the matching is it between is between

validators and services and AVS okay or av's right and uh a a kind a validators match to a service if they're if they're reaking for that service right and uh normally in a matching market right each each let's say you think about a job market so you have a bunch of these uh people that want jobs and a bunch of these employers right and you can kind of normally think of it being the case that uh I guess it depends on the on the setting but in most common settings uh each match is a little bit it's kind of separate from each other right like if I get a job at at some company uh someone who's getting a job at a different company like it doesn't really matter to them that I got a job at this company now if they're competing with me then

maybe it matters right but uh you know in a sense the payoff that I I got is somewhat separable from the payoff that they got and those things also led to positive externalities right like I worked for this employeer the employeer thought it was profitable for me to do this I thought it was profitable for me to do this there was like you know welfare was was receipt you to buy both parties what's interesting about the reaking matching Market is that uh when a validator is matched to a service and then a validators also match to a different service right so let's say a validator is already matched to ethereum in the sense that they're a Staker for ethereum and now this validator now chooses to also restake that eth for another service them making that choice

has consequences for ethereum right because it it imposes a negative externality on ethereum in the sense that that other service that re taking for now has the rights to slash some some eth and depending on how this implemented this could lead to a negative externality on ethereum in that staked eth could disappear due to the actions of other services whereas in a world where you couldn't match to these things that couldn't occur interesting how are the avss like are they offering I'm guessing abss offer unique incentives to the validators to get them to validate them yep what are these just native tokens on the abss usually so actually actually this is this this is

sort of uh what the my followup paper that kind of took naveen's framework and tried to analyze what happens when uh there's this reward mechanism so I think if you look at it from the abstract graph problem there's sort of some natural way that validators choose Services right you're just like you're given a graph like ZK validator has chosen e Oracle and IG da and none of the other services right that's that's a choice you made and then given that you n paper analizes like how do you attack that on the other hand in practice what happens is basically Services give you fees plus native tokens which you can you know you could think of like blocker

words from staking right yeah and so they give you some amount rewards and then you as the validator have to choose the subsets of services you want to operate right so like if zkv was running a bunch of reaking nodes you would say hey Igan da is giving me 5 eth of yield a month e Oracle is giving me one e I'm just taking the top three from the live right now we should make it clear zkv is not currently running any of these these are just theoretical I'm just I was just I was just using you this is an example so use me I don't mind and and then let's say you were I'm just going down the list of the live avss like witness chain was offering 15 eth a month well you say have 100 eth that you want to

restake how do you determine where you want to put it naively you might say hey I want to put it all in witness chain right because return they're offering me the most yield but witness chain might just be very slash now for to witness chain people I'm not saying you are this is all hypothetical I'm just saying like suppose it turns out it's the most flashable we have to be careful we should almost use like ABS a absb or something I I just want to make it more concrete yeah yeah because I think like if you make a concrete it's like a little cleaner so suppose you get you say okay I put all my 100 eth into the witness chain thing I expect to get 115 eth after a year but actually I got slashed on 90 of it my loss is not just

the 90 eth I lost there's also the future profits I lost like oh I didn't lose I I didn't get that 15 cuz it's expecting 115 do you get kicked out when you get slashed is that why like you would never or you don't receive your dep that's actually up to the service that's up to the service the other thing is I have the opportunity cost of the Lost eth staking yield right because before I had 100 eth that was say Ste like St de and it was earning 3% so I expected to get 103 after a year but now that I've gotten slashed I've lost that 90 and to Nine's Point there's this negative externality on ethereum itself and that ethereum lost 90 e that was staked in ethereum so lost 90 of security due to something outside of

ethereum yeah cuz actually who slashes you it's the is the ABS that slashes you no like who gets to take the eth does where does it just get burned what happens to it the main well it depends but for the purposes of this paper it's burnt it's burnt okay no one gets it it's not like yeah someone take okay it's gone but remember it's reducing the steak Supply right so isn't that yeah but isn't it kind of good for a network when you burn tokens it it I guess I mean there is a macroeconomic effect of you're Distributing if if you know the total market cap stays fixed and you burn tokens and you're Distributing wealth back to token holders but on the staking front you're losing stake so you're losing security you're making it easier to to attack in some sense right

right and so the idea there's this tradeoff between the two but but honestly the the thing is it does allow these services to get security as if they were themselves L1 in some ways and so there's this question of like if I'm a new service let's say I'm an oracle let's say I'm a new rollup that wants a decentralized sequencer and doesn't just live off of multisig forever I need some way of enforcing penalties on the offchain actors who enforce some State transition so in the case of rollup a sequencer in the case of say a ZK prover the people generating the proofs in the case of Oracle the people aggregating the thing is I could start a new L1 right but then I have to bootstrap

everything I have to figure how to get liquidity I have to get stable coins I have to do all of this stuff that's like very hard to get new validators but but the question is like when are the economics such that it actually makes sense for me to to join a shared network even though I have to pay rewards such that it is above this over collateralization threshold versus just completely starting a new one and the idea is is that that tradeoff is a combination of understanding the economics of how much you have to pay in rewards combined with understanding this notion of C security that naven made right because you want to pay enough such that you don't have those kind of risks so you want to be over collateralized that much but at the same

time you also you know don't want to pay so much that it's like cheaper for you to start your own L1 does that make sense and so that's the that's the tradeoff that's the balance that and so yeah our paper is about that and like how much you have to pay like to get that type of security that's what your paper was to and then which is sort of following on n's paper cool but I think this idea is just going to is more fundamental to blockchains that like there's some things that can share security and there's some things that need to be isolated it's very much like lending or perpetuals like the defi aspects of things and I think the interesting thing about reaking from a theoretical lens is it it Blends a lot of the analysis of proof of stake that

has existed for a while that people in consensus think about with the analysis of things in defi uh and sort of like it's like a exactly in the middle of the two um and I think because of that it sort of is a superset of a lot of things people are working on it like subsumes doing all these offchain services like rollups it subsumes some of like how should you analyze the economics of data availability it's like a more General framework to analyze all of these many different economic problems mhm which is one thing I think people don't you know I think reaking obviously people think about the Yi and whatever but like if you zoom out the real thing you should be thinking about is this type of stuff I find it so wild that like last year two years ago when the like ZK rollup world was announced I

remember ter you being like man everything's all paths lead to polka dot but actually there was a lot missing from the rollup model to be more like polka dot and now I think what you're describing just going back to what you were saying before it just seems like it's a different version maybe learning lessons from how polka dot went especially in terms of like launching these abss and like incentivizing them but it has more I feel in common with like that polka dot model than rollups do for sure but you have to remember that the reason that we don't see rollups doing that is that the rollups haven't delivered on all the promises that they were supposed to make right like they haven't they haven't gotten and when they do that they have to have

ways of penalizing the sequencer and there is no way penalizing the sequencer right now right fundamentally I think that is the thing that is missing like even in a centralized sequencer world if a fraud proof executes I should be able to slash the sequencer I cannot do that right now right I there's no penalty and my point is the reaking stuff if you look at it from this matching Market lens kind of subsumes a lot of these different things that people are like making specialized economics for including ZK proving markets CU I think there's a lot of this like service versus consumer versus producer uh relationships in ZK provs right where there's like many networks that might be sharing some Network and like they have to figure out how to price things

theoretically doesn't exist theoretically theoretically but my point is reaking is the first live market that is actually trying to to solve this directly as as the economics problem I I mean I'm not sure the people working on it think of it this way but this is what's emergently happen actually maybe to to give like a second framing of I guess all of that in terms of the of the matching Market I think tr's basically been outlining I guess two separate problems one of them is you can talk about what's the current state of who's RIS taking to whom and then you can discuss the security the security of this and uh you know maybe from the robust security this this is what I was working on and then what tr's been talking about is okay now what actually

is that graph how does that graph get generated and what are the processes that dictate who ends up actually reaking to whom so that you can actually make sure that that graph does satisy you know the original properties of of what's what's safe and not safe yeah and and and and I think this notion of like what the safety and not safety is something that especially from this economic standpoint is something completely unmodeled or kind of given a lot of kid delicate care in all these other networks like da proving rollups because mainly because like right now there's just this hunt for demand for any application to actually use them right and generate fees whereas with reaking there actually are more

applications that like are are like you can imagine paying fees like there's a lot of things that are closer to defi looking things like oracles that like have business models and because of that I think it it it's sort of like the proven ground where you can test out like you know how you should think about the economics for all these other areas actually this is a pretty decent seg into resource pricing perfect yeah sorry uh yeah so maybe maybe we should talk about you know you're working with Tim uh on your PhD who is kind of I think one of his first crypto papers was was about resource pricing and ethereum and how to think about it and you while working at ritual have been thinking a lot about resource pricing um in a more

abstract sense which kind of maps to this uh kind of thing of like there's service providers and there's capital and how do they get match together and what's the price they should pay so maybe we could could jump into that yeah maybe to give a little more context on that too so currently there's a lot of literature on transaction fee mechanisms and you know this is like 1559 or 4844 and there the idea is okay you have a blockchain and it enables some amount of compute but obviously there's a limited supply of compute uh and so you need to figure out how do you price things so that you match demand and supply and uh you know ritual is a train that's trying to enable more complex computation

notably trans actions that that might require AI inference and so there if you're thinking about running this stuff a lot of it will have to be run offchain and you know perhaps through ZK proofs or other stuff it'll be reported back to back onchain and you have this issue of how do you actually price this stuff right like if I'm an offchain entity or even if I'm trying to incorporate this onchain I'm just some node that's doing a workload that is very different from other nodes right I shouldn't expect to be paid the same I'd be pretty dissatisfied if I was being the paid as pay the same as every other node and so there's kind of this General problem of uh on one side you got all these transactions and uh they might have very very different computational demands and

on the other side you have these these nodes that can perhaps process very different workloads in this heterogenous setting and then there's a question of okay in this world how do you figure out who does what work and how much do they pay and so that's that's this resource pricing stuff that I've been working on is that like connected to this like is it in the reaking work or is it totally different totally separate from the reaking work I mean I guess there is there are similarities in that they're both matching markets and they both also apply to similar entities in the sense that just like how uh uh T was talking about uh reaking being helpful for coordinating incentives for interoperability and stuff like that um in terms of what those exact incentives should be and how they should be

determined this is relevant there but largely speaking these are separate separate things got it I wonder okay so this is is a a more like perception question but in the last year you know early on in the year I think everyone is very very excited about reaking and then Ian ler launched and there was this kind of frustration and anger about all sorts of things but it seemed like there was almost a sentiment change against reaking at least on Twitter and I just wonder like in the research though and like in the real world is there actually cuz I also know about a lot of teams that are like now as's and they weren't before so I'm sort of curious looking from the research side if you like see

any of that or have some opinions I think on the research side it's kind of become more clear that a there's a lot more interest in reaking as as things are starting to go from idea to actually deployment um and come coming with that is more of an interest in okay how do we make sure things are actually secure so from a research perspective I think interest in research on secure reaking has definitely become more relevant in terms of like posit or negative spin I don't know I mean like I think there's just a lot of parties that are interested and a lot of parties that uh now care a lot more than the did before about the Practical implementations of these things playing out properly I can see when I asked this question tun was kind of like annoyed I'm not annoyed I think I think

it's funny cuz it's like it it does really remind me of the early days of defi in like 2019 when like unisoft launch and like had $10 million liquidity which at that time I think it was the only onchain contract that had anything close to that to that and everyone was like a it's a scam it's going to be a Ponzi scheme it's going to blow up and there's a lot of fud the second thing that people also always say is like okay the slightly like you know the that's like the the 30 IQ take the 80 IQ take was like well there's so much adverse selection because it's a trading strategy that shows you the price all the time so everyone will front run you which you know is not untrue there's there's definitely but for some assets you don't have a choice especially back

then because exchanges were not listing things anywhere near as aggressively so there were no other primary venues and then I think eventually once people started actually building more things that were used around it and composability became the benefits of composability became way more clear you know for instance for making lowering the LTV for maker and compound and stuff people started being like Oh okay I I kind of get what what it's doing and I think reaking probably just has reached the point where haven't been any abss that have like had that impact yet um but I think like I said I think the fundamental thing if you zoom out and and don't go based on like Twitter where everyone is just like oh I I got 5,000

free points for yield whatever like like ignore the this is what I'm saying you have to like zoom out Beyond this like people who are just like yield maxing on Twitter and then getting angry that they got fewer points whatever that that that doesn't have to do with the mechanism that has to do with just like the the projects Ponzi funding things but uh I I I think like if you zoom out the the main point is is that it's a way of deciding having networks where you can either decide to share Security in a very precise way that has some covenants and guarantees which means that you're taking on some risk for doing that but obviously you're getting some reward and it gives the end developer the choice of

do I want to start a new network from scratch and go through the entire process of doing that or do I want to be able to use an existing Network and I think the interesting thing about polka dot was like they had all these ideas there from the beginning like this idea that that the the main chain was like the layer zero for the pair of chains and they would share but the problem is I'd say the polka dot model economically was like too greedy it didn't welfare optimize for the whole network it literally welfare optimized for DOT holders and I think the weird thing about reaking is because eth already kind of had a working economic system adding this afterwards was like is more effective than trying to start from scratch with that if that makes sense and I would say like you know I think

again I think I think there is a lot that was learned from the lessons of of the pokka dot design and you know I mean reaking supps to prove itself but I do think you know when you get to this fundamental economic question of I have offchain services that are need to get matched to participants who will secure them how do you do that how do you do it efficiently what should the market structure look like and like you know I think we're finally at this point in crypto where like you know all of this stuff that people have talked about for matching markets it can happen like live and iteratively right now uh you know another thing about the history of matching markets like hid the exchange or you know the The Residency matching programs or you know all all the stuff

that sort of matching markets have historically been associated with is they're usually thought of as these one- shot games right there's like everyone who's graduating from med school in one year in the US gives their preferences and then the hospitals give their preferences and then they get matched right that's that's like but that's a one-hot thing like once a year every year is a different student so it's like a totally different thing but these crypto matching markets are very unique and that they're iterative and like constantly and evolving and like that's that's the thing that I think is very unique about them that even in normal Finance you don't see as much like sort of online advertising you kind of see this but even then it's not quite the same because the user doesn't have control over their matching the matching

is done effectively by the platform like the Google Facebook Amazon they're auto bidding on your behalf they're kind of doing most of the work you don't really can't change exactly how it works but a unique thing in crypto is that this is happening live on its own with billions of dollars and I think yeah and I think what's what even complicates that further is that it's like halfway in between being a batch system and a continuous streaming system in that we've discretized the world into blocks and there are weird consequences of that being a bashed process while at the same time there's a continuous stream of people having demand wow for use of a blockchain that's funny as you describe all this I I start to wonder how or if

me changes at all through this like in the reaking world does it open up new places where there's crazy Arbitrage have you looked into that anyone I think that definitely opens up there's a huge uh huge amount of thought on you know interoperability in general and consequences for that for me so there's the people working on shared sequencers and other stuff like that so definitely definitely there there definitely implications for it but yeah I I wait wait wait wait I I would have thought you'd have more of an answer because I feel like your the broker mechanism stuff for resource pricing is kind of thinking about that so there there's there's the connection it's that's I that's less tied to reaking but yeah I mean meev in terms of interupt between

different parties you can talk about maybe different types of meev there's uh M that comes more from Arbitrage of lag times between uh different parties that have different information and I think this is definitely a thing that that is relevant um maybe harder to study theoretically and then there's also me that comes from just having a richer ecosystem to work with and M is often thought of as a pure net negative but uh you could even think about me in a different different light as uh it's a type of uh just in time just in time coordination that's going on um currently all of these types of coordination that's that's happening

because most of what's happening in crypto is a zero sum game it's just extracting stuff from the parties from from other parties but uh in cases where uh you know there's there's not a zero sum game you could perhaps imagine that that this leads to some some net positive that's funny funny you just mentioned just in time because that is a late '90s early 2000s management term and I'm amazed to see it now once again that's cool navine do you have any like research in the works that you want to shill or share anything coming out sure yeah yeah um the reaking stuff is already out um but also have been working on this resource pricing related work uh that we very lately touched on

that's probably going to come out uh in the next I don't know month or so if I were to guess cool well thank you so much for coming on the show sharing with us sort of your early research and how it led to you working on this reaking stuff and then going deep on the reaking stuff and letting me ask a lot of questions about reaking that I've always wanted to ask and it's this has really helped me to understand much much more how you're thinking about it but also how one can think about it a little differently than maybe it's initially been described so thanks just ignore the the the the 5,000 points for subscribing to my cuz like that stuff is not the the point of the mechanism yeah yeah yeah thank thanks so much for having me it's been it's been awesome to be here hey

thanks thanks for for coming on I uh I'm happy that hopefully we can we can dispel the myths about reaking uh slowly over time yeah okay I want to say thank you to the podcast team Rachel Henrik and Tanya and to our listeners thanks for listening