Navigating the Complexities of Restaking in Web3 - Marijo Radman | Alterscope

hello everyone um I believe there are not so many topics at the moment in the web three space who manag to get so much hype as uh risk taking it started already last year when uh senar was RA raising his first financing round they came questions what does this mean for ethereum what consequences should we expect risk taking could basically unlock unlimited potentials for the ethereum ecosystem yet there are also risks involved and basically I believe until now no one really understand what kind of risk do we have and what kind of consequences should we expect there are many even who believe that the next blow up in crypto is going to come from risk taking um the whole discussions are evolving around risk taking is a ticking bomb up to risks are over estimated in the risk-taking space and basically everything culminated that even vitalic bin um went out and wrote guys don't touch the consensus mechanism um for materium I will try to give you here a rough overview what kind of risk do we have in RIS taking what are the PO potentials what could happen in the space why it's so relevant and what possible mitig ation strategies could be used that we do not have a similar event like teruna before that let me introduce myself I'm the co-founder of alter scope uh I'm the space for the last eight years uh mostly working in the defi space before that spent more than 10 years in Corporal corporate development settings alter scope is a startup from Munich with uh offices in Zagreb and London we are more than um people uh company and mostly focused on the risk side um so basically we are building a decentralized risk Market enabling everyone to build verified uh risk models in simple terms anyone who works in the thread f space they know about Black Rock Elin and that's the vision what we want to aave we want to build a base infrastructure four protocols risk taking protocols um which can be utilized to make them more safe and efficient U our network is already like the first phase is already live we slowly increasing our um traction working with major protocols in D5 space but also on the institutional side so just few words about alter scope and why actually I have some expertise around resticking so what is actually the main problem of what is what is actually I layer doing before we go into the risks of it so igir is nothing complicated just few smart contracts who take the ethereum which is used for staking of U staking uh purposes on ethereum side from the validators and reuse it or reuse it for other means so basically ethereum has at the moment something like 1 million validators um billions in ethereum is stake there and San the founder of IG was asking himself can we use this ethereum also for some other purposes the main issue currently we have in the space whoever wants to build a chain custom chain protocol whatever the hardest problem is actually how to decentralize it um it costs the costs are very high to build the whole infrastructure the validator network and so on and basically if we could use the ethereum validation Network this could slow solve the problem of decentralization casy so that's the high level idea of it so basically we have the validators as I mentioned roughly m one million of them are currently valid in ethereum those are big uh companies like coinbase Cloud figment corus one and so on they're taking their ethereum put it in the igen layer smart contract and from igen layer smart contracts um there are so-called actively validated Services who are using that so what could be POS possible use cases uh of these actively validated Services simple terms ZK coprocessors data layers rollups Bridges um oracles and so on so we can with very few clicks from the actively valed Services point of view get um billions in Economic Security um of their um underlying protocol which is kind of amazing uh um before that when we just look chain link like no one actually managed to disrupt or create real competition to chain link and when you look what chain link is doing actually it's actually quite a simple idea the main problem is like how to build a network who is decentralized enough to make this working and I for example like building a Oracle on IG era could be possibly a solution for that um on top of that we have so-called liquid risk taking protocols um and they are actually simplifying the whole process of bringing this ethereum from the validators to the right actively valid service so the issue is like these validators they are nothing else than a hardware company who is running these nodes and they really don't want to deal with the the issue like which activ reled service to secure which one of them is safe and so on and basically those services like ethery Renzo swell and so on say to them hey guys give me your ethereum we will do we will do the jobs for you um then they take the ethereum as a collateral which they have create their own stable coin and bring it to defi so basically they manage then to get additional yield in the defi space on a compound and all the other protocols and not only that the ethereum from the validators is actually transferred to D5 which is amazing they can additionally create um some yields for those validators to validate these actively validated Services um of course these uh these validators are not going to do this kind of validation for free there has to be some kind of reward for that uh and the RVs is usually paid that in ethereum or their underlying token that's the whole idea it's actually amazing so basically we can really scale all these new middleware related applications in very uh few simple clicks and make them truly decentralized what is the issue here um all of them are working in their own environment like and this environment is dynamic all of them have there underlying risks and those risks are interconnected when we look just the validators um I said 1 million validators is on ethereum so basically if few of them blow up actually it's not a issue at all but that's far from from the reality uh 95% of ethereum which is used for validation on ethereum is coming from 1,500 operators and roughly top top 10 of them have something like 80% of the ethereum network so basically in simple terms if coinbase Cloud would do some kind of shitty job we could lose 200,000 Val leaders just with one ofs this is the issue so basically there are 10 validators basically who are validating the whole ethereum network and they have now to take this risk management really seriously um they have to think about diversification they have to understand how they are going to uh allocate their ethereum not to blow up ethereum only but like the whole ecosystem the issue here is that when we look at the possible risks so basically we have on ethereum side the slashy mechanism uh and now these actively valuator Services they are saying okay give me your ethereum you should behave according to my rules and I will create slashing mechanism from my side that you follow my rules and this is the issue so basically every single of these Bridges oracles whatever I mentioned uh have their own underlying logic if they make some kind of mistake they could could possibly in worst case slash all the ethereum which is used for for securing them and this could have consequences then for the whole ecosystem actually it could have major uh consequences also for ethereum uh because for validating ethereum you need 32 ethereums so basically imagine you are a validator and you got slashed you fall below this 32 ethereum and so basically that means immediately Security on ethereum side is lost so that basically means those validators have to behave like hedge funds crypto managers they have to decide and they have to build Advanced risk management capabilities to secure um those avss now we have all these lrts like Renos uh eery and so on they are taking this ethereum as I mentioned creating their own stable coin and bring it to defi the issue here is this ethereum is used as collateral and we have again these slashing mechanisms if some of those AVS is slashes some ethereum so basically that means collateral is lost on the LRT side so basically we have a very volatile stable coin and this stable coin is used in defi space and we have if you followed my talk from last year at e belgrad you know that def is also a composable Tex St this stable coin is all over those protocols from Pendle up to gearbox up to a compound and so on Moro and if there is a deeg possibly we could have consequences in defi space but also risk taking space actually it could be like also that the risk taking space totally fine but we have some kind of liquidation wave happening on defi side out of reasons similar what we have so far like teruna very good example and this could create a very huge cascading wave which could go back to race taking space and this is the problem so basically the complexity in the whole Space is magnified to a new level um I mentioned the lrts I mentioned the validators I didn't even mention I layer um there are smart contract security related um concerns so basically if there is a exploit of the smart contract I don't want to even think what could possibly happen at the moment I has roughly 12 I believe believe ion tvl and I really can foresee that something like 120 billion could go in rcking space I layer provides this e economic uh Securities uh shared security layer but they have also in plan it's not yet long launched to bring their own insurance of attributed security that basically means providing security for avss and you all know that basically SEC uh insurances do not really work work in defi space so far and if there is some kind of event happening what those kind of cons consequences could have then overall for ioner and the whole Rising space I believe the founder in itself is not sure how is he's going to manage those kind of risks um and besides that lastly we have avss active lity Services I mentioned the slashing mechanisms which should go live in the in the next to six to n months we have another issue here the question is how much security do they need imagine you're building a bridge what is the right Economic Security uh amount of ferium for your uh use Keys what is the cost of corruption what is the cost of profit all these questions are coming now on the AVS side and the it's getting even more funny like usually those avss are all new projects like they are bootstrapping they don't have like billions in ethereum to to pay these valers they have the underlying token imagine some kind of alter scope token if you are active service and you are paying actually with your token and together with ethereum for your security and these validators now have to think also additionally about this token like does it make any sense is it stable like are there any kind of St staking mechanisms does this all stuff makes sense in three days should I sell it immediately is this actively valid service actually liquid enough to pay me for the security and all these questions are actually issues which risk companies are at the moment uh dealing with starting from Gauntlet C SLB Al alter scope I believe I believe basically every single company is at the moment which is dealing with risk involved in one or other uh way with risk taking now we understand the problem we understand the connection to defi we understand the POS possibility of a blowup what this could lead to um besides that I didn't even mention that these active levor services are all all interconnected like if you are building a co-processor usually it needs also a data availability layer and there is also a bridge and Oracle involved so basically in this Cascade just one of them has to break and all the other ones are going to be affected it could be the case that basically one slashing mechanism could be activated and then four other ones are because of that activated like those kind of possibilities are all existing or coming to us with r taking the question is like how to deal with that we know from defi space so far how these Landing protocols were dealing with risk management as a good example there are companies who are dealing with security auditing real-time monitoring and there are economic security companies like Gauntlet and cow laabs setting up the economic parameters every few weeks and we still had all these blowups and the main reason is of course you cannot really react if something happens in a second so basically some kind of liquidation is activated because there is a exploit something was hacked you are not able to change the parameters but not only that like even if you see some kind of exploit is happening and you have a direct exposure to this exploit the protocol cannot do anything it's a static smart contract written it doesn't have kind any kind of input data to react on that side and that's basically where we as alter scope come in I believe we are one of the first one who build a real realtime risk infrastructure so basically building verified models who are updated in real time ensuring that the data and the models which are used for that are not uh exploited so basically we are using the ZK magic be behind that I will not go in the complexity of that uh inside and helping this protocol to react in real time um we launched not only that these models like the question is like who should build this kind of models even the alter scope research team started to post constantly new models new possibilities of the risk Marketplace we are inviting everyone to build those kind of models so far for far in the first phase we enable non-verified models and soon the verification will come um we made it intentionally easy to create those kind of models so basically there is a no code risk uh modeling engine and the reason is actually very simple we learned very early that researchers analysts they are not really the best coders in this world and developers are not really good Risk Managers and to solve this issue so basically we created a no code risk engine to enable creation of statistical model models in a very simple few clicks way um or some kind of linear regression models yet there is going to be also a scripting um part where more advanced models can be created on that side uh besides that like data which is used for the purposes of those models we are focusing more on the fundamental data like security liquidity related data we are also inviting third party companies who have amazing data very good example could be like for example block analytica or cyers with whom we are collaborating to provide those kind of data to make those model more um Advanced and the use case um is basically making defi smart data driven uh not only defi but also risk taking protocols simple cases could be like automatic rebalancing of security of those kind of U abss so basically calculating in real time the cost of uh of corruption cost of profit and adjusting the the value Network in accordance to that automatic slashing mechanisms unverified data sets and models but also more advanced D5 models which we do do not have yet at the moment um life and basically coming to the end of the presentation we at alter SC believe that we are at the moment at the merger between the thread and defi space we are living in a world of financialization of everything it starts with tokenization so basically all value is somehow um bought onchain we believe that the complexity and the risks in the space are going to be exponentially going up due to the modularity of the TX Tech from the def5 space actually from the whole web3 Space Race taking is going to be a major risk related topic due to the amount of liquidity which is going to go through risk taking so basically it's almost creates competition to Lio um we believe institutional demand for D5 will come over the years they are now at the moment investing in Bitcoin ethereum and the major Blue Chip tokens but at one point they will also start providing liquidity it started already to happen through hedge fund and crypto asset managers and at one point these bigger names are going to do that also and lastly We Believe for Defi and restake or general web three space to manage to get mainstream adoption we have to move from those statistical uh statical smart contracts we have to have Dynamic smart contracts in place uh data driven smart cont in place and we at Al scope are going to enable that um that would be from my side if you want to follow what we are doing go on our website you can immediately play play with our risk engine or just join our telegram group uh if you want to follow the updates or if you have any kind of questions thank you any questions from your side yes hello so the reaking and the liquid staking isn't this technically leverage leveraging the whole Market I mean if you look from the staking side of ethereum and you take risk taking in in a account it is so basically you're using using the same asset for two purposes and the consequences of one uh use case and the other use case has the almost same underlying uh cons consequence for the market um you can see it from that base level and then on top of that you have all these stable coins uh coming from R taking so it's additional leverage coming to the space so it's it is a leverage in place but I not compare it to the classical uh defi related trading protocols it's a little bit different there is for sure leverage in the space are there some numbers how much leverage there is if you look it that way I believe it's not a lot of Leverage at the moment because AVS is who are active they don't have any kind of slashing mechanism in place um they are taking some kind of security numbers they are not able to calculate how much security they need so basically take billions because they are not many of them and they are not paying much for that the main issue is actually the stable coin which could be like deack we had a very similar actually deeg event already happening on the Renzo side due to air drop reasons but slashing is not active so basically this stable coin is not really that uh volatile as it's going to be over time so I would say like even I cannot quantify it I would say like The Leverage is not that high but it's going to go up with time that's for sure how to calculate that uh this is going to be a very very interesting exercise so basically you have to take the ethereum validator side take the risk in account you have to take some kind of ethereum uh I layer minimal risk possibility especially when the insurance comes up and then you have to take the avss the interconnection of the avss you have to create LRT then you have to look the exposure of the LRT in the defi space taking all this stuff together and then you will get some kind of number who is able to calculate that at the moment no one um there is not even data for that available so basically all these risk companies are struggling to get data so I believe my personal opinion we should expect actually two cases the first one slashing is going to be very conservative not many of those AES are going to activ activate slashing mechanisms very soon that's the first issue and the second one I guess we will have one or two blowups until people realize okay we have to rethink uh the risk and risk taking in general around it thank you yes we we plan to run uh we at alas plan to be a active levelor service and the reason for that actually is a different one um for risk modeling you need besides onchain data also offchain data and you have to ask yourself how are you going to validate this offchain data like even if we bring it to our infrastructure and we do the verification of the storage the computation and so on if we bring in the database then the goes out and we are going to use the validators adding additional services to them so basically the scrapers of data coming together they're getting the same data points and then on a consensus based mechanisms proving that the data is not manipulated to ensure whatever is written in the database it makes sense and for risk management really offchain data makes a lot of sense security data is offchain data you look social media data Twitter like all these Community these panics what we saw so far they are really mattering for the risk management side on that on that level and so basically the data Agents from our side we are working with companies like auton noas the data agents are going to be run on those validator Network uh can you explain the sentence that the contracts should be dynamic and not static like right now I mean it's very simple like you have a smart contract who executes some kind of logic on a predefined code level yet and this logic is always executed in the same way the problem is like maybe you have to change your strategy very good example is actually Landing protocols they have their economic parameters in place these economic parameters are constantly like hardcoded inside the protocol every few weeks depending on the exposure of the whole Market you do a lot of simulations on that side and the issue hereby is basically these Protocols are executing what they are doing what they are supposing to do yet the problem is like they are operating a dynamic environment that basically means if something happens so imagine a case what we had last year um Oiler hack happened 200 million were lost 200 million activated some kind of liquidation wave uh you are having exposure all over the whole ecosystem and maybe uh if you would have a intelligence in your protocol you could mitigate some part of it and how this possibly could work even if you are not directly affected with the oiler heck you have indirect indirect exposure and you can see already through models that some kind of changes that you can already rebalance um the economics of your um protocol so basically you can change the lending borrowing rates and reduce the amount of harm which could happen in worst case like there are also protocols who have this pause button you can utilize Al also these kind of things like these possibilities are going to be possible but for that you have to get verified models in place the issue and why we don't have that at the moment in place is that we actually at Al like imagine you're using alas scope verified models the reality is like we try to help those protocols to be smarter to reduce the amount of those kind of events yet we have quite High incentives to exploit it by ourself and the reason is very simple like imagine we are providing verified risk capabilities for a a has more than 10 billion tvl I would say like some of my developers in in my team have quite High incentives to mly manipulate something in in our infrastructure to exploit it one or the other way around and that's the reason why you have to verify everything from getting the data storing the data in computation of the data but when we lose some compability if will we lose some com com composability of the contracts if we the contracts will become more Dynamic I uh if we add more Dynamic to the contracts we will lose some composability between them right right of course I mean that's true the question is like I mean you should ask yourself how often should we have these kind of updates like I don't like those kind of things what we are doing here it costs like if especially if you're doing s car verification utilizing ZK technology you cannot verify every minute or every hour like those kind of things are going to come are going to be questions of the future that's sure for now those verification is going to happen every few hours every few days depending on the use cases so basically making the models Smarter with additional like triggering events uh in place what is going to happen over time once once you have this models who are constantly bringing updates I believe this is going to be a question for protocol developers uh research analysts and so on of those protocols how to make this working in any like like what makes sense um the current state doesn't make any sense like you cannot have a landing protocol who change its parameters every few weeks in a dynamic environment what we have at the moment on the other side you don't want to have every second updates it's insan Insanity what you get in place and what is the right rateof we have to figure out I guess there are no other questions thank you