190 - Sergey Nazarov on Chainlink’s CCIP

we want the whole world to run on blockchains which means the whole world's value has to go into blockchains which means Banks asset managers Sovereign wealth funds family offices everybody has to be connected to and utilizing blockchains for their favorite flavor of financial products so this is you know about as big an opportunity as I can imagine for banks because they're going to be able to generate huge amounts of assets put them on chain and

give them to a global market as long as that Global Market can connect to their chain which is by the way what ccip does welcome to bank list where we explore the frontier of Internet money and internet Finance this is how to get started how to get better how to front run the opportunity this is Ryan Sean Adams I'm here with David Hoffman and we're here to help you become more bankless bringing Banks to blockchains how do banks and crypto networks communicate and why why should they what's the prize available for both of

them we brought on Sergey nazarov from chain link to help answer that question a few takeaways for you in today's episode number one why Banks and crypto live in separate worlds and how to bridge them number two we talk about Banker Tech what is Swift what's the dtcc have you even heard of that acronym number three what do decentralized Oracle networks like chain link have to do with any of this number four why uniting banks and blockchains can produce a $100 trillion Market maybe higher why what Sergey calls

securitization 2.0 number five what makes a decentralized network Sergey has five levels of security that you need to hear and number six will crypto happen fast or will it happen slow of course you're listening to bank list because you think crypto is inevitable but how many years do we have to watch it all play out David why was this episode significant to you during this episode listening to Sergey I had the old bankless thesis the protocol sync thesis in the back of my mind briyan and I

think he was articulating um a version of that when we asked about why there is this incentive for banks to bridge their assets into the world of crypto into the world of blockchains uh Sergey has this Vision that as soon as these two systems can talk to each other and of course he thinks that ccip the new protocol from chain link is the way that these two systems talk once that happens all of a sudden there's just this gravitation flow of value onchain where Banks go go from being offchain to onchain and to me

that's the protocol synesis playing out uh blockchains are these new lower levels uh lower lower ledgers than Banks they're a ledger deeper than the banks deeper than the Central Bank even uh and all of a sudden the value of whatever Banks hold on their ledgers can be expressed on chain uh and that is the protocol thing thesis eventually credibly neutral decentralized larger networks will eat up all the smaller ones uh and banks are the smaller ledgers and ethereum and all this layer twos are the larger ledgers uh and we

just need some sort of path pipe to bridge these two forms of communication of uh communicating states around ledgers uh that's the entire thing I was thinking about while while Sergey was going on but I didn't dare bring up the protocol syn thesis because I'm not sure he's familiar with it uh but we'll have to talk about that in the debrief yeah I agree there's very much a trojan horse uh going on we're getting them hooked on block I think there's a ton of wins for um traditional Finance once they realize this and I think Serge's thesis is all we're missing is the glue how do you

connect Banker Tech and blockchain so that is the source for today's discussion yeah and David we have so much more to unpack during the debrief this is our first chain link episode ever which is uh somewhat interesting and there's maybe a backstory there we'll save that for the debrief though and if you are a bankless citizen that is available for you now on the bankless premium feed some disclosures before we get into this episode nothing to disclose neither David nor myself hold any assets that are relevant to the show of course we are long-term investors we're not journalists we don't do paid content there's always a link to all bankless disclosures in the show notes

at bank list.com disclosures we're going to get right to the conversation with Sergey but before we do we want to thank the sponsors that made this episode possible including bankless Nation I am super excited to introduce to you Sergey nazarov the co-founder of train Link train link as you know uh provides all the prices in for defi providing defi with the data that it needs to know what prices are but really chain link is much more than just prices for defi chain link is aiming to become the interoperability layer between all of trafi and defi the Gateway between

offchain and onchain data and we are going to explore that vision of chain link today with Sergey the co-founder here today on Bank list Sergey welcome to bank list thank you for having me great to be here so I started off this conversation talking about I think what everyone understands chain link to be inside of the crypto context uh it Powers a lot of the information that many dii apps need to work even at all like we need to know the state of prices for many many defi apps to even function but I think if we just focus on that conversation we would really be selling

the chain link Vision kind of short uh the chain link vision is much more than just prices for defi or even just data for dii maybe you could start to um open up the conversation of where chain link what the ultimate vision for chain link and why it's created such a powerful movement behind it and what what it's really looking to do what is the the chain link Vision sure so Oracle networks are a environment that does computation in a trust minimized way about everything that you want to make trust minimized or have consensus about

that is outside of a blockchain right so blockchains generate this unique computational result of this decentralized consensus where multiple independent nodes agree um you basically do the same computation and the commutations all match up up so they arrive at consensus on on what is basically a trust minimized computation because no one can manipulate that computation no one can change it no one can modify it as long as it's done according to the protocol of that chain but blockchains only come to consensus

about three things private key signatures token ledgers and state machines so these are the three things that blockchains arrive at consensus about they they this like the universe of of of what they're able to do and this is why blockchains for so long have been focused on tokenization because that's a native capability there is however a lot of other things outside of blockchains that would benefit from consensus proving things about them and

this extends to random numbers various computations any and all data communication and connection between chains and this is what Oracle networks do so all Oracle networks don't have a chain they they they don't make a chain chain link doesn't have a chain it is an oracle Network kind of framework that has generated over a thousand Oracle networks live on production securing value and processing over $8.5 trillion dollars worth of transaction value and the way it's done

that is every distinct problem that you have you can generate a unique Oracle Network to solve that problem so for example you need Market data about a certain price of a certain asset or you need a random number computation or you need um some connection to um another chain in the chain link world what you do is you make an oracle Network that generates a service basically a a single

um specifically focused service that's made up of multiple independent nodes with multiple independent keyh holders that are following the protocol of that service if it's a data service then the Oracle network is Computing the aggregation of data from multiple different places and making that data in chain Link's case resistant to various manipulations such as flash loan attacks and others and this is why chain link data is so widely used it's because you can generate Oracle networks that

reliably prove things about Market data but Oracle networks extend to basically every computation outside of a blockchain that you would want to make trust minimized through the use of consensus and so that's what Shan link has now expanded to do both in terms of going Beyond Market data to be the largest provider of proof of reserves the largest provider of various other types of data and then moving on to various um computations for gaming

automation uh very soon we're releasing uh into mainnet something called functions which is going to have a lot of very critical capabilities for people to access various external systems um and most recently what has gone live is the cross chain system which once again uses Oracle networks as basically a validator set to compute cross-chain communication and that communication includes both the movement of value and the movement of messages basically data and then the very important thing that

ccip can do is it can combine the value with the messages so that when you send value you can tell it what to do when it reaches its destination which means you don't need to be at the destination with your own wallet or your own address you can just instruct the system that you want the system to do something with your value when it reaches the destination which is you know just a an advanced feature that Oracle networks once again do computations around so um I think this is the thing that isn't

super clear to people because the main use case of oracles for the past couple of years has been um defi Market data basically because Defi boomed and so the Oracle network computing method that you know chain link invented and originated is something that um was very widely used for that but actually it's it's very widely usable for everything else that you need to do a computation on that you need to prove with

consensus and and that extends both to all data all computation that's offchain so computation that chains aren't able to do and now all cross-chain Communications and the important thing is that this all uses the same chain link um kind of Oracle network security model that has already secured so much value and that's why people are also prone to to rely on it because it has been it has St stood the test of time for over four years on production and the amount of value transacted and the

amount of protocols secured but it it isn't it isn't just about Market it's kind of like you know AWS was at some point about S3 but then it had a lot of other things that cloud computing could do that's kind of where chain link is now where it originated by doing the computations around Market data because there was a lot of demand for that but now it has expanded far beyond that um to many other types of computations as we progress in this conversation I I want to fill in that gap between where train link has been

where it's like frequently known to be and and the Very expansive vision for itself which is almost like Limitless I'd say um the starting with train link and just the price oracles for Defi and then expanding to just like General offchain computation well offchain computation is like everything that's that's almost anything it's a very open-ended use case and I'm wondering if you could try to just shed a little bit more light on just like when you talk about offchain computation well what what what does

that mean like is that how can we Define that a little bit better sure so what I mean technically speaking is trust minimized offchain computation so it doesn't mean every single computation that you don't have any trust Dynamics around so if you want to run a private GPU for yourself to play a game you don't need trust minimized offchain computation just like you you don't need a blockchain to run a GPU for you to play a game on your personal computer right so you don't need there's no trust issues there's no value transfer there's no big problem

um what I mean by by offchain trust minimized offchain computation is basically the spectrum of all computations outside of a blockchain that need to be made trust minimized to create what we call the verifiable web so I'll give you an example if you have a defi protocol and the defi protocol ingests um defi Market data basically well not def5 Market data like Market Market data crypto Market data from a chain link Network that isn't where the

di5 protocol's interaction with offchain computation ends right that's kind of actually where it begins then you need to automate the protocol right because the protocol doesn't know a lot about doesn't know what time it is there's various conditions that need to be triggered that the protocol doesn't know about itself so then you have chain link Automation and then the protocol let's let's say the protocol has some asset that needs um to be that want that they want to be purchased by a bank and the bank needs the asset to comply with

identity data so now you need another piece of data added to the asset right and then let's say you took the asset you made it you in injected proof of reserves into it you proved the price of of the asset that the that's the reserves are proven for you've added the data about identity so the asset can be sold to the counterparty that requires it to have identity data so they know who they're buying it from because they can't buy it otherwise and now now you need to move it to their chain because that's where they that's where they want it when they buy it right so they send

you a stable coin using ccip that transaction triggers the purchase of the token we just talked about from the defi protocol and sends it over ccip to them so this is just some examples of the trust minimized computation that can't be done in a single server by the creator of the asset just like the market data can't be fed by them the proof of reserves can't be fed by them the automation can't be done by them the cross chain can't be done by them more importantly they don't want to do it because it's an extremely complex and

security sensitive process which takes you know tens of millions of dollars to create any one of these services and then millions of dollars to audit it and and huge amounts of resources from nodes to run it right so generally application creators don't want to create infrastructure they want to create applications with the help of good infrastructure so if we if we just go to defi already today on production multiple defi protocols use multiple separate Oracle networks for distinct purposes for providing different types of data for providing automation now for

providing cross-chain connectivity so that's a very concrete example in today's world in the future World um you can see a lot of trust minimized offchain computation creating the verifiable web that the real question is what are the computations that people need to make trust minimized in order to create the verifiable web which is what we call this kind of vision of the internet that we're all excited about and how are they going to do those computations if a blockchain doesn't do

it right a blockchain will allow private key signatures it'll allow tokenization it'll allow various conditions around that tokenization such as dexes Dow votes from tokens um conditional statements about what would happen if an oracle Network were to provide a with data or to create automation for it or to do something else right or some other system but um the the the real kind of point is is

that trust minimization is a spectrum right in defi we have hyper sensitivity to trust minimization because there's a lot of you know distrustful paranoid smart people that understand how the world actually works in this part of the world then there are people that weren't um trustful until Silicon Valley Bank right and now we've kind of converted that more and they're like okay I understand what you were saying you know private keys are good and and so the the

the real uh thing that that's going to happen in my opinion is that more of the world is going to realize that a verifiable web and Trust minimization are what they want the reason they don't realize it now is cu they don't they don't have the option right like if stable coins were uous ly available if they had remained stable which they unfortunately didn't remain that stable during svb so if they were ubiquitous if they were easy to get and if they were stable I think that many people would

have migrated onto stable coins as a store of value during svb right because it has all these Superior properties and it has this trust minimization about actual control of the asset so I think that we are really in a in a kind of world where more and more um people will move towards the verifiable web I also think that AI will accelerate that big time because the amount of distrustful um bad dishonest information

and and stuff that will bombard people will increase and so computations uh through blockchains and computations through Oracle networks will become more valuable in that world does does does that all make sense does that answer your question or or not really no Sergey that that's really helpful and actually in in getting into this episode I don't think I was ready for kind of a rabbit hole uh to go down but but it strikes me that I don't think bankless is actually gone down the Oracle Network rabbit hole and uh I'm kind of prepared to in this episode that sounds like a rabbit hole

we should go down so I want to set like some overarching you know framing for this and and maybe zoom out right so we live on this um you know internet with a lot of unverified computation right so we don't know what exactly to trust and I think in kind of the traditional world uh most people don't see a need for verifiable computation because they already have trust networks set up and most of these traditional trust networks take the form of like reputation you

know it's coming from XYZ institution who has a reputation and I can trust it right so that's one trust Network that the trafi World Turns to another kind of uh source of trust is probably the legal framework the legal environment right I can trust this because um if this information Source because if they do something wrong they'll be breaking the law and they'll be fined and there'll be penalties and somebody will go to jail and in order to avoid that they they will not do that so therefore I can

trust this Source um I think what we have been focused on primarily at at bank list and in crypto so far has been really a a more trustless or let's say trust minimized onchain data set which is uh the blockchain essentially um but what you're saying is you're making the argument that blockchain data is is relatively it's very powerful uh and it's incredible and it can trust minimize a lot of things but its aperture is somewhat narrow in that we could do stuff with private Keys we can

do stuff with with tokenization and this there's this whole world of other data of other compute that also needs to be made more trustless on the Spectrum and you're saying that's kind of the void that Oracle networks really fill and so I think my main exposure to chain link going into this episode was a subset of that offchain data that becomes more trust minimized which is price Oracle data right so thank you very much chain

link poers so much of Defi and so we know the price of ethon chain thanks to chain link oracles and what you're saying is that's only the tip of the iceberg there's a whole bunch of other data sets that the world needs that the internet needs that need to be digitized and made trust minimized that where where it's not a good use case for the legal system or kind of the the reputational trust system to maintain and you're saying that's what Oracle networks are here for and that's what

you're executing against is that is that like a a useful regurgitation in in kind of that you know 30,000 foot view of what Shan link is is here to do I I think that is useful thank you um yeah that's that's that's a good that's a good summary with with a few small um caveats the one of the caveats is it's not just data it's offchain computation whether it's related to data or not including communication between chains so all trust minimize computation and

data aggregation is a subset of trust minimized offchain computation that has simply gotten significant adoption from the trust minimization sensitive defi user base and it would actually makes sense that the people in this industry in the blockchain industry are very sensitive to trust minimization because that's the unique defining property of the products in this industry so it makes sense that something like this would find its adoption in a world like blockchains so that's the first caveat

the second caveat is that Oracle networks greatly increase the use of blockchains and blockchains increase the use of Oracle networks because um as you start inter acting with Oracle networks you you'll find that a lot of the um data storage settlement value transfer aspects of your relationship with counterparties peers with using your system um you know institutions on your platform whatever it might be that a blockchain has an

important role to play so not only can Oracle networks feed information into blockchains and enable them to do much more in the form of defi game decentralized Insurance um you know huge number of other things but they can also connect existing systems to use blockchains so they they are a conduit through which existing infrastructures existing web 2 systems existing Bank systems and this is what we've been

doing with Swift and many big market infrastructures and many big Banks so it's a bidirectional relationship and the bidirectional relationship creates greater trust minimization leading to a verifiable web that is immune to various manipulations of you know keeping people's money reneging on contractual Arrangements being subject to AI manipulation basically manipulation resistant web that's what verifiable web is you want to be manipulated you want to take the chance the bank's not going

to give you your money you want to take the chance the insurance company's not going to pay you okay you can stay on the manipulatable web and chake your Chann your alternative now that we're working on and that I think our industry is working on whether or not it talks about it um is a verifiable web where you can verify that your assets are within your control at all times that the insurance company has no choice but to pay you and and to your point about the legal system it absolutely is about not leading the

legal system so I'll give you a fantastic example my favorite example of this which has nothing to do with tokenization it has to do with crop insurance insur so there are places in the world that will experience drought experience drought will continue to experience various exogenous weather events unrelated to anyone doing anything wrong in that place and that will affect various farmers and those Farmers have no way to get insurance basically because their local legal system does not allow Insurance to exist doesn't allow it to exist no insurance company can generate Insurance in the

legal Frameworks available to that country basically yet these people need Ure ins they have an Internet connected Android device with the help of stable coins with the help of smart contracts with the help of Oracle networks you can give them a completely parallel system of contracts the Oracle networks verify the weather triggering the smart contract the smart contract executes the state change which is the value transfer method the settlement method to take the stable coin from the capital reserves of

the insurance protocol and give it to the farm armor this is what I think will change in the world where there will be this massive leap frogging in emerging markets in developed markets um it'll be different because there's a huge amount of problems and and you're right that people don't realize that these problems exist but I can promise you they realize them when they go wrong and it's it's in those moments that our industry um I think shines and will

continue to shine and un based on the world the way the world is built today it's going to continue to go wrong quite a bit um which is both sad but in a certain sense it's an opportunity to create a better verifiable web and a different way for the world to work if uh if that makes sense understanding how this works uh technically uh I I'd like to go into that a little bit um in the the the base iteration of of chain link which is just like provide price Oracle to defi is a pretty simple concept to understand

there's just a smart cont contract that provides an output and then other apps consume that as an output and that's that's how it works when we expand chain links vision for just generalized offchain uh computation that we want to prod create trust minimization I start to lose my ability to actually understand how this works technically because there's there's so many different kinds of offchain computation that we would like to trust minimize how do we how do we have like a a nearly infinite spectrum of types of things

that we would like to come to consensus on computation that we would love to uh come to consensus on how do we how do we like make that into a standard how does that actually become trust minimized when there's so many different types of data out there how do we actually like make this into a consumable um system for applications and people just to like uh organize around does my question make sense yeah your question totally makes sense so I I would just clarify what it is that chain link does and and and what it makes possible um and and even to

your point about consuming the price data I I'll just use that as an example to show that it's not exactly as as simple as as that so so basically you need to create a node Network you need to create a validator set in the case of price data you need that validator set to be made up of high quality nodes with reputation that proves that they are high quality and reliable then you need that validator set to connect to various data sources for which you also have a reputation so what chain link does is it not only allows you to combine various

nodes into a validator set that you then Define as doing a specific computation from a specific set of data sources but it also allows you to Define how that validator set will manage various risks so I'll give you an example when FTX happened there were a number of data providers that were waited toward FTX as a price source so not only were the Oracle networks within chain chain link um operating you know correctly but the

chain link system and the chain link Network and the way the whole system works basically ended up seeing that there was um a skewed result coming from a few data providers that were heavily weighted towards FTX those data providers were removed and replaced with data providers that were not weighted towards FTX so Oracle networks are actually about um a few things they're about combining nodes into committees into validator sets that do very specific singular computations singular so a

single Oracle network will generate a single aggregation of a single price pair right and that's what it'll do and the chain link Network and framework it it it has launched over a thousand Oracle networks on production that have secured value successfully for over four years processing 8.5 trillion in transaction the majority of that for Market data so so Sergey uh chain link is an oracle it it is a network of Oracle networks it's not just one one sing abely not I we do not have a single Network that

will not scale and that's not secure and the the people that build single you know I I I hope we'll have time to talk about the five levels of cross-chain security while I'll explain the different types of networks you can build but but no chain link is thou is over a thousand Oracle networks and if you count the staging the test net some of the private instances it's thousands of Oracle networks and each Oracle network has a specific specific function has a specific set of inputs and a specific set of computation inputs and a specific set of computation

outputs it's a it's a way to create focused validator sets that are focused on generating a specific you could call it a decentralized service so just like you have centralized web services chain link is the way to create decentralized services and each service is very specific and and then what you can do is as the value secured by that service increases you can increase the amount of nodes in that service you can increase the amount of data providers the variety

of data corrected to that service you can increase the cost that the node provider is willing to Bear to broadcast transactions for example the chain link Network when ethereum gas went to over 2,500 was one of the few things that continued to broadcast without missing any updates and and that's because each of those Services were properly incentivized and because there was basically a focus for each service to do that so so each service can be um

configured with the amount of noes with the amount of data with the conditions that you it has to meet and if the service doesn't meet those conditions it s suffers reputational damage if the node doesn't respond It suffers reputational damage that makes it less likely to be included so it's not about um it's not about a single Network and chain link doesn't have a chain and chain link isn't us forking some chain and changing some configurations chain link is a completely separate method of generating consensus that we put