Euler: The DeFi Super App - Michael Bentley. Ep. 615

Transcript

early lending and royal protocols were restricted just like the short tail of assets. So like ETH and uh USDC and Bitcoin. So we started building Oiler as like a an integration with Uniswap at the time actually to enable people to lend and borrow not just the short tail but also the long tail of assets. If you get liquidated on Oiler, the because it's a fair auction, the bonus tends to be basically the fair market rate. Like however much it should cost to to do the liquidation plus a little bit extra, the actual cost of it might be a few hundred, whereas the cost of it on on another platform might might be literally hundreds of thousands or millions of dollars. When Tap is draining a protocol, they a lot of the assets come back in terms of like USDC or USDT, right? and circle on Tether. If

uh if they get a request from law enforcement or they know that funds have clearly been stolen, they have the ability to put freezes on asset. A lot of attackers will just effectively autocon convert any stolen funds into uh ETH or an asset that's like more decentralized and more difficult to kind of freeze basically. So that's what this guy did. the recovery period where we were tracking down and negotiating the recovery. The the price of ETH rallied and so actually it was kind of like the attacker put on a long position on behalf of all of our users. Welcome to Epicenter the show which talks about the technologies, projects and people driving decentralization and the blockchain revolution. I'm Ryan Crane and today I'm speaking with Michael Bentley who's the CEO and co-founder of Oiler Oiler Labs Oiler Finance which is

a very innovative uh DeFi lending protocol. So I'm really excited to talk with Michael about that. So just before we get started, we'd like to share a few words from our sponsors this week. >> If you're looking to stake your crypto with confidence, look no further than Chorus One. More than 150,000 delegators including institutions like BitGo, Penta Capital, and Ledger Trust Chorus One with their assets. They support over 50 blockchains and are leaders in governance or networks like Cosmos, ensuring your stake is responsibly managed thanks to their advanced MEV research. You can also enjoy the highest staking rewards. You can stake directly from your preferred wallet, set up a white label note, restake your assets on

IG or Symbiotic, or use the SDK for multi-chain staking in your app. Learn more at chorus.1 and start staking today. >> Hey guys, I want to tell you about Nosis, a collective of builders creating real tools for real people on the open internet. Nosis has been around since 2015. In fact, it started as one of Ethereum's very first projects. And today, it's grown into a whole ecosystem designed to make open finance actually work for everyday people. At the center of it all is Nosis Chain. It's a lowcost, highly decentralized layer 1 that's compatible with Ethereum and secured by over 300,000 validators. So whether you're building a DAP, experimenting with DeFi, or working on autonomous agents, Nosis Chain gives you a solid neutral foundation to build on.

But Nosis is more than just infrastructure. It's also tools that people can actually use. Like Circles, for example, lets anyone issue their own digital currency through networks of trust, not banks. And then there's Metry. It's their smart contract wallet that makes it easy to access circles, manage group currencies, and even spend anywhere Visa is accepted thanks to their integration with Nosis Pay. All this is governed by Nosis DAO, where anyone can propose, vote, and help guide the network. And if you want to get involved, running a validator is super easy. All you need is one GNO and some basic hardware. To learn more and start building on the open internet, head to Nosis.io.

Nosis, building the open internet one block at a time. >> Cool. Thanks so much for coming on, Michael. I'm really excited and looking forward to this one. I'm curious, tell tell us maybe a little bit about yourself and how did you get into crypto first? >> Yeah. So, first first thanks for having me on. Uh yeah, looking forward to it. Uh how did I get involved? Well, uh I used to be uh before before this life, I used to be a research scientist. I used to be an academic. uh my specialism was uh evolutionary game theory. So I used to do uh like lots of population modeling of biological systems and how they change over time effectively do uh I was more on the math side but I used

to collaborate a lot with like computer scientists and biologists and other people. Um I started as a bit of a script crypto skeptic when I first heard about it in 2015. Didn't think it sound very cool. I just mistook it. Mistook Ethereum for like some kind of boring database like solution or something. I I just yeah I wasn't wasn't interested. But by 2017, like maybe late 2016, early 2017, like the the the markets had kind of got a bit frothy and uh I kind of got drawn in at that time and started um I think later that year I started building uh trading bots and things for some of the primitive decentralized exchanges. uh at that time this is something called IDEX uh that I used to to to trade on and ether delta I don't remember that one

these were like literal order books on Ethereum um so but quite slow and quite plunky but they were kind of like a good introduction to like markets and and how Ethereum works and stuff and then uh yeah stuck around as as things like deteriorated in terms of the market conditions over the next few years and got more interested in a bunch of other stuff uh in 2019 uh and early 2020 like DeFi was sort of starting to emerge as like this kind of force to be reckoned with. It was like unis swap and um and year and things like that were becoming popular and that's when I I got sucked in and started studying interest rates and and and things and so I I actually started Oiler off the back of doing a hackathon project where uh I'd created a novel interest rate setting mechanism

that was more decentralized for like less dependent on on third parties than than the ones that were being used on on compound at the time. What was the original vision for Oiler? >> So, we wanted to make something where you could lend and borrow any asset. Um, a lot of the early lending and borrowing protocols were restricted to to just like the short tail of of assets. So, like ETH and uh USDC and Bitcoin uh and maybe a handful of others. And uh yeah, I guess uh I guess in my hackathon I was like I was sort of inspired by like how how do you set interest rates? If you want to expand that set to like everything,

you need an interest rate model that's that can adapt over time and and adapt itself. You don't want necessarily like a like a slow process of governance or third party like changing interest rate models. So we wanted to see yeah could we could we could we let lending happen for anything and so we started building Oiler as like a an an integration with unis swap at the time actually. So unis swap v2 had oracles for an inbuilt oracle for the value of assets. We plan to use that to enable people to lend and borrow not just the short tail but also the long tail of assets. Uh create create lending borrowing markets for all of all things. Um didn't actually work out like that. We had to pivot pivot a few times as is the punch route, but

that that's that was one of the early visions, I think. Yeah. >> Okay. Okay. And then I'm curious, is this it sounds like something where maybe your evolutionary game theory uh was that background knowledge very applicable here? >> Oh, I mean massively, right? It's um yeah I mean in that day job you often like what was what was I actually doing you know how do you build these models you you like consider like a set of things or entities people biological organisms and they're like kind of playing strategies in this big like competitive game and they interact a lot

and so you have to like model that out and think you know what what would be the the like uh equilibrium state of uh if if you've got like two different entities playing different strategies And that's very similar to what happens in markets, right? You have like lenders and borrowers. You have like uh people striving to like optimize profitability or optimize some kind of uh metric like riskadjusted return or whatever. And so um yeah there's a huge amount of overlap in terms of designing like a D5 protocol that can facilitate like markets and designing a model about how how uh yeah populations change in response to their environment. It's it's yeah if you just abstract everything then it's it's

actually very very similar kind of uh task. So yeah it was yeah definitely my background was uh very relevant to what I do now. So I I imagine you're pretty unique among DeFi protocol founders having like that kind of perspective and background. Do you think that, you know, did that just help you maybe design or learn a little bit, you know, have better models for different risk events or do you think it also ended up, you know, just in in a different protocol design uh because of that background we had?

I think um I think yeah I think it's uh I mean yeah lot I mean D5 founders tend to come from really varied backgrounds right and and there's like trade-offs to having different to to all of those backgrounds you know like I think famously at Oiler we're very like uh engineering driven a lot of the the things that we build are like really really strong on the engineering front and well like uh well architected and uh so we have like very strong attention to detail to like low-level processes and mechanisms and things. So like for instance liquidations on oil I think we have the best liquidation engine in in all of DeFi. A lot of uh a lot of liquidations when when they happen on other lending borrowing protocols

um effectively the borrower just has like some of their collateral like slashed and sacrificed and then that's used as a reward for the um for the people that that performing the liquidations and it's often really costly really really costly especially if you're a big borrower. you've got like, you know, tens of millions of dollars at risk. These bonuses, you know, this these kind of slashings can be and they can be worth like hundreds of thousands of dollars or millions of dollars. On Oiler, we use uh we use a different mechanism. So, we don't take like a fixed amount of the collateral. Instead, we we have this like auction that plays out. It's a a Dutch auction that lays out on the bonus that's that's distributed. What's neat about that is that if you get liquidated on Oiler, the because it's a fair auction, the bonus

tends to be basically the fair market rate, like however much it should cost to to do the liquidation plus a little bit extra to kind of justify it. And so like the equivalent if you're a large B, an oiler and you get liquidated, the actual cost of it um might be, I don't know, a few hundred whereas the cost of it on on another platform might might be literally hundreds of thousands or millions of dollars. And to that just think that that like process that like focus on uh on auctions and how they can drive efficient outcomes. I think was like very much driven by like who we are as builders at Oiler which is is effectively just a team of engineers or scientists. Uh yeah all of our backgrounds have like a strong influence over how how oil is

architected. I think it's uh yeah certainly certainly always been a bit different to other other platforms in that regards. And so the the high level vision then where you guys ended up was basically like a lending protocol where you can borrow uh like any asset and use any asset as collateral. That's kind of the >> I think that was the that was the original vision and uh and then as we as we adapted as we grow over time we realized actually allowing anything to use as collateral is uh there are other other like factors that come into play right like even if you could do that you need to find like counterparties who are willing to take the other side of that

trade and the truth is that there's just not a market for that like there's I'd love to use my like longtail meme coins or whatever that I've got in my wallet as collateral but who wants to lend to that like who's going to take the the other side of that trade. There's just not really anyone available. And it's kind of like, you know, the that's that's true in real life, right? Some some assets just make really good collateral and you can go to a bank and say we want to take out our loan against a certain stock portfolio or or a house or car or whatever, but but some things just, you know, don't make for good collateral. Even though their paper value might be quite high, it doesn't mean that they're strong collateral. So yeah, I think there were a lot of learnings we made in those early years

uh as well about like actually there are certain things you can technically do but whether or not they they have like product market fit is a very different question. So if if you kind of go back to you know sort of history of oiler right so you talked about this hackathon and then how it evolved into this vision of having uh this you know very flexible lending protocol um at I know at some point you guys had a big hack I don't know when when did that like can you tell a bit like how did things progress uh from that point onwards >> yeah so we uh there was a there was a big exploit unfortunately in um uh yeah in 2023. So it was actually at the time

when the oil protocol being like growing very rapidly and taking market share from incumbents at that time. Oiler v1 was like a very very heavily audited platform like we'd had more audits than most people. Uh we you know we work with sora who do do formal verification. Um we had the largest largest bug bounty of anyone in that kind of vertical. So we we took security really seriously but uh what happened effectively was that we had this uh we had this bug bounty open and somebody came and reported a pretty minor bug on oil. So there's this uh smaller smaller issue that that first depositors in a brand whenever you open a brand new market the first depositor has some funds at risk and because it was a funds at risk uh bug we awarded

them a bounty uh a pretty small bounty um because the the amount of funds at risk wasn't huge but because there were some funds at risk our security partners uh when we talked to them said yeah you should probably fix this right so we developed a fix and uh it looked uh it looked good on paper and we had it reordited. So, we were with our security partners, they audited the fix, but I think ultimately the fix probably when you when you do a full audit the first time around, you're considering absolutely everything. You know, you're not just looking at them a small part of a protocol. And so, when you know full audits are fantastic, right? But when when they were auditing this smaller fix uh and when we were developing this like solution for this this much smaller

issue, it was much more of a contained uh you know it was more like a focus on that thing and we weren't both ourselves and the auditors weren't seeing that bigger like holistic vision. Turns out that by developing that fix, we'd actually inadvertently introduced a much more fundamental error uh or much more fundamental issue into the protocol. And that's how 9 months after we deployed that fix uh somebody was able to come in and effectively uh exploit a large amount of funds and take a large amount of funds from boiler v1 uh which brought down the type protocol.

>> What was that error that you guys introduced? So uh the solution to the problem uh well just very briefly I mean the problem itself was caused by an uninitialized exchange rate. The very when a very very brand new markets created and there's no users yet or anything there was one one exchange rate between the receipt token and the deposits that was uninitialized. It turns out that for technical reasons that an exploer could potentially frontr run a deposit if they set up a bot or something. they'd have to do quite a bit of work for it, but potentially they could like frontr run the first deposit and only that deposit and then use that to to take some of the first depositor's funds.

So the fix that we we developed was well uh you know why don't we take the first depositor's uh deposit and we will before they actually add the deposit to the account we'll use like one way or like some really small amount of that deposit uh to initialize the exchange rate. So there was this uh function that was added to the the codebase called donate to reserves. And the donate to reserves function was only intended to be used as a uh as a as a as a way to initialize this exchange rate. Uh and you know first depositor wouldn't even know it's happening. Now what what it turns out so what someone realized uh unbeknownst to us was that you could use donate to reserves um as a borrower.

If you donate your collateral to the reserves, if you do it in in some unique circumstances, you could drive yourself towards like a less collateralized position. And then if you can force yourself into a liquidatable state, you could then lose money uh and and and and liquidate the account. And so you might be thinking, well, why would anyone want to lose money? It turns out, we remember we've talked about those liquidation bonuses earlier. Turns out that there's a there was some like parameter space effectively where you could lose uh lose less money on the liquidatable account than you could make as the the person liquidating. So the person liquidate would get the bonus. And so uh it wasn't a huge difference, but it doesn't need

to be because it turns out that the attacker could then effectively kind of like loop to amplify the amplify both the losses and the gains. And so with this uh this looping strategy, they were able to um to withdraw a lot more and drain drain more funds from the protocol than normally it would allow. And so that was uh yeah that that's how it happened. >> What was the like TVL back then? And how much did the hacker manage to steal? >> So we had um I think total deposits were were sort of around five 600 million. uh the the taco was able to take out 200 million uh from the protocol.

>> Yeah, big big numbers and um and you know I can talk about like the the recovery process but we did re I should say up front we recovered all of the money and more. So we were able to recover 240 million uh from the exploiter over a period of several weeks following the attack. So we were able to track them down and uh negotiate the return of uh of of all the stolen funds. >> Oh. So yeah. So tell us about that. Like how did how did you guys track the guy down and how how did you why did they agree to return these funds?

>> There could be an entire like I I'm not kidding like an entire documentary series on on this uh exploit. It's one of the most uh harrowing moments of my life, but also retrospectively now if you look back on it, one of the most interesting sort of exploits in Tai history, I think. So, um yeah, I mean it firstly the reason it was particularly awful uh I mean it's awful for everybody involved when something like this happens, but for me personally, it was only 4 days after the birth of my son. So, uh I was on Yeah. like the the weekend my my son was born at like you know the early hours of Friday morning uh by Friday afternoon uh there was a banking crisis in the US in circle asset

USDC was deping so I was called into action to like deal with this crisis issue with USDC over the weekend so didn't really get chance to spend any time with my son then and then on uh on Monday morning I heard some alarms going off and I assumed it was related to the USD DC DPEG event, but uh it turns out it was actually an Oiler specific attack. Um and so yeah, my uh yeah, the next few weeks were were were then spent with him, my son like in the background and me like trying to negotiate return of all this money. Um the actual I mean there was so many twists and turns. the exploiter like tried to uh yeah well you know we track them down through various means like I

can't can't really reveal all of them but you know you have to do a like a huge amount of data gathering and uh >> so you figured out the identity of the person >> eventually yes uh although not not initially I mean we we had like a I think we had a list of around 10 candidates can't remember but like we we developed a list of people that that we quite quickly over about 24 hours we had a list of people that we thought could involved based on on available evidence. And um but yeah, we didn't actually know the identity of the person until quite a long long time after. Like even when we first started talking to them, we we we wanted them to believe that we knew who they were, but we didn't actually know which person they were out of our list

or whether they were on our list at all. in fact. Um, but yeah, they were they they started getting I mean they tried to do a donation to the a North Korean address to make it look like it was a North Korean uh exploit because uh if North Korea exploits the protocol, the funds don't come back, right? You don't >> then they're they're gone forever. >> Negotiate with North Korea. >> No, no. I mean uh so so it's over. So, I suppose uh I suppose he was uh his gambit there was, hey, if they think it's North Korea, then maybe uh maybe they'll just like go away. But, uh at that point, there was already like strong indications from our side that we that we knew it wasn't North Korea. So,

we were kind of able to call his bluff on on on those uh on those donation gambits and uh and then try to provoke more of a response from him and then try and try and just get closer and closer whilst uh trying to negotiate the return. And then the funds started coming back in drips and drabs. It wasn't like they all came back in one big block actually. They were probably they came back in maybe a batch of batches of like I don't know. Yeah, maybe like eight different times like different amounts of funds like would come back and he tried to play games like trying to p pretend to be multiple attackers at one point and uh and do all sorts of like really crazy stuff. So it was a it was a wild goose chase and uh it lasted for

weeks. Uh yeah. >> Wow. Wow. Interesting. And then in the end he said he recovered more than what he had stolen. >> Yeah. And I mean uh in dollar terms uh because the he when when when the tap is draining protocol they a lot of the assets come back in in in uh in terms of like USDC or UST right and circle and tether if uh if they get requests from law enforcement or they know that funds have clearly been stolen they have the ability to uh put freezes on assets. So, a lot of assets, a lot of attackers will just effectively auto autocon convert any stolen funds into uh ETH or an asset that's like more more decentralized and more difficult to kind of freeze

basically. So, that's what this guy did. And then um in the period of the recovery period where we were tracking him down and negotiating the recovery, the the price of ETH rallied. And so, actually, it was kind of like the attacker put on a long position on behalf of all of our users. on heat. >> Yeah. Whilst whilst it was going up in price. So when um when we finally recovered all you know all of the funds uh the the users got back there was an extra there was a surplus a $40 million surplus of of assets basically. So uh yeah.

>> Wow. So then you guys paid that out to the users or how did you >> Yeah. I mean yeah we uh yeah it was paid out to the users. I mean it's a very complex calculation figuring out how much uh users you know have in a protocol you know uh lending borrowing protocol like they users have like a net asset value right snapshot point in time but that that net asset value changes a lot over over a period of three weeks that can depending on what kind of positions you got you got loans you just got deposits just lending and so on but yeah it was all distributed back to the to the users uh uh so that was the recovery uh period which it itself took you know a long time and was very very difficult

>> and cuz the TVL went down to when did it went down to pretty much zero know or >> I mean the protocol wasn't viable after after that had happened. I mean um yeah so it wasn't like there was it wasn't uh it wasn't like there was a live protocol anymore. It was basically a dead a dead protocol. So it had to be uh the the the option was effectively it it turns out by the way that there was the the the the way to prevent this was like a there was like a a single single thing missing find this donate to reserves function.

The donate to reserves by itself wasn't the wasn't the biggest uh wasn't wasn't the issue. It was just there was a a missing like uh check in this in this function. And so uh had had that function had the missing check then it would have been fine. So in principle at that point uh we could have just if we wanted to to continue from there we could have just made that that fix to donate to reserves and then we launched oil v1 uh because it was a very good protocol and I think it's there's actually a fork of it uh somebody else has been using on one of the other networks I forget its name that's been going strong ever since. So, uh, yeah, it's it was a very good protocol aside from that that one crucial flaw, but uh, we we decided for for for other

reasons actually to not to not go with, uh, relaunching V1. You know, we decided as a team to stick together after that. Uh, it took us a while to come to that decision, like a few weeks, but um, you know, >> did you guys also think of like just shutting down or >> Oh, yeah. Yeah, of course. I mean, it was uh, extremely traumatizing for the team. Um and uh and obviously going to be really hard to recover from that. But uh we had a very good reputation in DeFi. I think up to that point you know it wasn't like uh this had happened out of carelessness or or because you know there was we hadn't got the audits or whatever. We were we were like above we went above and beyond at the you know with all the all the gold standards at

that time were kind of followed and so we had a lot of support from the border defi community and from the security community and so on to to continue. And the question was could we could we restore the faith to like potential users though, right? Like is it's it's one thing having a security reaches re researchers say well yeah they they you know they did their best and and still still this happened but um it's another like finding that faith with the users again. So there's a lot of concern I suppose that like even if we wanted to we wouldn't be able to come back reputationally from from from this. Uh but uh yeah, we we talked about it a lot and uh decided that uh we all liked each other and wanted to work in DeFi still and try and try and try to keep moving

the space forward and that we had had kind of a special team. And so we thought we we would rather than disperse and go and like join new projects or and try and like rebuild a new project, we thought we'll we'll give it a go. We, you know, we still had uh we still had the finances to do it fortunately from our investors and we had the backing of all of our investors to go out and rebuild. And so, yeah, we said uh so let's let's give it a shot. And so that's how we started building oil V2 at that point. >> And what were the biggest changes in V2? >> Yeah. Uh quite a lot to be honest. I mean V V1 was um we had a lot of we we learned a lot about from V1 about like the market and the product market fit

like we discussed earlier about you know what what is it like it's one thing to just build something but like does does it really have demand the main thing we learned I think was that there's there's not just like a one-sizefits- all in DeFi for lending and roaring like every single user has a slightly different riskreward preference uh you know some people like really conservative like simple markets that are low lower yielding but they do one thing and do it really well. Others like uh things like Ave, which are more like larger monolithic protocols, probably like more capital efficient but like with an elevated level of risk. Um and so like it's not clear that there's there you know there's there's just this like one perfect way to do credit markets um that fits all users. there's

there's a diversity of preferences. And so we thought how can we and by the way also like an increasingly like diverse number of assets as well. So there was like lots of you know new stable coins and and like state teeth and things were emerging and so on. So lots of different asset types. And so what we realized was rather than building like this one protocol that's like designed as a as a particular product for a specific user set, we said why don't we build a why don't we build the the infrastructure for making credit market products. So, we decided to build more of like a modular a modular protocol uh where you could have these like plug-and-play pieces that you could fit them together effectively to

then rebuild products. And so, you can rebuild Oiler V1 using this toolkit that we've got for Oiler V2, but you can also build other types of credit markets as well that might have a different design. They might be more conservative on the risk spectrum or or whatever. And then the the idea with V2 is that uh other people can then come and build products in their own image and tailor them to their the user bases that they have in mind and the riskreward uh demands that they they have on the protocol. So Oiler V2 is is not a product a single product. It's a it's a it's like a rainbow of products. There's just all sorts of all sorts of different types of uh things going on with V2 and and different operators as well for the for those products. There's lots of thing lots of

entities called risk curators or or curators that come and build their own build their own credit markets in oiler and that wasn't really something we had in V1. So does it make sense to think of like today if you think on like lending credit protocols you have like a that's you know of course the best known and uh which is you know very simple protocol and then I guess morpho is a bit more flexible and then you guys are even more flexible even more modular is is that like a reasonable way of or or like how do you compare Oilers sort of in the landscape of different D5 lending protocols.

>> Yeah, I think um I think I would say like at its core it's a toolkit for making D5 protocols. So you can build Morpho with Oiler or you can build a with Oiler or build something in between those two. But you there's that's not true the other way around, right? You can't um you can't build the primitive using their their toolkit. So like morpho fundamentally uh has these things called markets and markets in morpho are just like collateral debt pairs the simple isolated pairs and then uh they have many many pairs and then on top they have allocators. So capital allocators will then push capital into the different pair into the different pair markets and borrowers will will borrow from those pairs. Ave by contrast is this big monolithic market. So it

it's a market but it's rather than having just a pair of assets, it's got, you know, 30 or 40 assets in it, right? And they're they're crossc collateralized and so on. So it's it's more capital efficient, but but higher on the risk spectrum because now you don't have this um yeah, you don't have this isolation essentially. So if an asset >> it makes it more capital efficient because basically a has like one pool of USDC and you know and there's like one borrow rate and uh versus like for example Moro there will be like various different types of USDC pools and maybe some are like used to the max and some others are barely used some >> exactly and that kind of Yeah.

>> Yeah. Yeah. So on Moro um I mean just like the if you think about the design trade-offs there a little bit like there's there's there's fragmented USDC pools. Um now Moro tries to solve that for the lenders by effectively having like uh a higher level allocator ball. So most lenders deploy into the like allocator ball and then that disperses those funds into these lower level markets. Um and that's that's fine for the lenders to a degree but on the borrowers on the borrowing side it still means you have this huge huge amount of fragmentation which means still leads to like variable rates and um and uh and less capital efficiency there. The other thing about morpho is that the collateral is always held in like an escroed state which means it's not it's not rehypothecated at all. In ave uh the

markets crossclassize and often if you're taking a loan let's say you're using ETH to borrow USDC um the ETH is rehypothecated which means that somebody else can be borrowing the the ETH from you whilst you're using as collateral and that makes it more capital efficient because as a as a if I'm borrowing uh USD on Ave or rehypothecated market like we have an oiler as well it means uh that I'm earning interest on the on the ETH side and then paying interest on the on the debt. And like sometimes the two car cancel each other out. Sometimes it's actually even profitable to kind of uh take loans like that. Whereas on morpho your your ETH would typically just be sat in in kind of escrow and it wouldn't earn extra yield. So it might cost you

more. Um if you look at the what's called the utilization rate of these markets which is like the the the the average if you look across all possible positions uh of like the amount of um you know the basically the fraction of the fraction of assets which are borrowed on oiler the fraction of assets which are borrowed is around 50% or above uh on on uh on morpho by contrast it would be like low30s so there's a big gap between the capital efficiency of of markets which have relocation or enable that as a feature and the ones that don't. And uh that's one of the yeah one of the trade-off and design decision differences between Oiler and Morpho.

>> And so Oiler you also has have rehypothecation like in a >> Yeah. So on on Oiler you uh our primitive unit is not a market but actually even smaller than that. It's just a single vault and users can deploy assets into vaults and then they can use those assets as collateral or they can lend and borrow them from the vault. Now if you want to recreate a morpho pair you you basically connect two vaults together. You you have like one vault which does the lending and boring and one vault which just holds the collateral. It's got a single connection but on oiler you can because it's a toolkit and it's the modular toolkit you can actually extend that you can say that uh I'm going to act I'm going to

accept multiple class walls and then I would have like you know different types of class for my learning borrowing role. But you can go you can also do the relationship both ways. So you can say that A can borrow B and B can borrow A. Um so if you start to build up markets where you have uh lots of vaults which all cross reference one another as collateral that's effectively what what Ave is. Uh so yeah under the hood uh you can reconstruct an ave. The simplest the simplest possible AR would be something where you have just yeah people can deposit ETH and borrow USDC and people can deposit USDC and borrow ETH right.

Uh that would be like the simplest possible outfit with rehabocation. Um there's a protocol called silo which allows uh those kind of markets to be developed and uh yeah oiler has those too right we have some markets which are just that simple they're like the simplest possible a we have some markets where there's there's also many assets and they start to you know the market gets much bigger and starts to resemble more like a itself >> and where do you see the most traction today >> mainet uh in terms of networks we're we're deployed across 12 different networks and uh ETH mainet I think is the going to is is is the home of finance today or home of DeFi today. But there's uh we've got a lot of traction on Avalanche more recently as well. uh

recently deployed to Lana which is growing uh quick quickly and uh arbitrum um yeah there's it depends on which network you're on what what the asset bases there that all users have different profiles on different networks we found and so you see very different types of markets developing on different networks uh and uh yeah I mean stable coins yield yield bearing stable coins and stables more generally is has been the thing that's driven the explosive growth of both Oiler and and also Moro I think uh there's like just a diversity of yield bearing assets and uh and yeah demand to borrow >> so yield bearing stable coins where people will like deposit the yield

bearing stable coin and then they borrow against it. >> That's right. Yeah. >> So like USD or or like what are the top yield bearing stable coins in in oiler? I think uh USD will be like a massive one. Uh we also have like some open need and stuff as well. Lots of pendle tokens. So I mean the the basic trade that is super popular these days uh on oiler and other lending protocols is you deposit your yield bearing stable. Let's say it's paying 10%. And then you borrow a non-yield bearing stable but from a lending market and maybe you have to pay I don't know 7% to borrow. So um yeah if you >> that >> and then you loot that. Yeah. you so that you rather than getting just your

10% you can uh effectively earn the 10% plus the number you can get that 3% or 4% spread but on a loop. So you amplify your your uh you amplify your exposure to the interest rates and you take on board risk if uh if the if the stable coin dgs and it falls in price your collateral decreases in price and you risk liquidation. Assuming you're yeah assuming you're you're not worried about that risk, then you can effectively amplify the interest rate you earn by by kind of looping uh a yield bearing stable against a non-year bearing stable and and try and perform what's called like a carry trade. Uh it's very popular in trades uh been the the big trade of the the

past year I would say is is is some kind of carry trade on yielding non-yielding stables. What do you think the future looks like? I'm curious. I mean, I guess in in the traditional financial market, right? Like fixed fixed rate lending is a huge part. Um, I mean, I would guess probably a majority of debt is fixed, although you you probably know better what the breakdown is there. Whereas in I think DeFi, right, we have vast majority of lending happens with variable rates. Do you think this is going to change?

>> I think it will change. Yeah, for sure. Uh I mean there's we've got new products coming on oiler which are more targeted to uh fixed rate, you know, fixed rate borrowing. Uh we've got two two products coming actually uh in that regard, but there we're not we're not alone, right? Uh I think the other competitors have products coming too. The reason why they're in such huge demand, I believe, is that uh it's not so much, you know, lenders probably fairly happy with variable rates, right? Yeah, a variable rate is probably the the like the fair rate that you should be getting at any one point in time. The rate reflects like the the demand to to kind of borrow in the market and then like compensation you should get as a lender based on the risk that you're taking at that time. So I think for lenders they're they're fine

with it but for borrowers these carry trades uh become a bit unwieldy uh to carry out if uh if the thing you're borrowing is very variable right you you might have your yield going stable that's at 10%. And yeah, maybe maybe you start and you open up a trade and you're borrowing at like 7% or 8%. But then the rate spikes now it's actually more costly to borrow the thing than then you're than the thing you're using as collateral. So now you're in like this negative carry trade uh where you're actually losing money on on on a loop as well. So you're actually like amplifying your your losses. And so that's that's you know borrowers are getting increasingly frustrated with that, right? they they would much rather maybe pay an pay a premium initially to kind of lock in at a fixed rate. So maybe

they started maybe the variable rate pulls 7%. But maybe they're happy to pay 8% as long as they know that they can lock in and it won't go above 8%. So uh huge demand on the borrow side I would say for for for borrowers to be able to lock in uh lock in fixed cost borrowing positions uh for for short to medium to amount you know amount of time. >> Okay. And then and then I guess for the on the lender side they would also then accept or have to accept the fixed rate. >> Yeah. And this is the question, right?

Is uh is the are the lenders going to be just as happy with this? Um the fixed fixed rate products are are um definitely favorable for borrowers. Are they favorable for lenders? And do you have product market fit on both sides of the on both sides of the the plane here? I think on the uh for lenders, they would expect a premium, right? because they're also sacrificing like uh they're sacrifi sacrificing this like fair rate that they're getting. Uh the borrower should be prepared to pay this premium because it's it's clearly advant you know um advant providing an advantage for for that that party. So they yeah they're effectively trading like uh the f

they're playing a premium to like lock in lock in a uh a profit. So yeah that premium goes to the lenders and maybe that premium is enough but maybe it's not. The other the other thing the lenders forego I guess is that flexibility to withdraw. Um a lot of lenders um yeah a lot of lenders like to have that flexibility to just like withdraw whenever they want right a lot of lending protocols are designed with that in mind. So they they often have this uh unutilized portion of the pool or idle capital in the pool which allow the lenders to withdraw at any one point in time. if uh if you lock in at a fixed rate, however, you typically kind of get locked into a more of a fixed term as well, which means you you're you're losing that flexibility to to access

your capital when you need it. So, yeah, there's there's that like clear tension between lenders and borrowers and and one thing we know is that markets a really good way to like find like a resolution for that tension. So, it'll be interesting to see when these products ship. Uh I think I think it sounds like all all protocols are developing slightly different types of fixed rate products. Be interesting to see which ones get the best product market fit and satisi satisfy the needs of both the lenders and the borrowers in a way that's like kind of balances out and is fair. Do you think the demand on both sides will be mostly sort of short like I don't know a month or maybe even shorter duration or do you also see markets

developing for more long-term debt? >> I think initially it's going to be driven by um most borrowers are kind of like traders and they're working on like time horizons that are like months, not usually years. um doing you know as a borrower if you're doing these strategies uh you could in principle be like rebalancing every day and you can probably optimize yield by like finding the best strategy every day or you know rebalancing very frequently but then you have like the cost of rebalancing right which is not just uh is like gas cost but like your time like planning like then the added risk of like repetitively rebalancing like you might make a mistake in one of those things. So

rebalancing too frequently is kind of not something people tend to do. So yeah, usually borrowers like in my opinion will be looking for to lock trades in for weeks or maybe like several months at a time. So yeah, we see this kind of thing with uh Pendle, right? Pendle offers fixed rate products uh and um allow allow uh allow borrowers typically or like more like lenders I suppose to kind of like lock in for three month to six month periods as well. And I suspect we'll see fixed rate products that kind of match that sort of time frame. So that like yeah one month to six month sort of range will probably be where I put my money on on being the most popular.

>> So you mentioned Pendle, right? I mean Pendle is another D5 protocol that has gotten like a lot of traction. I my understanding of Pendle is that it's a lot of um around you know points and and basically you know someone will earn some unknown amount of some other token and then they basically they they basically sell those for like some fixed uh some fixed interest. Um, and then they get these PT tokens out with a specific maturity and and I guess they could then use that that as collateral in something like Oiler to to

borrow and to loop it. >> Yeah, exactly. >> Would you also think or I don't know if it's possible to build something like Pendle directly in Oiler? >> It would. Yeah, I mean you could definitely definitely do that with the kind of fixed rate products we've got coming. But like you say, I think pendals where pendals are exceptionally powerful is by it allows people to effectively convert intangible um intangible like incentives and rewards into like a concrete fixed rate. So that the you have people earning earning points on other things which are like quite hard to value. Um, and so they don't really know what they're worth and they like sell them off.

They'll kind of like lock in and like so for like make it real basically like make the make those points real and turn them into like something that they can actually trade on a proper secondary market. And that's where like Pendle services that need. Um and uh yeah, as you said, then once you've once you've converted that into a kind of like uh a real kind of yield, then people want to loop that yield and like amplify it on places like oiler where they'll come in and use collateral as a pendle token and do a carry trade and borrow non-yielding stables there. The loopers end up being kind of like junior capital. they're taking they're like seeking higher risk but like taking uh sorry seeking more reward but taking on higher risks uh liquidation and performing all these

loops and all the rest of it and then the lenders on the other side of that trade become more like senior capital where they're yeah they're not getting paid as much but they're they're kind of protected and they're protected through uh effectively the overcolateralization of the of the borrowers that's been the big popular trade I would say on that front but uh yeah you can you can build pendal in oiler but it's not like we're when where there's no plans to build that pendle product in in oiler frankly um it's a different kind of fixed rate I think people want to lock in in for what what do you see as the future for oiler >> uh I mean I think uh I think all of finance is coming so I've been saying

this for years I think all of finance is going to come on chain I think uh if you use swap products and you're a swapper you have no loyalty to who you use, right? You just put it through one inch or cow swap or whatever else and just get the best possible swap rate you can. But on credit markets, people don't do that, right? There's not going to be one winner. There's going to be uh there's going to be uh a diverse selection of uh of protocols because people want to uh diversify the risk. Uh if you're putting assets into somewhere or taking out loans uh over extended periods, you're exposed to extended periods of risk. and the the you know the there's this saying right that the um the only free lunch in finance is

like diversification. So I think oiler is going to be uh one of the largest uh credit protocols in the future of finance uh frankly and I think uh we won't be the only one. There'll be others other big protocols too but certainly we'll be up there and over time I imagine that we'll all slightly uh specialize. Usually what happens in markets is you find specialization. So Oiler will will develop a specialization for for certain uh classes of trades probably and do those better than anybody else whilst the other competitors end up specializing on other things as well. Yeah, in the short term I think uh right now we're still dealing with very cryptonative forms of like

credit, but I think there's increasing amounts of more traditional forms of credit coming on chain and I think uh oiler is very well uh set up as you know the architecture of oiler is very welcoming for uh yeah real world assets I would say and so that that's something that I think we'll be pushing forward on in the in the months and years to come. So you think in terms of you mentioned that different protocols will specialize in different ways. So you think for oiler it will be more around rwas. I think real world assets will be a big part of it. Yeah. Uh we also have like a swap protocol that's built on top of oiler uh oiler swap uh which um which which works extremely well with real

world assets and can open up growth opportunities for real real world assets that aren't aren't really there today. So yeah, I think that's today that's something that that we see as like a competitive advantage. Uh certainly. >> So when you talk about real world assets, what even got the type of real world assets that will get the most traction? >> Uh well there'll be the in the short term I think like DeFi has always been a very risk environment. So I think like tokenizing uh credit funds will be popular because they will be they'll generate more yield and those yields then get will get passed on to to lenders uh through through looping where we get like senior junior trenches

um just like we have with the DeFi assets today. Um I think initially >> well that's like credit funds uh where people basically I don't know for example lend a fund that lends dollars to for example businesses like that kind of thing and then they somehow like tokenize the fund and put it on chain. >> Yeah, that's right. Yeah, I think I think those kind of things they're definitely higher up the the risk spectrum. Uh but um but that's that's the kind of DeFi is a very risk environment generally and I think like with the the the people we have here today that'll probably be quite popular but over time I think the space will continue to

mature. We see like all sorts of fintexs coming on on and more like traditional institutions coming on chain as well and uh so then we'll start to see the emergence of of much much lower risk uh types of um of of real world assets coming on chain and being popular as well. I mean, there's there's already some of those here today, but they're just not widely used, right? We like see Bidd, it's um yeah, tokenized uh tokenized T bills, all those kind of things. Like, they they don't provide juicy yields uh and therefore they're not that popular to to trade among uh in D5 protocols today. Um and there there's friction to using them as well. There's like extra constraints with moving these assets around. Uh Oiler and Oiler Swap

helps lower those frictions. I think makes them more efficient. But even with those efficiencies, you still need the types of people that want to trade those assets. Uh in in, you know, Tradfire, you have like repo markets and and money markets and things where you see like really large sums of money like changing hands over like short short periods of time. Uh and there's just nowhere near enough liquidity in in in DeFi today to support those kind of trades. But increasingly that's changing. I think we'll that will that will be very popular in the in the years to come as as the space matures.

>> I I know one of the things that was often discussed as a big bottleneck to getting u more financial institutions on chain is like privacy. Do you how what do you think is the role of privacy in future of DeFi? >> I think it's going to be it's it's an interesting one. There's a huge tension, right, between um between the the the desire for privacy, but then the the desire for transparency as well. Um and where where you sit on that spectrum uh I think depends a lot on your own personal background, like your experiences, like how you think. It's more of like a political thing, right? I mean uh I think if you look historically at like

protocols that have provided uh like a really high degree of privacy in in DeFi uh it's been fairly heavy use uh of like illicit finance by by by bad actors and that that puts traditional finance off actually like they don't then you're not going to see uh you know institutions mixing funds with uh with with elicit fund. like funds that were taken in in hacks and like you know mixing with other sorts of elicit finance. So that's that's one form of privacy that's yeah I just I think it's going to struggle to get adoption honestly. But um on the other hand institutions also don't want to you to always know what they're

trading right like especially if it's a directional trade. I don't want to be putting on a trade on Bitcoin or whatever else and let everybody else just like pick me off because they can see that I'm like got I'm long and I've got certain types of exposure. So um yeah there's a huge tension and I don't know where it will resolve. I don't think it's it's not like it's not that we we don't have the technical capabilities to make these protocols. It's more of like a political thing and how much how much how much should we trade off between transparency which tends to make like fairer markets versus uh privacy which is like argu arguably fairer for individuals and their personal freedoms, right? I um yeah I just don't know in in like most

regulations in in Tradfi seem to be driving towards more transparency not less but if you ask any individual person like should I have more privacy I think most people would would want more privacy right and that's that's only right so yeah there's this huge huge societal and and tension there between those two things >> yeah and I guess the other thing I mean maybe that is a solvable issue right but of course one of the advant advantage of DeFi on the transparency of DeFi is also much easier to assess where the risks are.

>> Yeah. >> Uh I mean hard enough, but at least the the information is there and someone can try to do it. But then if if you add more privacy that that probably also makes it much harder to understand where the risks are and you know if this fails, what else does it result in? >> Yeah. I mean, I was um I entered the like labor market in uh August 2008 and I joined a bank in the UK called the Royal Bank of Scotland just as the entire system was like collapsing, right? Um, and I remember hearing from that that you know after that after people like went through the went through everything afterwards and I tried to like piece it

all together and it was just like an absolute nightmare and I still don't think people completely understand what happened because like actually mapping it all out and like you know looking when all these different banks collapsed and everything like what what actually went wrong what was the the trigger and like where were you know where were the funds flowing all of it was very hard to like process it and I think that at least indeed that's all easily mappable on chain and if someone wants to do it they can uh it's not not necessarily like easy as you say but um it's much much more transparent in that regard and I think that will be healthier for financial markets in the long run uh even though it does pose its own challenges there may be you know we may yet see technologies emerge which actually are able to kind of balance

that tension out where um yeah you you can kind of have like uh like privacy is like the de facto norm, but like it can kind of you can like under certain conditions like reveal certain things or whatever if it's if it's essential to do so to make sure that you aren't uh you aren't like mixing funds with elicit finance and so on. But I don't I don't know. I mean it's uh it's not something it's not something that I'm working on personally at the moment. We we're working with what we've got which right now is it's mostly like max transparency I would say on uh open public blockchains.

>> What's your biggest focus right now? >> Oo uh I mean personally uh we have new products coming. We've discussed on the fixed rate side. We've been uh working on growing our swap protocol. And uh from my side like a lot of my time is spent like going and and and meeting uh new new types of users like institutions and other people that and talking to them about how how oiler works effectively and educating them about how how the system works and seeing if they can be encouraged to come and use the protocol, you know, new types of asset issuers and so on.

There's just almost not enough hours in the day at the moment. like it really there's uh yeah there's just a massive massive shift uh in the past year uh and growth opportunities absolutely everywhere in DeFi right now I would say >> what do you think about the the impact AI is going to have on DeFi. Ooh. Uh I mean um as we've mentioned, I think like one thing that's that's certainly certainly likely to happen. You know, we've talked about like rebalancing strategies and all the rest of it and making sure that you're uh able to able to kind of monitor positions and so on. I imagine that that's something that's going to be handled by by AI.

We've already like internally like played with some some like machine learning algorithms for like optimizing yield basically by like rebalancing like lending positions and so on. I think you could do that with AI as well, like have like smart aing basically where you rather than you having to sit at the computer like with all your spreadsheets like the uh you know the financial analysts going through things pouring over data like right really intelligent bots that can basically under some constraints move move the funds on your behalf to either optimize the yield as like the the senior trench or to to optimize the lending and borrowing positions as the the kind of junior branch of things. So I think that will be yeah will will change things a lot uh and lead to more efficient markets

overall. There's huge inefficiencies in DeFi today. Rate optimization is not really a thing. Uh it's there's there's protocols like IOR and others and and that's you know do it like it's the kind of what year introduced like when I first started back in 2020 like think was we're doing this kind of thing but it's still it's still a very inefficient market you know by and large and I think AI is going to change that. Cool. Well, thank you so much for coming on. It was really cool to dive into Oiler and um yeah, I think I think what you've pointed out before like you know with uh Trafi and and RWA is coming on chain a lot more capital

coming on chain uh this is going to be an you know I think TFI will continue to be an extremely interesting space. So I'm really excited for what you guys are building. >> Thank you. Yeah, thank thanks for having me on. It was yeah, a really good chat. Lots of different topics. It was great. There's just so much to do right now. It's a very very exciting time.