DeFi Utopia by Eugene Chen

Transcript

all right thanks for showing up everyone uh I'm Eugene I'm a co-founder of ellipses Labs uh today I'm going to talk about our vision of how to build a defi ecosystem that is actually better than what exists in traditional Finance so at Ellipsis Labs we're core contributors to Phoenix which is a limit order book spot deex on salana so so this is part of our UI here this is uh the live order book updating and the entire state is fully on chain so the matching engine is on chain the entire book is on chain uh all on the salon of blockchain like I said fenix's fully onchain order book decks uh today we've

done actually over $24 billion doll in cumulative spot volume uh daily volume usually between three and $500 million so that's in the top five five of all spot dexes across all uh blockchains and the most important thing in my mind is the market makers themselves are profitable and there's no incentives to market makers so we have a really credible story that the liquidity on Phoenix is sustainable so now I want to talk a little bit about uh what the path looks like going forward and before we get into that we'll need to talk a little bit about the history of defi so early defi was really enabled by the launch of

ethereum mainnet which was the first generalized uh generalized decentralized compute environment and it enabled early D5 Pioneers like Unis Swap and maker and a to deploy their protocols and they had these properties like because they were built on top of ethereum mainnet because the contracts themselves were immutable uh interacting with them did not require any trust in the system uh all of the protocols were auditable and sensorship resistant at least to the extent that the underlying layer one was they also showed there there's some net new use cases so Unis swap enabled liquidity for a new coin without a centralized

exchange listing or without a uh sophisticated Market maker and lending protocols like a and compound showed that you could actually have these completely trustless lending protocols where you don't need to trust your counterparty but all of these Protocols are actually quite primitive they uh were designed to be safe they were designed to be proofs of concept and they're really constrained by the underlying infra of the ethereum L1 those guys all came out before the year 2017 it's been seven years since then and the underlying blockchains have gotten a lot better but the D5 protocols themselves have not we've seen a ton of marginal improvements we've seen unisoft

go from V2 to V3 we've seen a lot of ways to reh hypothecate risk to uh stack yield with more leverage but the core protocols themselves borrowing and trading have not gotten much better and even when we go look at these new blockchains that deliver much more performance we still see these original D5 protocols getting copy pasted which doesn't really make sense because if you have a blockchain that is much more performant than ethereum it stands the reason that you should be able to build protocols themselves that are a lot more sophisticated a lot more efficient today in the market we're seeing this new school of thought that I

call shitcoin nihilism which is this very pessimistic view for defi it says that defi is never going to outperform trafi we shouldn't even aim for that uh it's it's too lofty of a goal the only thing that's going to come out of defi maybe the only thing that's going to come out of crypto is these these shitcoins that we've made and the only value add for defi is Casino that's enabled by regulatory Arbitrage which yeah for sure is the the main use case that we see today and so we're at this Crossroads for defi where I think this pessimistic view which I term defi dystopia is it's clearly where a lot of the market is looking today just optimizing

the casino that we have but I think the future is a lot brighter where there's a ton of potential to actually build a system that is better than what exists in traditional Finance there's a few key reasons why this ought to be possible rather than just this p in the sky vision and so that's what I'm going to get into next so first we're going to take a look at some of the problems that exist in traditional finance and what the core advantages are for def5 so Trad is this really big system is global in scale um but it moves really slowly and the core reason for that is it's been around for so long so trafi traditional Finance is

a system that has evolved iteratively over a hundred years starting from a point where an equity in a company is represented as slips of paper in someone's safety deposit box and from that point we've like upgraded it to where it is today and so it's this really big system it moves really slowly it has a lot of scar tissue uh there if anyone wants to make any sort of changes there's a ton of different entities with different interests that all have veto powers and so this really slows the pace of innovation where something as simple as implementing a new order type on the New York Stock Exchange is a type of process that actually takes 5 to 10 years whereas in defi that's something you know a new engineer could Implement

in in weeks and just put it out into production and I think the most important piece here is there's no opportunity for new ideas so we've been stuck with some similar architectures uh for a very long time there actually no way to there's not even like a Sandbox for which we can try new things in traditional Finance so it's very easy to get stuck in these local Optimum so when we talk about defi most teams really like to emphasize permissionless access and auditability trustless auditability and composition and I think these properties are quite important but when it comes to competing with traditional Finance I think the most important piece is actually

permissionless innovation this is the thing that enables defi to iterate significantly faster than the traditional Financial system again using the example of the New Order type uh that's the type of thing will'll Take 5 to 10 years on a centralized exchange especially a regulated centralized exchange whereas uh for something like like Phoenix we built the original prototype in six months and then when we were ready we just deployed it to production and it was in this environment where we could see we could see if it worked or not there was um you know an environment that had real assets on it the salon of blockchain and it's worked pretty well and we can also uh deploy new versions if you want to

deploy new versions we get really fast feedback from the market and this is the type of thing that's just totally impossible in traditional Finance so when I talk about competition there's two main Dimensions I care about and these are really micr structure things obviously there's a ton of other pieces that go into having d5b competitive but this is sort of like the backbone these are the things that are required before we can even go out and like do the BD to bring on the real world assets uh this comes first so the first is efficiency for a trading platform this is usually going to refer to the quality of the liquidity so how deep is the liquidity and how tight are

the spreads and then we also need to be able to compete on sustainability where we have all these D5 protocols today that sort of depend on this infinite spigot of incentives that is unsustainable and when you take away the incentives oftentimes the the usage just completely dries up and we need to look at these in a very sober way I think there's a often times this desire to have this rah rah rah moment in crypto where we all pretend that we have product Market fit when actually we have token Market fit and we need to look at okay how do we actually build the better Primitives and the first piece here is better infrastructure uh on the application

side like I said before Phoenix is really just this proof of concept uh there is no way that Phoenix in his current form is close to good enough to displacing traditional Finance because we haven't innovated whatsoever on the liquidity primitive so it's the same primitive that exists on coinbase the New York Stock Exchange it's a vanilla limit order book and it's on this infrastructure that's like 10,000 times less efficient it does prove though that active liquidity can outperform passive liquidity so I do view that as a step function Improvement in in D5 so when we take a look at the the infra you know ethereum mainnet today does around 20 transactions per second salana main net does around excuse me

salana main net beta does around uh a th000 transactions per second and then the cost to land a transaction is anywhere from um a single Cent maybe a little bit under a single Cent up to Dollars and this is about a thousand times worse than centralized infrastructure when you think about the cost to place or cancel an order on a venue like coinbase or finance and we just need like really big improvements here I mean I don't think we'll ever get down to 1x while maintaining the properties that we care about uh on decentralized infrastructure but we should be able to get it down to a factor of 10 to 100x where we can get get blockchains that do tens of thousands of transactions per second

where the block space is so abundant that the cost per transaction is under a tenth of a cent latency is also extremely important we haven't really talked about latency much today but um market makers getting pre-confirmation under 100 millisecs ideally under 10 milliseconds that's super important and we need to do all this while still maintaining this trustless auditability still maintaining permissionless Innovation obviously if we don't care about those things then AWS has already solved this problem for us luckily for us there's a bunch of teams that are working on this infrastructure scaling problem there's two main approaches today the first is uh to scale the layer one and this is done by I guess the market leaders today are

like salana and monad where under the proof of stake architecture just really trying to squeeze as much performance out of that as possible uh through hardware and software improvements and then the other approach is to scale via Layer Two where we're going to toss out uh short-term sensorship resistant and get a lot of performance on um uh for that trade and the path to scaling these is quite clear where instead of using ethereum da you can use uh da from igen layer sorry Igan da or or Celestia and then with uh centralized or rotating sequencers you can really crank up the performance requirements there and as the the cost of ZK proving comes down uh

we can see these pretty clear paths for both layer ones and layer twos to get to tens of thousands of TPS and then the question is can we get the cost down low enough such that it's actually uh possible to provide active liquidity and then supposing we get the infer there and I I expect us to in the next one to three years let's say then we really need to take advantage of the infrastructure and this is what is really missing from the market today where the core D5 Primitives need to get more efficiency from better infrastructure so on the trading side uh I think like the limit order book active liquidity makes a ton of sense but there's still going to be room for things that we can put on chain that we

couldn't put on an exchange like coinbase for example uh programmable liquidity curves uh make a lot of sense orders with side effects uh these things can only be done on a shared State machine and so there's actually a compelling reason for why it only makes sense in defi on the Bridge side frankly all of the borrowing mechanisms we see in crypto today are pretty much copy paste of the a model from 2016 2017 but when the blockchain runs really fast when the Oracle update is really cheap when the Oracle comes really really reliably like every half a second every one tenth of a second you can make a much more efficient leverage mechanism uh you can even have active liquidity there where

things look more like actively traded uh bonds fixed income products on an order book or on some other trading primitive and then your liquidations can be a lot less aggressive which is just a better end State for the user so there's a lot of uh room to experiment here um and I think the most important thing is when we build these big blockchains we can copy paste original protocols that exist on ethereum and then they'll be cheaper on the new infrastructure but the real step function improvements come from building better protocols that are uniquely enabled by the new inra in the same way that Unis swap was uniquely enabled by ethereum actually going live

and going to mainnet so we're working on all of the above at alyus Labs uh we've been around for a year and a half um and we are hiring across all job functions all business functions in our New York City office so if you're interested in building the future of Finance not just the future of defi but we hope we can take a real whack at building a better Global Financial system uh feel free to hit me up thank you one two three okay thank you a lot of

actions for my small question thank you a lot uh you mentioned an interesting uh topic about Protocols of Dey and we understand that right now there are a lot of Dey Solutions and what about about some common protocol about this because if we speak for example about Cosmos inter blockchain communication so there is only one protocol only one way and speaking about Dey for ethereum for salana there are a lot of ways and what do you think about future will we have only one approach or only one protocol or still a lot of of them yeah I think it's pretty difficult to predict uh I

think in general in crypto today we overestimate sorry we underestimate the difficulty of this like cross chain interoperability problem especially if we're not going to uh trust somebody in the middle which in the current crypto world looks like a basically like a centralized exchange um on the protocol design side I think no matter what you're just going to end up with a bunch of protocols and there's going to be this longtail of protocols that has you know not very much usage and not very much tbll and I think that's totally fine that's sort of the what you expect to see in an environment with permissionless Innovation where you have like a 100 well-meaning teams try to build a a good liquidity primitive or a thousand well

meaning teams and at the end of the day only one or two of them is going to work out but it's going to be not obvious a prior which are those um so I think that's just the nature of the hyper capitalist onchain environment thank you lot does it means that yes just smaller amount yeah yeah uh yeah I expect there to be plenty of blockchains with not very much consolidation and plenty of protocols with not that much consolidation for the foreseeable future thank you Lord have one more please thank you so much for the speech

was super interesting to listen uh a small question am I understanding correctly that active liquidity means active Market making uh where are placed yes so uh if we move as an industry on like even defi to active Market making doesn't this remove uh in a sense the decentralized nature of for example like pools on current amm like Unis swap where you put the money in the pool and unless it's fully drowned out it exists there forever and people can do swaps like forever without any personal involvement of anybody yeah so there's still uh permission less access to the liquidity anyone can anyone can be a taker on a limit

order book and actually anyone can be a maker as well so anyone can place limit orders but yeah it's definitely true that the barrier to entry to being a competitive market maker is higher which I view is a good thing because I care about quality of liquidity uh and yeah I don't think uh you know myself clicking one button should I necessarily be competitive with jump trading like probably not you have all these high frequency trading firms that have spent tens of billions of dollars in R&D over the last you know 20 years 30 years getting really good at this business of electronic Market making which is trying to provide better spreads tighter spreads more depth to win the retail flow and that competition is really really good for the end user whereas I

would view the you know the XY equals K World almost as a communist world where you have no freedom to provide liquidity where you want to provide it you're stuck to this curve and that leads to much less efficient liquidity I think there's also other ways to enable um uh more retail participation so one pretty clear retail basing product that one could build would be lending to market makers where if uh I'm Citadel and my hurdle rate on the capital is 50% and then my market making strategy on Phoenix only makes 20% I'm not going to do with my own Capital but maybe I can borrow from retail at at 15% and there there should be a way to make that sort of trade happen just here the main question is that I may be wrong here

please correct me but in this case the market maker is uh the point of censorship that can be used uh if they actively have to do something and they can press a stop button to stop doing it uh this is the point of failure uh yeah but there's multiple market makers in the same way that un centralized exchange there's many market makers and probably all of them face some amount of downtime some of them go down for very long amounts of time and the end user doesn't really notice it most of the time because there's plenty of market makers and the reason the market makers exist is because the business of being a market maker is quite good so there's incentives to make this the case yeah of course I'm not speaking about like efficiency I'm speaking um censorship in a sense that like the problem of

centralized exchanges is in essence that you have to convince somebody in the sex to get you listed and then you have to like pay market makers and if the market makers go out of business then you have to find new ones and like it's a lot of points in which the system stops being uh decentralized in a sense I mean I think the core architecture like the social layer here does end up much more centralized where you realistically speaking have like somewhere between four and 10 dominant market makers who make money and a bunch of others um uh and and there is some Reliance on these guys but when they go away they get replaced very quickly because there's so much business

opportunity there and uh there are like a couple big differences here with the centralized exchange where it's probably a little tougher to provide liquidity on the decentralized exchange but uh you don't have to worry or your risk profile looks very different where okay The Exchange can't you know run away with everybody's money and um you know that no other Market maker has an advantage over you so actually in some sense that's lowering the barrier to entry compared to a centralized exchange with all like you know the fee tiers and whatnot okay so it's kind of in between in a sense yeah and uh yeah just to be clear I don't disagree with the premise of your question I think it's a very reason point of view well thank you so much man uh amazing uh presentation cheers okay thank you a lot of actions

for my small question thank you a lot uh you mentioned an interesting uh topic about Protocols of Dey and we understand that right now there are a lot of Dey Solutions and what about some common protocol about this because if we speak for example about Cosmos C blockchain communication so there is only one protocol only one way and speaking about de for ethereum for salana there are a lot of ways and what do you think about future will we have only one approach or only one protocol or still a lot of of them yeah I think it's pretty difficult to predict uh I think in general in crypto today we

overestimate sorry we underestimate the difficulty of this like cross chain interoperability problem especially if we're not going to uh trust somebody in the middle which in the current crypto world looks like a basically like a centralized exchange um on the protocol design side I think no matter what you're just going to end up with a bunch of protocols and there's going to be this longtail of protocols that has you know not very much usage and not very much tvl and I think that's totally fine that's sort of the wa what do you expect to see in an environment with permissionless Innovation where you have like a 100 well-meaning teams teams try to build a a good liquidity primitive or a thousand well meaning teams and at the end of the day only one or two of them is going to

work out but it's going to be not obvious a prior which are those um so I think that's just the nature of the hyper capitalist onchain environment thank you lot does it mean that yes just smaller amount yeah yeah uh yeah I expect there to be plenty of blockchains with not very much consolidation and plenty of protocols with not that much consolidation for the foreseeable future thank you Lord have one more please thank you so much for the speech was super interesting to listen uh a small question am I understanding

correctly that active liquidity means active Market making uh where orders are placed yes so uh if we move as an industry on like even defi toess Market making doesn't this remove uh in a sense the decentralized nature of for example like pools on current amm like Unis swap where you put the money in the pool and unless it's fully drowned out it exists there forever and people can do swaps like forever without any personal involvement of anybody yeah so there's still uh permissionless access to the liquidity anyone can anyone can be a taker on a limit order book book and actually anyone can

be a maker as well so anyone can place limit orders but yeah it's definitely true that the barrier to entry to being a competitive market maker is higher which I view is a good thing because I care about quality of liquidity uh and yeah I don't think uh you know myself clicking one button should I necessarily be competitive with jump trading like probably not you have all these high frequency trading firms that have spent tens of billions of dollars in R&D over the last you know 20 years 30 years getting really good at this business of electronic Market making which is trying to provide better spreads tighter spreads more depth to win the retail flow and that competition is really really good for the end user whereas I would view the you know the XY equals K

World almost as a communist world where you have no freedom to provide liquidity where you want to provide it you're stuck to this curve and that leads to much less efficient liquidity I think there's also other ways to enable um uh more retail participation so one pretty clear retail basing product that one could build would be lending to market makers where if I'm Citadel and my hurdle rate on the capital is 50% and then my market making strategy on Phoenix only makes 20% I'm not going to do with my own Capital but maybe I can borrow from retail at at 15% and there there should be a way to make that sort of trade happen just here the main question is that I may be wrong here please correct me but in this case the

market maker is uh the point of censorship that can be used uh if they actively have to do something and they can press a stop button to stop doing it uh this is the point of failure uh yeah but there's multiple market makers in the same way that on a centralized exchange there's many market makers and probably all of them face some amount of downtime some of them go down for very long amounts of time and the end user doesn't really notice it most of the time because there's plenty of market makers and the reason the market makers exist is because the business of being a market maker is quite good so there's incentives to make this the case yeah of course I'm not speaking about like efficiency I'm speaking um censorship in a sense that like the problem of centralized exchanges is in essence that

you have to convince somebody in the sex to get you listed and then you have to like pay market makers and if the market makers go out of business then you have to find new ones and like it's a lot of points in which the system stops being uh decentralized in a sense I mean I think the core architecture like the social layer here does end up much more centralized where you realistically speaking have like somewhere between four and 10 dominant market makers who make money and a bunch of others um uh and and there is some Reliance on these guys but when they go away they get replaced very quickly because there's so much business opportunity there and uh there are like a couple big differences here with the

centralized exchange where it's probably a little tougher to provide liquidity on the decentralized exchange but uh you don't have to worry or your risk profile looks very different where okay The Exchange can't you know run away with everybody's money and um you know that no other Market maker has an advantage over you so actually in some sense that's lowering the barrier to entry compared to a centralized exchange with all like you know the fee tiers and whatnot okay so it's kind of in between in a sense yeah and uh yeah just to be clear I don't disagree with the premise of your question I think it's a very reasonable point of view but thank you so much man uh amazing uh presentation cheers guys thank thank you so much uh this is all thank you pleas Round of

Applause