Curve Finance - Fireside Chat with Michael Egorov

hey everybody how we doing um I'm defi Dave uh interviewing Michael from curve uh he needs no introduction how you doing Michael how's it going hi hi good all right let's let's get this started um for those that may be unfamiliar with yourself can you give a brief introduction and what your philosophy of building is because you primarily build products for yourself right yes pretty much so um well I guess just to introduce myself for those who who don't know I'm started curve in 2020 um and before then I did new Cipher which is now um which is now doing the T tbdc um and um before that I actually had an education as a physicist so I'm using pretty much experimental physics approaches to to curve in some sense um uh but uh yeah anyway and I'm I'm building uh products for for myself because uh originally idea of curve the first idea was the exchange for stable coins uh appeared because I was a very active user of maker D and I needed to uh exchange between essentially D and usdc at that time um and you know centralized exchanges couldn't really do a job nearly as good even for small amount so uh that's why uh I started curve uh but right now it's uh quite a bit more than just a stable coin exchange although this is still a very very big fraction of what curve is doing yeah exactly and with all your new products especially CRV USD which is one tenen of the tvl but the vast majority of the revenue can you go into how what CRV USD is how the Llama mechanism works and why it's been so successful yes so Ser USD is a CDP um over collateralized stable coin and uh in principle it's uh well in a nutshell any CDP is where you mint stable coin by borrowing and there are also uh liquidations and stuff like that but uh what's interesting is what's what's so special about CV USD is this mechanism of soft liquidations so um it works this way that if price of uh collateral goes down uh for what uh for whatever reason due to market conditions then uh it uh the system essentially converts some of your collateral to CVUSD and converts back on the way up and it does it so efficiently that all of these little market fluctuations they don't uh uh don't drain too much value from the system so if you choose very conservative parameters for the loan which you can um even under liquidation price you may stay there for like months without getting actually liquidated and then when price of collateral goes up again you get your collateral back uh with the loss of uh you know some you know maybe few per of collateral on these inevitable uh inevitable losses in these conversions back and forth so uh essentially uh it appears that you uh can avoid being feared of uh of a deep where you get liquidated and then uh you know having maybe your loan forgiven but uh you don't have a collateral anymore so when uh prices of crypto go up again you don't have your precious crypto which is uh kind of terrifying for me so and that's why uh that's why this algorithm was um created um or in addition it appears that if prices of your collateral do go down and don't come back up you still can um you can close your position at at the bottom because you already appeared to be um to be U essentially sold your collateral at higher prices so which is also a nice property it's like a stop loss in some sense um so that's um essentially how it works and many people use it for trading strategies um like when they don't know if crypto is going up or down but they can leverage and essentially the whole system works as a stop loss for them uh much more efficient than our traditional stop losses so or uh they can do this strategy when they can uh stay uh under liquidation price for months and uh if that becomes necessary and still be okay but that of course depends on how they configure their alone and that's both CRV USD and uh the curve Landing product Lama land it uses exactly the same algorithm and actually code yeah and um going to The Exchange now uh the most volume pairs on curve have really a minimal amount of CRV rewards I noticed so would you consider this liquidity officially sticky and what does this mean for CRV rewards if the highest revenue is CRV USD anyway yeah that's that's an interesting one because I think this appeared with all this raking uh narrative as well as points and uh things like that uh which came both with RIS taking and Etha right so um essentially the points are very very good for bootstrap and liquidity and it doesn't really uh well doesn't really require uh CRV emissions there is also a good point that maybe it is possible to give these points for votes to get actual CRVs for liquidity but I don't know most uh mostly projects played uh giving points uh directly to liquidity and that worked um but there is one other thing which I noticed these uh all these tools uh which you know with the projects with points they have um a very high trading volume to LTV to tvl ratio so uh trading volumes are high and the APR which goes from Trading feys is also high so some of the pools actually have a natural APR being pretty big maybe in addition to points maybe not but uh it's all related to all to these new projects and uh that's pretty exciting to see so it looks like they do have a good uh natural um um like natural appeal like even besides points there are also very big trading fees going on so that's that's an interesting thing which I'm observing and uh um yeah uh trying to explain explain it for myself like how how can it happen that trading volumes are so high in this uh uh in these cryptocurrencies yeah and going uh further on exchanges compared to a few years ago uh some some of these highest volume pairs are factory pools and they're not base pools um was this expected and what are your thoughts on it yes that's that's absolutely expected and uh this is why we switched to fact is because uh you know uh you you don't need to uh we just didn't have this mechanism originally so it was just uh pools being deployed and accepted to the Dow so it's a little bit of a convoluted process but it's uh obviously easier to make it on self- sered basis and it was done and it's it was just a matter of time that all of the pools become factory pools so even uh some of the the core pools like three cryptos or our our factory pools because like Factory is just the way to uh uh the way to make it easy for people to to create new pools yeah now going to l2s and the proliferation of different l2s and chains what has been C strategy for deploying there and what have you noticed in the process yeah we didn't really push overly too much on l2s uh after after the multi-chain uh multi-chain disaster well on l2s we we actually did but not so much on other chains I would say uh so we uh we were doing essentially major l2s like arbitrum but uh didn't make uh kind of that the main story uh yeah and actually with the ethereum uh with ethereum gas fees being low it's uh um well it's probably reasonable but of course L2 narrative is uh is an interesting one so uh so yeah yeah and actually going back you mentioned llama in the be beginning can you go over how it works how it fits into the curve ecosystem and how it compares to other lending markets right so Lama land works with exactly the same soft liquidation mechanism as svusd um and uh it allows well the only difference is that the supply of CRV USD goes not from minting but from other people depositing right so the idea is that Lama land serves a supply sync for CRV USD uh this is why Lama land doesn't have admin fee at all but um the idea is uh that people depositing CRV USD for other people's people to borrow will create some um you know some CRV USD held there so and because of that this uh uh this kind of balances uh supply and demand side of CV USD something uh which uh which was mentioned in liquid talk so Lama land is serving uh serving uh this purpose but for the user it's just a lending and borrowing protocol so people can borrow and enjoy exactly the same soft liquidation mechanism not necessarily just with major uh cryptos but with uh with other cryptos as well and um on on the other on the other side uh lenders can choose which uh collateral uh to to give to borrow against and uh uh lenders uh can uh can choose that by uh uh let's say by safety preferences so maybe given to borrow against uh ether is inherently safer than giving to borrow against uh some uh uh some essentially lower cap uh crypto um all markets are isolated so you essentially if you um give some something to borrow against one crypto it doesn't really affect the other markets so and and vice versa um so unlike AI where everything is uh interconnected or here it's uh very isolated Landing uh maybe can be seen as a protocol for having some other protocols on top of it so for example it is possible to build a state CRV USD product uh which balances between different uh different markets on lamaland depending on uh safety preferences so uh but yeah for User it's the same as uh as CRV USD so one should definitely check and uh see where the borrow rate is uh more preferable whether is it on lamarand or is it in sery USD for some reason on larland borrow rates are lower uh consistently uh which is I don't know U hard to explain but um that's what it is maybe some people who use serusd uh didn't really figure out that Lama land is essentially the same thing for for them yeah uh so a year ago you were at the first stable Summit and since then a lot has happened to say the least uh what has been some of the biggest lessons you learned and what are you most excited about in the future mhm um yeah so um I guess with uh with everything happened we uh um well one of the uh one of the lessons is that uh we probably need to well we we figured that before we need to support uh safety of uh infrastructure which we use and uh essentially this was done with Viper so Viper is a very very safe language right now um and um we have uh we did have an exploit due to uh due to a bug in one of early Viper versions uh which and after that Viper was completely um completely refactored and audited multiple times so I'm pretty sure it's probably uh probably the for me it's looks like the safest language at the moment um but yeah so it was very good that we uh and I personally invested in safety of uh of Viper uh so right now pretty confident in that uh another one uh is uh with regards to uh to lamarand uh the uh kind of the uh with the uh huge liquidations uh recently it was there was a lot of data obtained on uh how how this system with soft liquidations Works in under extreme conditions and uh it can allow to introduce maybe uh maybe further uh improvements uh on let's say how to determine uh how to determine limits uh for non- major markets whether it's just um liquidity limits or whether it's uh something else uh and also some interesting data about how CRV USD pad uh uh works when when big liquidations happen and uh how to make uh how to make uh all of the markets not really sensitive to uh to CRV USD uh pegging up or down so this a lot of good data about that was obtained and uh um I think uh will be applied in the nearest Future um so I probably will uh will publish some research about that but uh yeah very uh very very interesting and for the curve road map uh what should we look out for in the future well I think um uh one uh well there are multiple things one uh very important thing is um we probably should look at incentivizing um Supply things for CRV USD automatically because their uh borrow rates for CRV USD are naturally very very high so is making uh a lot of a lot of Returns on that sorry but I think it's it's important to incentivize uh to incentivize the uh Supply sink for CRV USD so that um uh so that this system automatically grows again something uh which relates very very well to to the liquid talk half an hour ago so that's um was a good one um and another thing is we uh we can actually do some uh very good improvements to our crypto pools algorithms um the work is going on right now for for that but uh it sounds like they can be uh the crypto pool algorithms can be made uh work much better especially for uh for foreign exchange for let's say Euro to USD stable coin swaps so that's uh uh that's something uh we are most actively working on at the moment yeah and now looking at the big picture where do you see defi in a few years and what will it take to get defi to a trillion dollars in tvl and Beyond yeah I definitely see it uh in in those tvls in the future um but it's a good question how exactly we get there um my my opinion is that defi is very capable to replace the existing uh Financial system essentially um in which form is a good question because of course there are some um some you know regulations and whatever real world uh issues but I think uh technically it is very capable to do it and probably will do it because it solves uh like it it removes a lot of unnecessary intermediaries and might uh remove uh a lot of uh uh let's say a lot of uh issues which are connected to intermediaries like political issues uh defi is very neutral so it's it's just a tool which works autonomously and I think that this is why it it actually will replace uh a lot of uh like a big part of traditional Financial system but how exactly it will happen is a good question whether it will uh have some sort of centralized push or whether it will uh grow from uh from the wild from the D5 we have now but in any case uh like in any case it starts with the D5 we have now so that's why I'm more excited about defi than about uh anything else in crypto thank you so much Michael for joining us