Meet the Nation - RAI & Reflexer Labs | CEO Stefan Ionescu
15 Jul 2020 · 20:51
Reflexer Labs CEO Stefan Ionescu joins Meet the Nation to introduce RAI, a non-pegged stable asset designed as a more trustless alternative to DAI. He walks through how RAI dampens price volatility without targeting a fixed peg, comparing its architecture to MakerDAO's and detailing where the two systems diverge.
Speakers
Topics
Transcript
all right guys i'm here with stefan yanescu who is the ceo of uh reflexer labs this is a project that uh we saw kind of circulate around twitter a while ago uh and it sounds like it's in the early stages of development and so we're bringing stefan on to meet the nation which is this brand new format that we are offering to the bankless world to the bankless community into greater ethereum to help projects get introduced right and so we're going to kind of go through what rye is what reflex or labs is and kind of where it fits in the grand scheme of things so staphon if you want to give yourself a little bit of a quick introduction in how you came to be working on working at reflexer that'd be great
sure so i started my stablecoin research last year in september when i started to be a finance grantee so i worked for a couple of months with binance to develop kind of a stable coin for them but then i in march i saw that the design that dye has i wanted to kind of get inspiration from them it's not that resilient when it comes to shocks in the market so i remember that they didn't implement the rfm in dye but yeah the thing is with the rfm is that it changes and stable coin it kind of changes the asset type from a stable coin to something that follows the price of its collateral but in a dampened way and that's something it's is a bit more kind of a good fit as a collateral because it kind
of inherits the properties of its collateral so it kind of can do what its platform can do but in a dampened way okay so you saw this um mechanism and you saw that you could apply it in a more general way and so that's kind of what reflexer labs is is that right yeah yeah so basically we're kind of developing a framework we're working on a forking maker and developing a framework for another type of asset that we call a reflex bond and we hope that this asset is complementary to anything else in b5 from stable coins to synthetics to anything else that needs collateral okay so so if i can understand this correctly the way that um rye works or and rye is a generalized
architecture that works for any asset but it produces a new asset based off of the original asset that moves in similar in the same um direction in price as the old asset but it moves less it just is the same but less for any asset not that right not really less so not we're not really removing uh volatility we're kind of just lengthening the duration of the price move so we're mirroring the exact price move but it's kind of over a longer period of time if the price moving ether is sustained going kind of 50 up or down for a couple of months then the bond will also go 50 up or down but if it only goes kind of for one second one minute one hour it will
probably won't move much okay okay that's that's news to me okay so like if ether goes from one hundred dollars to one thousand dollars the rye version of ether would start at a hundred dollars and then slowly move to a thousand dollars yeah so we're not we won't really start at a high value so we'll probably start with let's try at i don't know five bucks two bucks something like that and rye will move kind of by the same percentage as ether or whatever gradual sustains it but again kind of over a longer period of time okay cool so so the cool things about this are that it's generalizable for any asset and it offers a different volatility profile for any asset right and so it allows people to um retain exposure to
the asset of their choice but perhaps allows them to use that their asset of their choice in a different way than what their the actual true peer asset would have been yeah i mean the the point is you cannot really describe right without describing the system so if you want kind of a gist of what right is it's basically fixed or capped loan a fixer loans with fixed or capped interest rates but with variable borrowing power so even if the price of your platform stays the same for 10 years the borrowing power that you have on your loan will change according to kind of a variable that's set by the system itself okay you're going to have to unpack that
a little bit more um so so alone so there are there multiple parties involved with the creation of rye for an asset it's exactly like maker so we're basically getting maker but we're improving on top of the system we're adding stuff um we still have stability fees so you still pay a fee per second fee for your cdp but in our case we can kind of fix that because we don't want to move multiple levers at once it's really hard to make a controller that's autonomous and can also work with multiple levers so we can kind of fix the stability we can use the stability feed to kind of like pay for oracles or to have a buffer as insurance in case some liquidation doesn't
completely have made the system solvent we can use that insurance buffer and pay to bring this to bring the system back into kind of equilibrium um but we also have this second lever that we call a redemption rate and this is this lever is kind of like an interest rate it's still applied per second and it first every second it kind of changes your borrowing power it can make you um you can make kind of your cvp borrow more or less it kind of gives you basically it reprices the bond internally it basically system says okay the bond is x dollars right now in one second from now it will be more or less if it's less it's basically cheaper
compared to one unit of collateral in this case ether so you can generate more of them because it's cheaper on the other hand if it's more expensive you can generate less bonds and you might even need to pay back your loan your cdp so you don't get liquidated okay so so let's go through the process of is it minting rye that is the right word because that's how it is you can use the exact kind of terminology as maker okay bolts i prefer cdp's okay i also kind of prefer cdp's in a way okay so let's let's take ether as the example because why not that seems like the right place to start so you deposit ether into a a reflexer cdp
is that is that what we're calling it reflex or cdp uh it's kind of the framework is kind of we prefer generalized ethereum bonds so it's kind of geb gem okay jeb you deposit ether uh and then you have collateral inside of a system and then you can mint rye and rye isn't it's not a stable coin it's not pegged to a dollar it's and is it self-referential as in it discovers its price from ether itself it's not rep it discovers its price in relation to itself actually because when we start the system we say okay you are five bucks uh and from there according to market forces if if the system needs
to compel users to generate less bonds it will reprice up and if it needs to compel users to generate more bonds it will reprice down so it makes it cheaper compared to eat okay okay and when you say it reprices it you're we're talking about repricing the rye the asset that is pre the synthetic asset that is produced is it a synthetic asset is that right you could say so okay yeah it's kind of okay yeah it's a synthetic that mirrors its collateral okay okay so you we've deposited one eth uh and eth is a thousand dollars uh and we uh then mint new rye and rye is priced at some lesser amounts like a
hundred dollars or something yeah uh and then let's say for example ether goes um sharply from one thousand dollars to seven hundred dollars it loses thirty percent uh what are the mechanisms involved that uh contribute to the repricing of the rye that i've minted which i still own because i haven't sent them away i have i have both my eth in the uh in the uh c cdp in the cdp i still have my in the cdp i have the rye that's based off of eth that i've minted in my wallet ether started at a thousand dollars and just dropped to three hundred dollars what happens in the background to rye and how does rye discover its new price
so this case is very similar to black thursday in this case people initially at least would rush to get right to pay back their cdps because they don't want to get liquidated also on the other hand you have the keepers the scripts that protect the system and kind of liquidate tdps they will want liquidity to participate in auctions and basically buy the stress collateral from the system so initially right might actually go up in value exactly like dye did a couple of months ago but then uh the system will given that right the market price will go above what the target the target that we want the system will try to compel users to generate more bonds by repricing down so repressing um what right right okay
initially so inside the system it would say okay every second you'll get cheaper and cheaper compared to one unit of collateral okay okay so this is right yeah so there's two prices there is the actual price of rye on the secondary market which and then there's also the price that uh the system targets and those are different right yeah and so how do we price is the actual price the target price we call it the redemption price because it's kind of the price at which you redeem your collateral from a cdp okay okay and so over this if the system works over time the market price will always target the um the target price yeah right okay uh so let's go through
that one more time um if uh if the rye price is above the target price where are the incentives that bring that that actual secondary market price back into the target price the incentive is that basically i can allow you to generate more of these bonds because they are cheaper inside the system so rescue will say okay now i can lever even more so or and over time you'll you'll be able to leverage more and more so that i can say okay why not do it that's the idea and on the other hand if you reprice up and if you reprice up the rye will get more expensive compared to east your c ratio will start to go down
and at some point it might go so kind of it might go down so much that you might be close to liquidation so you can say okay maybe i can uh deal yeah i can get more leverage and get less leverage sorry and kind of close my position and kind of get that right from the market repay my loan and this way kind of um there will be less right on market okay so the uh incentive if i'm hearing this correctly the incentive incentive is baked into the cost of capital and so if if if rye is priced too high there will be less cost of capital to mint new ones to add to the
market and then the reverse is true uh to get it pulled off of the secondary market yeah yeah okay yeah so uh i mean i think this is a very brief question but like this isn't just for ether this is for anything can you kind of talk about or comment on how any this will work for any asset um it's very easy to do it when you have only one asset because you can say oh you can kind of make the link okay this is ether this is right so rai is basically ether but just less volatile ether you can also have a basket that kind of backs rye and at that point you will basically get kind of you kind of dampen the average of the volatility in the basket according to how the basket is composed
okay so so um there's one rye rather than one rye per asset right so i guess this model is generalizable but the way that you guys are trying to actually build it there's only going to be one rye right uh yes and no i mean no no one stops us from deploying multiple systems with multiple rye versions but in this case the first asset will just be ether backed and it's one ri for all the classroom in the system okay so this is kind of like maker dao's single collateral dye where they just use ether now we have single single collateral rye and that's just going to be ether uh do you guys have plans to spin this out into a multi-collateral
ride version yes and no um no because i hate control so i'm a developer i generally i hate management um right now the plan is technically kind of built right make it work uh after six to 18 months post mainnet kind of remove 95 of our control over the system just leave it work uh then maybe build a second system on top of ride to show that hey this actually works as collateral so you can use it in your own stuff and then shut down the company um and apart from kind of shutting down we want to have a dao that right now is called metacoin that's kind of the
distinction so reflexer is the company metacoin is the tao that we want to have and leave that dowel kind of live on on its own um you might need to ask me again you might need to ask this question again i mean two or three years after you build this and see if i actually want to build another system that's backed by chain link bats or whatever we'll find fair enough fair enough okay so if i remember this correctly uh rye was originally announced in the heat of this debate with maker dao as to whether maker dao should have centralized assets inside of its inside of collateral backing die and there has been this strong movement which i would say is has never died and is actually perhaps
coming back with usdc being uh blacklisted from somebody's wallet and ustc is in a collateral for backing die which a lot of people sound talk about as a risk so rye is supposed to be this or could be this uh government or governance reduced governance minimized system for producing stability on ethereum can you talk about that the motivations to produce that yes i can actually link it to your article about the protocol sync uh the idea with this so the idea with rye is that we believe that we can bring even more capital in the system compared to single class will die simply because we remove most of our control and if we automate
as much as possible so you can kind of make plans on top of the protocol i cannot call me stefan or amin cannot come and say okay now we're adding link or now we're adding gold you always know what to expect from the system and that's the idea you can bring more classroom now because you have more security very cool very cool so uh yeah the idea is to have to strip away so a lot of people will point at maker dow and and claim there's a bunch of weaknesses the centralized assets uh that back die is a weakness the centralization of the team itself is a weakness and so what you guys have done is you guys have compromised on pure stability right so
rye is not stable it's stable er it's stable ish and then it's but also there's less risk with both the collateral and the governance which in cell in a way is stability just in a different aspect of the system it's it's removing volatility in governance and volatility in centralization of assets and you're allowing that volatility to be expressed in rye a little bit in comparison to die is this all right yeah fantastic yeah it's pretty close i mean even if you check if you look at die now it's kind of starting to look uh kind of like a reflexive asset and the community doesn't really like that they just want
kind of a dollar peg stable coin and that's it so that's why i say hey this can actually be complimentary to die we can maybe even help them but kind of come back to the initial question um nothing stops us from creating a system later on that's backed by mortgages but this system needs to be separate from rye with rye we kind of built an mvp that we think will kind of slide to the bottom of the protocol sync we can help every d5 protocol out there and with something else like mortgages we can make maybe focus more on the fact that we are kind of a credit facility we can we allow you to leverage your assets more than kind of focus on the platform side
okay so like kind of in in a way where you know unit swap v2 came out but you know swap v1 still exists and still can be used excuse me there's a motorcycle outside and so in the same way there's going to be original rye and that's always going to be there because that it's that it's part of its design it's supposed to be very dense it's supposed to drop down to the bottom of the protocol sink but yeah but you guys are using rye as like a proof of concept saying hey this works it still works it's working now it's working today and we're going to build a whole entire new protocol a new company on top of this model because we know that it works yeah i mean again it everything depends on whether in three years from now i'll actually want to
switch from being kind of a pure dev to being the corporate manager that you see everywhere i don't really like that thought right now but i'll see are there any um like attack vectors or or chinks or weaknesses that we haven't discussed yet yeah and the first i'm very open about the fact that in the first six to 18 months we'll have kind of complete control over the system and the reason is that this is extremely experimental technology it can blow up at any time and i would rather save users than just decentralized from day one um leaving that aside anyone exactly like maker anyone can kind of kind of sacrifice their protocol tokens to shut down the system at any
time so they need to sacrifice their kind of capital to shut down this system whatever kind of ride or whatever the second system that we might deploy that would be kind of the biggest problem with this mechanism okay are there is there a mkr version uh inside of the reflexer system that's like a there's a capital token yeah yeah okay cool all right well there's potential there definitely in this current current environment many people kind of call it a governance token but i think that's not true because we again initially will have control but after that i just want to remove 95 and bound the remaining five percent so the
governance part won't be that important maybe we should call it i don't know a protocol token or protective token or whatever because you can mint this system can mint the token when it needs to recapitalize the system right so it's just like mkr but without with less governance right and so yeah this is this isn't a good name either but i would call it a value capture slash risk token so you get the value and you also take the risk yeah yeah kind of kind of yeah we'll we'll pin down these names better and better over time okay uh stefan what can people expect from the next six to 18 months of development from the reflexer team and and if they want to
learn more about reflex or labs where should they go so you have a metacoin project on twitter that's kind of our dowel twitter uh we also have a discord i can share that with your listeners if you want uh i think that's we also have a telegram group but we want to move everyone from telegram to discord so this is kind of it and you can expect interesting news even this month looking forward to it all of those things i will put in the description of the youtube video uh so you guys can find those things there uh stefan thanks for coming on to the very first meet the nation and thank you for meeting the nation thanks david