Fixing DeFi with Open Data - James Corbett | Superchain

uh-huh here I am um so I'm James I'm the founder and CEO of super check and I wanted to talk to you about about today some problems we kind of see in the industry a little bit of History um around decentralized finance and kind of where we're going into the future and what we're doing to kind of fix defy with open data so a short history lesson but one from kind of what we think of the important perspectives so we kind of given an opinion based score so don't read too deeply into the 10 number this is more kind of uh just to represent the different experiences of the different parties and the different kind of features of defy so we've got a user score which is how easy it is for a user how nice how pleasant life is to use a particular tool or platform we've got liquidity provider score so this is someone who's bringing liquidity to defy who's someone who's maybe providing liquidity to unisoft V2 or V3 and how that experience is for that user and how sophisticated they need to be and then we have a decentralization so this is fairly obvious but you know how decentralized actually is this process in real life not what people say it is or the theory but the actual practical decentralization and then the efficiency of that market and efficiency I'm talking about is not like cod efficiency but it's efficiency of Finance markets so how fair is this price how much spread is there how close are you to like a real price and how much are you paying um compared to say a centralized exchange which often are very efficient because they have a lot of liquidity and they don't have the overhead of decentralization so we've got some time points along the bottom so unit swap V2 this was really really easy for users it was really easy for LP providers it wasn't very efficient as soon as things the way these things work means that they have a curve and when you come off of the curve it basically means that your your price changes quite dramatically so you can get a large amount of what's called slippage which means that people were tending to go more towards even though amms were useful people still using centralized exchanges for the predominant volume and in fact only recently have decentralized exchanges started exceeding the volumes of some centralized some centralized properties like coinbase the decentralization of unisot V2 was actually very good um in general it was a smart contract you could interact with it was very simple to understand for users developers and to integrate within the wider defy ecosystem but as I mentioned before the efficiency is fairly low so with unisot V3 we saw a big bump in efficiency so unisotv3 works in a more complex manner which means that you can basically get better prices so it's called concentrated liquidity and that means that you can as a user get really really good pricing but as a liquidity provider you're incurring a lot of complexity you need to manage all of your liquidity positions you need to choose where and why you pick that particular position now this means you need a level of sophistication you can't simply go on to Unisource website and add liquidity and hope that you'll make money there are risks from impermanent loss which you have to hedge and often non-centralize exchanges you would have to hedge and there are risks of simply not getting enough fees through picking to broader concentration or too narrower concentration and missing trades so either getting a small percentage of all trades or no percentage of um no trades so a large percentage of no trades so this actually reduces the decentralization score in our opinion because you now need to be able to read data from the chain you need to be able to build trading strategies simply to manage liquidity positions so it's not like univ2 which you don't just add and yes you have the the impermanent loss risk but you don't have the kind of more complex risks that uni V3 introduces um but you know as I mentioned before it is actually a better pricing model for the user so this is why you still see tvl increasing but most of that tvl is not from you and me it's not from retail it's from professional market makers these are people who have lots and lots of data at their disposal to make really good decisions about where they put that liquidity moving further on to more probably closer to recent times we have fragmentation so when I talk about fragmentation I'm talking about lots of different networks or mostly ethereum base which are copying unit swap V2 onto their chains be that something like pancake swap and uh also within chains you obviously have competition between between dexes and this further fragments liquidity now this has had a really big effect in terms of efficiency because the more liquidity you have in one place the better price you're going to have the less lipid you're going to have um so that's meant now that people have to build routing on top of all of these pools for you to get the best price the best price is not necessarily one pool of one pair on one particular protocol maybe you'll have to go through a few different protocols maybe even cross over chains so that fragmentation is only getting worse as people use incentives or funding to start a new blockchain and move a bunch of liquidity there and then the rewards run out and a bunch of liquidity moves back and all the time the people that are not speculating we're actually just trying to use D5 as a utility are kind of getting screwed and caught in the middle so the solution to this has been primarily to use things called aggregators as aggregators are actually a great tool which basically look at all the different uh decentralized exchanges and try and root your transaction in the most optimized way possible these are examples like kyber swap zero x one inch they all have these kind of Big Data warehouses full of all of the information about everyday centralized exchange and then you ask them please root me from usdc to arbitrim and they will find all those routes evaluate them for you price them for you simulate them and suggest how you should execute them and then you have the choice and they'll take a small fee for that service although I say a small fee in aggregate is actually quite a large fee that these Services accrue over time now this is a centralized service so what now you have to do with how fragmented liquidity is you have to use a centralized service a gatekeeper of data to be able to do an efficient swap you may be able to do a swap on unit swap V3 for example without data without context but you're not necessarily going to get the best price you're going to get some price but perhaps someone else is going to get something better or perhaps mevu or something like that in the process now we have an evolution which releases good competition so you know classic web 2 business we're going to have a bunch of aggregators all competing against one another and then we're going to aggregate the aggregators so this is kind of like Insurance comparison website but for crypto so promise decentralization you know I I personally want to be able to run a node on my own machine a light node I want to have the internet and I want to have a simple web browser perhaps an ipfs enabled web browser so that I can browse to the unisot website make my swap without relying on a central party perhaps I'm in a sanctioned regime or simply uh crypto businesses are being shut down by regulators and therefore can't provide their apis anymore so I actually have a really quick question for everybody here who has used the ipfs UI for uniswap not their default UI but the ipfs UI well done sir they've got one so if unit swap Labs disappears which is something that could in theory happen given regulatory Frameworks in the US they already have to filter tokens and adhere to sanctions so there's only one more step uh they have an ipfs UI which is fully decentralized until you go and use it and you realize it's not because it turns out if you block the uniswap website from the ipfs UI which is hosted somewhere else it the UI actually has to call out to uniswap to do the routing so you can't actually make a swap without that API now it falls back to some cached data that unisop have injected in the UI but it's very old and it doesn't have any new pools and it isn't really able to root because it isn't reflective of the current reality so this is a huge problem so suddenly someone like myself is thinking of course I can swap I can buy my particular asset I can manage my portfolio myself you can't you actually cannot do that not without relying on some other party um so you know this was actually a surprise to me when I was doing research for this talk that this is still such an unsolved problem so the gatekeeper problem comes in many forms you can look at it as RPC providers routers and aggregators and data warehouses these are all web 3-1 businesses we're going backwards we're going to we're going back to web 2. so all of these guys are kind of making money off something that's already open and already there and should already be available the data for these blockchains is complex and annoying believe me we know we build data indexes but it is there it is available so it's really disappointing to see the incentives are just such that Gatekeepers are are the natural progression of this and users suffer because users are going to get worse prices because someone has to take their cut it's not cheap to build a data warehouse to build the infrastructure of something like one inch so you know at the end of the day we're the ones suffering we're the ones exposed to centralization risk so how can we look at solving this particular problem and obviously with open data so this is a conceptual design for how a swap experience could be given open data for all liquidity pools um on univ2 V3 on Sushi swap on die the pasm which is a way of one for one exchanging usdc for Dai um and this particular scenario that simulated here is the usdc d-peg because my company super shown it was exposed to usdc that's what our investment was in so we had to manage that very late just before it depict to try and mitigate our losses from this event now that's not a nice position to be in when your company and your employees are relying on you and relying on usdc in this case to be there tomorrow so you can pay bills now obviously it's very sad it's also Silicon Valley Bank itself collapsed so you know startups in the US couldn't pay the bills either so what did we do well we had to sit on an aggregator in this case it was defy llamas swap aggregator which is about the best experience for swapping you can get at the moment and sit and wait and to try and watch when the best price was because it wasn't just us exiting usdc we're now competing against the entire Market of sophisticated players with terabytes and terabytes of data and we're trying to look at this aggregator view to press the execute button at the right time on the right platform to try and get the best money and you know when your size is size on amms you have problems so you have to really really watch it um because it and it it's just not a fun experience and it's saying I hope no one ever has to do in a panic and a rush because it's just not not fun and there's a huge what's called information asymmetry which means that other people know more than you and can basically use that against you so for example there's a lot of subtle differences between something like Cal swap and something like um a zero X so Xerox will quote you a price and you'll get that price within slippage but Cal swap may or may not execute if it's moving quickly so they could actually just say no we couldn't actually get that price within your slippage and we're just going to return you that or they'll do a partial fill there's lots of different rules and it's not immediately obvious that which button you should press even though you have the quota prices so to address this with open data we're going to give you the data so in this case we're going to give you the data in a web browser so instead of calling an API and saying I want to make this swap we'll give you the current liquidity positions across all pools so that you can root it yourself and we'll keep you up to date as new blocks come in and we can even do this across chains but for now we'll just look focusing on ethereum and this will allow you to visualize all things that are happening on the chain to have context when you swap so one of the big problems when you swap is you probably want to go on Dex tools info.uniswap try and see some historical pricing because the swap screen just shows you what's happening now so you can see on the screen we've got a variety of different routes that can be taken to go from usdc to usdt which is our example for exiting and they're color coded and they relate to the graph below so the graph below is historically what were these prices so they've all started off at one dollar which is great and they've all kind of taken a term for the worse so in this case we know during this time that the curve pool became really unbalanced the three pull which means that basically there was a lot more usdc than there should have been because everyone was running away so the price for usdc has to go down so you can see that's depict first in our example um the best route in this case was the top one um although that's probably not realistic in real life as it's uh two pools but it's it depends on your size and different different properties um but this gives you the ability and obviously this would be updating in real time to make a decision based on the context based on the history you'll know if you're getting a good price compared to 10 minutes ago an hour ago Etc this is really important because you know we're normal retail Traders we just use the tools we're given or we try and augment ourselves with some centralized services like Dax tools and this kind of tooling being integrated is really really important to give you context no professional Trader is going to trade without historical data in front of them without at least an idea of what the price was now 10 minutes ago yesterday so this kind of stuff is the default for normal Traders and yet we as retailer are expected to deal with Swap this number for this number best of luck right this is this is not a good experience so this is visualizing the historical price and the price you're getting now and in this case it's suggesting a particular route where you know we're going to basically lose just shy of two percent plus minus the gas fee um to exit into usdt now for some companies at that time that's kind of acceptable depends on your risk appetite but at least you can make the decisions around risk so this empowers users um and to be clear like when I say open data you own this data so you can do your historical analysis you can use this with data science tools you can use this with a local SQL interface you don't have to use this from like a website that someone else controls so you can go back to kind of a more pure decentralized experience so going back to my earlier slide super swap I mean of course it's going to be awesome so it's much more decentralized much better for usability starting to add efficiencies back into the market even for retail not just for sophisticated players and yeah this is uh what we think the future can look like with super chain we're very excited to you know keep building and this is one of the demos that will actually bring to life over the next year so thank you very much [Applause] so any questions thanks for the talk um do I trust the data or can I verify the data it will be fully verifiable um so when we push it out to you you can verify how that got to you where it came from which validators agreed on that and and you'll be able to understand in the web browser actually whether that is something you can trust or not so you don't have to trust our API or something like that does that come with a ZK proof no it's without zico thankfully um so how do I how do I verify it because it comes with what attached um so it's a little bit TBD but basically our because we deal with data in a certain way we can deterministically produce it so it can be easily produced by multiple parties so you can verify through simple hashing okay so I don't have to recompute the state no no um this actually basically is a replacement for like RPC and reading data you're no longer going to need to use RPC to get any kind of data and if you're a Dev that's had to use something like the graph you'll realize that you know you don't want a combination of RPC some other service your own centralized service this is a One-Stop shop for for this kind of defy data I was more concerned from you know if if you would look at prices from past blocks and you give me that data then I have to trust you that these were actually the prices actually historical prices yeah so you have and if I if I can't verify that then I can't bring them on chain but if I could verify them I could bring them on chain and use them yeah so um so there's a few questions wrapped up in there so kind of how are we guaranteeing the data so so this application runs in a web browser more like Integrity yeah yeah so this application runs the web browser which means it has limited capability um but effectively we have like a list of hashes of all of the the files that you would receive so there is some kind of prerequisite to have a verifiable thing like a blockchain or similar that um we're already synced up to you or to the protocol you're using um we're not 100 sure how that works like decentralization for us is like a very aggressive um but kind of slow process because we'll basically start by running it in multiple places at once and then will allow people to run it themselves and then we'll start to add incentives and mechanisms in there as well but it is kind of a long journey but we generally avoided ZK through a deterministic production of data so you can grab it from you can basically know what the hash of it should be from multiple different places and then download it from one guy and you'll know that you got that so it's more like a simple consensus mechanism on a file basis but this is all TBD it may work slightly differently by the time we get there so don't don't hold me to this but that's the current intention um thanks any other questions of life important because I know that so the question was is there a live demo um soon we will build this as like one of the use cases of our data we're basically building right now um all of this kind of infrastructure we're a rust development shop we're building everything from the ground up we don't use like open source databases we're building our own stuff it's it's taking some time which is sad from a market perspective but it means the product will be like way better and have a lot more longevity but soon watch this space go to Twitter Etc and then you will be the first to know yeah another question about so you mentioned the graph for instance uh there are some issues for instance with some the graph data sources because they provide like incorrect data for instance we tried once to rely upon the graph data from our version 2 and yeah there were a lot of different problems with that so how can we like I don't know rely up on you having in mind that like I don't know in a year everything won't be broken and so the proof will be so we how how this proof will be maintained so that we can like rely on you sure um so most blockchain networks have longevity because of incentive mechanisms um so our plan and this more relates to I'm going to give you a tokenomic style answer and then I'll give you a more technical answer if that's okay so tokenomics wise we're building a product that people will use prior to it being decentralized and then we're decentralizing that product so we will have and then we will hand that Revenue will then move to the decentralized product that means that the network will have a base layer of usage so we're not just going to deploy some core L1 which no one uses there's already users paying for data there's already incentive there so that's the first reason so validators are incentivized to run it we're incentivized to maintain it people can build customizable data in our platform we call that toolboxing and that means that uh the incentives are aligned correctly for that Network to continue to exist now going back to more of the kind of Maintenance style things like especially in relation to to some of our competitors um there's a whole host of reasons why those things happen um they've had to make a lot of technical choices that we've seen like for example being able to provide incorrect data which to be fair to the graph is a fairly rare occurrence but obviously when it happens it's really bad for users now they have to rely on something called fraud proofs whereas we don't send non-verifiable data so we can't answer you incorrectly basically unless our entire network is wrong somehow which you know fundamentally could happen like any network could make a wrong thing and break um but obviously we think that's highly unlikely it's much more likely we'd fully halt than just send you like some wrong stuff um so yeah welcome any other questions nope well thank you very very much for your attention [Applause]