On The Brink with Castle Island - Aditya Palepu and Frédéric Fortier (DerivaDEX) on aspiring to a fully decentralized exchange (EP.119)

Episode Date: August 31, 2020

Aditya Palepu and Frédéric Fortier, the cofounders of DerivaDEX, join the show to talk about DEXes and prospects for their future development, and to introduce their own project. In this episode:  How Adi and Fred chose to build a Dex Why traders tend to love Dexes despite their shortcomings Dexes versus noncustodial exchanges Adi and Fred's Dex taxonomy and where they fit into that How Derivadex plans to decentralize the operation of their orderbook Are today's AMM-style Dexes sufficient? How the Derivadex team is thinking about the shape of the liquidity mining distribution curve How the distribution curve actually affects which users end up owning a piece of the DAO Why Derivadex struck a more pragmatic approach than certain other Dexes – and how they're going after CeFi users Why the team opted for USDT as the default collateral type The lessons Derivadex took from the MakerDAO exploit Their approach to capitalizing an insurance fund For more, see this introduction to DerivaDEX.

Hello and welcome back to On the Brink. It really feels like we're on the brink of something here. Anybody else getting that feeling? Anyway, we've historically neglected discussions of Dex's on this podcast. That's probably going to have to change a little bit here. Just recently, Uniswap exceeded Coinbase's daily volume. So there is a bit of a changing of the guard here. Or at least a recognition that Dexes are a mainstream way to trade tokens,

and they're not going away anytime soon. The most popular style of decks right now is the automated market maker where you pool liquidity. There is no order book. These kind of come with their own unique set of challenges, but they're also fairly simple to use, and people tend to like them. We wanted to better understand the decks landscape and try and understand the direction. They're going. So we brought on Aditya Pilepoo and Frederick Fortier from Derivodex.

Derivodex is a decentralized. derivatives exchange built on top of Ethereum, backed by Polychain, Dragonfly, Electric Capital, CMS holding, three arrows capital, really who's who of the biggest investors in Defi. And their objective is to really replicate the functions and feel of a centralized exchange in a way that is non-custodial and mitigates the risk of holding your collateral with a third party. This is pretty different from the uniswops of the world. There is no automated market maker. And it's going to be a considerable challenge to decentralize or outsource the functions of a centralized exchange. So we talk through where they sit in that kind of taxonomy of Dex's.

Why traders might prefer Dex's to centralized exchanges and how they plan to architect this system. I learned a lot about the Dex ecosystem in this episode, and I think you might as well. Let's jump in. Brought down by bad mortgage investments, Lehman, which has 25,000 employees will be liquidated. The federal government loans American International Group, AIG, $85 billion. This is a different kind of market, and the Fed is asleep. The federal government is stepping it to stabilize Fannie Mae and Freddie Mac, the two mortgage giants that have been threatened by the housing crisis.

The Bank of England has pumped 75 billion pounds more to Britain's ailing economy with a new round of quantitative easing. You print a couple trillion dollars, and all of a sudden people start to worry. So out of this worry, we have something called the Bitcoin. Bitcoin. So Fred and Adi, thanks so much for joining. and welcome to the show. Thanks for having us, Nick.

Yeah, it's my pleasure. I think what you are working on is something pretty novel, which as far as I know, doesn't really exist in the industry today. So I'm pretty excited to dig in. It's a very ambitious vision, and I kind of want to decompose it

into its constituent parts. Maybe before we start, can you both tell me about your crypto origin story, so to speak, crypto and then how you chose to build a decentralized exchange. Yeah, definitely. So I can kick things off. So I'm a DTIA, first time podcaster here, by the way. So thanks a lot for having us on, Nick. All right. Yeah, let's do it. Yeah, I'm from the East Coast. I'm from the DC area on the Virginia side. I graduated from Duke back in 2013 with the ECE comm side degree. I then spent the

majority of my career at DRW, commonly in the crypto world. You've probably heard of the Cumberland arm they have as a quant trader, trading pretty much everything under the sun there. I left in around 2017, 2018, that time frame, and I started Algo trading cryptos on my own. And it was a really fun, super fast moving volatile market, you know, with ample opportunities, of course, with all types of strategies. But amidst the trading, I got really into the underlying tech itself. So I started working at a project called Enigma as a blockchain engineer working on privacy preserving smart contracts and that was my first foray into the crypto space, certainly on the technical development side of things. And I had the wonderful opportunity to meet and work very closely with Fred,

who is obviously my co-founder here as well. Yeah, hi. My name is Frederick Fortier. I'm Canadian, but I'm in the Bay Area at the moment. I have degrees in computer science and finance. So I've sort of always been interested in the intersection of both these things. For over a decade, I worked as a software consultant for financial institutions, bank, investment banks, some insurance companies, essentially building different internal systems for them. At the same time, I've been running my own algorithmic trading fund because that was always a passion of mine, trading mostly biotech products. And in the last five years or so, I started adding cryptocurrencies to that. And that's how I came to also join Enigma because at the time they were

working on an algorithmic trading platform, which used a similar stack as mine. And we basically joined forces and I joined them as an employee. And that's when I started to do more also blockchain protocol development. And that's where I'm at ID. And what was the genesis of the idea to build Derivodex? Yeah, that's a great question. So, you know, exchanges and markets are integral to the world around us, bartering and trading have existed in one form or another and have underpin society's evolution since the dawn of time. So the significance of the problem space is certainly not lost upon us. But, you know, in particular, one of the great ironies that we are all aware probably in the crypto space is that you have these theoretically decentralized assets that are

almost exclusively transacted upon centralized venues. And additionally, right, yeah. So like if you look at traders also in what they're looking for. They're looking to speculate on and hedge with a wide array of products in a secure, censorship resistant, and efficient manner. But unfortunately, there's kind of two choices that they have right now. And this is a little overly reductive, but they can opt for either centralized venues that, yes, have a wide product offering an efficient UI and U.S. strong liquidity profile. But they're very weak when it comes to security of funds, opaque, dare I say, even shady often business practices and high regulatory requirements and really kind of unfathomable barriers to entry.

And then on the other hand, you have decentralized exchanges that paint the opposite story, right? So they're theoretically very secure. They have a low barrier to entry, but they have a very poor product offering, very poor liquidity, and an unforgivable U.S. So Fred and I, as traders by trade, no pun intended, you know, we empathize with the end-use. user greatly. And so we aim to solve and remove this choice that traders are currently left with and instead provide them, you know, the safe, performant home that they deserve and need. Well, it's certainly a grand vision. And, you know, I think it's interesting that you, you know,

have decks in your name. There's kind of been, I'm sure you've noticed as a move to like, move away from the Dex jargon and refer to these things as non-custodial exchanges, take, being mindful of the fact that not everything is decentralized, but your proposal is actually maybe more ambitious, you actually are trying to decentralize the functions of an exchange. Is that kind of a fair way to put it? Yeah, like some of these other sort of texes, at least at the start, our technology looks a lot like off-chain execution and on-chain settlement. we are executing, we're building the matching engine, price feed, and other components of the execution of orders inside of what we call operators.

And initially, there's going to be few operators in the network. It's going to be a bit more centralization in that area. But the settlement of orders is done on chain. And the way that we're building the operator, our purpose is to develop an infrastructure that can easily be decentralized as time goes so that we can easily add operators to that mesh, growing mesh of operators to achieve full decentralization. And in addition to that, like in addition to how we're building our technology to be decentralized, we're also considering the business aspect of it, which is sometimes not present in, some of these other dexes where we're going to run Derivodex as a DAO from day one. It's actually called Derivada. And that business aspect of decentralization is a priority for us,

and we think is very important not to have a single entity that can be targeted and then ultimately shut down the exchange, where all of the strength of decentralized code is essentially useless if the business entity doesn't also have decentralized attributes. The component that is done by centralized exchanges, you guys are effectively planning to have that be done by just a decentralized network of operators. Is that correct? Yeah, that's correct. Have there been instances of this before? Is this kind of a novel thing? Well, I think maybe you can talk about a little bit like the story on Dex's currently. Obviously, there are decentralized exchanges that are being decentralized, but they all use maybe different attributes to achieve this as what we're using.

So maybe we can talk about the taxonomy of Dex's then. So like you have kind of fully decentralized in theory ones like Uniswap, which just exists effectively as smart contracts that are permissionless and you can opt into them. and then you have kind of hybrid models. So maybe talk me through the way that you guys think about the established Dex marketplace and then where you fit into that or how you differentiate. Yeah, great, great question. So when we think about the characteristics of successful Dexas, you know, these include several different things.

So security, transparency, privacy and onboarding ease, regulatory side of the equation certainly plays into this, you know, innovative components. possibility, performance, anti-fragility, and liquidity, and also democratic involvement of users and participants. So I'll dive into each of those and say a quick note on each of those points and why they're important and how we fit into this. But before I say that, I think it's worth mentioning that, you know, many of these are inherent characteristics. As I mentioned them, you probably think that these are characteristics that are pertain to any successful exchange, whether it's centralized or decentralized, right?

I think Defi has this knee-jerk reaction, almost compulsive need to reinvent every wheel possible. But there's a lot of C-Fi that actually has done quite well. And so that's one thing that I think is worth noting that we pretty early on realized. And then also, I don't believe that we've really seen a quote-unquote successful decks to date, which is why we're so excited about what we're building. Certainly no solution checks all of the boxes I mentioned, or quite frankly, even several. of these in a meaningful way. So to elaborate on these, you know, security, right, like non-custody of funds is very significant.

Hacks like Mount Gawks, coin check, you know, where $500 million were stolen, these are real. And users should have their funds, you know, safely in their control. Transparency, right? The code dictating and exchanges rule should be out there and open. There should be no fuzziness around what's going on. Exchanges are notoriously opaque with how they conduct matching, as you pointed out, liquidations or what they're doing with users funds. Privacy and onboarding. So we're, you know, permissionless K-Y-C experience. And so this allows users to participate without fear of identity

theft, as we've seen, can be a huge problem with data breaches on like Bitmex, for example, with emails being released, et cetera. And it's also a material UX improvement over centralized avenues. And then on the regulatory side of things, you know, a truly decentralized platform can operate freely and robustly, whereas CFI, you know, runs the risk of adverse jurisdictional authorities, halting or undermining the product. And I think to more specifically talk about some of the products that you would outline, two areas that are really important to know about how we fall into this equation. So one is on the performance side of things. A big hindrance right now in DFI is the on-chain price oracles that, you know, Maker or Chain Link utilizes. And so most are all,

all DFI platforms use some form of these price feeds. And that results to very, very slow liquidations. This has a major impact on traders and on an exchange's solvency. So we're excited to bring real-time price feeds that enable the operational performance that you find on DFI platforms. And then the other point I want to mention here as well, and certainly we can dive into these, is on anti-fragility and liquidity. Right.

So this is a particularly important point. Most of the systems that are driving interest in Defi right now employ what's called automated market makers or more simply abbreviated AMMs. These are essentially algorithmic agents or provisions of liquidity in the markets. So things like Uniswap, as you had mentioned, they employ this and essentially set rules in place that define the relationship between quantities of tokens. And so while this is an interesting technology enabled by the decentralized tech around us, the issues with these approaches. are not to be taken lightly. For one, slippage tends to be quite large with even reasonably sized orders. So the cost of entering a position can be dramatically, can dramatically hurt a trader. And so as such, the types of strategies that you can reasonably employ in these systems are

necessarily naive, actually. And additionally, liquidity providers can get, quite frankly, fleeced with large market moves. You may have heard the term impermanent loss, for example, thrown around. So all of this put together. means that these systems actually don't perform well with large orders and large market moves to hallmarks of crypto trading. So we're excited to employ the familiar capital efficient order book with novel ways to bootstrap liquidity with tight spread. So now on top of that, when you start tacking in the Dow that Fred was alluding to earlier, you basically create this very performant, usable exchange that also now allows users to participate in the governance and in the

evolution of the exchange, which is an extremely powerful proposition for them. So there's a lot of pieces there, and I want to work through it. So maybe for specificity's sake, it almost sounds too good to be true, not to make any cast any aspersions or anything. So how does the order book work in a practical way? You know, like who effectively is managing it? And how does price formation occur on the exchange? Yeah, so we're using an open order book, so similar to what you would see on centralized exchanges and not an automated market maker style of orderbook.

The order book is hosted by the operator. So it's very similar to some of the dexes that have, again, this off-chain execution and on-chain settlement. So each operator has its own order book. The key difference with some of these other exchanges is that the operators sort of work in concert and submit essentially batches of match order and liquidations to the smart contract.

And the smart contract verifies that each operator has submitted the same result. So that's a little bit different from Citi0X model with relays, where each relayer also has a copy of its own orderbook and so forth, but it's expected that relators with Open Order Book would share their liquidity across each other. Well, for us, it's sort of assumed built into the protocol that all of the operators are going to share liquidity with each other, and we use a non-chain consensus to make sure that all operators behave the same way.

And the trick for doing that is just to make sure that the operators all have the same inputs so that all operators receive all of the orders and all of the price feed updates. that are relevant for matching the orders that they are responsible for matching. And again, at the beginning, it's going to, just admittedly, it's going to look a lot like some of these dexes that have centralized matching because we're going to have maybe one or two operators, and we're going to slowly ramp that up. We're doing this just to be cautious at the beginning,

to be able to work out any kink very quickly. But in the first few months, we're going to be going to, start adding operators and we're quickly going to get to a point where there's going to be a network or a mesh of operators that is across the globe and that removing a single operator will not affect the liveliness of the exchange itself. I can also talk about the price feed a little bit if you want. Yeah, just to dig into the operator quickly. So who do you expect to become an operator and what is the actual process for becoming eligible to be one and and starting to help run these order books? Yeah, good question. So it goes back to our Dow Derivada,

which we haven't described in a lot of detail so far, but essentially we are operating the exchange using a on-chain governance system. And that governance system for now, at least at the beginning is going to work by vote, similar to, for example, the compound governance mechanism, some of these other defyre protocols as well. And for operators, they will have to be whitelisted by the DAO. So when a new operator wants to join the network, there's going to be a vote as to whether that operator is essentially trusted by the network. This is done just as a preventive measure. There are also mechanism like, for example, this M of N consensus that we do on-chain to prevent even a trusted operator from malicious behavior. But operators would be whitel.

This is kind of similar to how MakerDAO kind of whitel kind of whitelist their price oracles, for example. And when an operator is whitelisted, then the operator has to submit a bond, which is another kind of backstop security measure, right, where if an operator sort of misbehavior, not only it's going to be blacklisted and jailed, but also his bond would be revoked if the DAO can, in a dispute resolution process, if the DAO can prove that the operator acted maliciously. So he's get whitelisted, submits a bond, then he has to run the, you know, basically the operator software stack, which for, at least for the price feed component of it, requires trusted hardware. So essentially we're using trusted hardware to establish a TLS tunnel essentially between the exchanges where we're getting our price information from. So we're getting our price information from an array of different centralized exchanges, similar to perhaps how bidmex does it, but maybe a little bit more open where it's verifiable which exchange we're getting the price from.

And the operator has to do this using trusted hardware so that we can establishes. a TLS tunnel between these data sources and the trusted enclave that the operator is running. And then ultimately that channel between the data sources and the operator gets signed by the operator using what's called the remote attestation feature of Intel SGX, which is the trusted hardware vendor that we're using. So that's how the price feed essentially works. So to be an operator, you need to have this sort of hardware available and do it and run it. And there are two types of operator roles, if you will, like an operator could fulfill both. You could either be a price feed operator or a settlement operator or both. And the difference maybe it's a bit obvious, but the price rate operator is the one that connects to these data sources, fetch price feeds,

submits it to all of the other operators, and the settlement operator kind of receives this, so it receives a stream of price updates and a stream of orders, and does the matching and liquidation business logic. So assuming that all of the operators that have a certain, say, market in their scope receive the same data as input, then they should all arrive at the same liquidation batches and the same batches of filled orders

that are then being verified on chain using a consensus between multiple operators. And how do you actually hold them accountable if they misbehave or they don't behave the way that you expect them to? Yeah, that's a good question. So basically, it's sort of three different layers that I just mentioned that sort of guarantee that operators will behave properly. And we think this is very actually, we're very careful about this. So the first layer might be like the trust in each operator where operators are already kind of voted in the dial. This is not a very robust layer, we know, but we think that it eliminates some obvious

bad actor from getting into the system. The second layer of trust is this trusted execution environment where operators are being shackled by the trust execution environment and can only execute a specific code. That's a very robust layer, but we're aware that from a perception standpoint and maybe even from a real pragmatic standpoint, there have been some attacks against this type of trusted hardware. Obviously, we're using the latest patch software that doesn't have any known attack, but we don't know the future. We don't know if there are going to be future attacks against it. So we have also other layers of security. One is the consensus, where not only each operator needs to run trusted hardware, but also we establish a consensus across

multiple operators, some kind of n-of-m consensus. So let's say, for example, if we have 20 operators in the system, we might say, well, we would require five or ten of these operators to submit the same like batch data in order to accept that batch and settle it on chain. Another layer is that we are very aggressive in terms of what we're validating on chain. For example, when users place order on the exchange, they have to sign an order intent using EIP 712, kind of similar to how zero X orders work, if you're familiar with that. And instead of just verifying that signature at the operator, level and kind of rely on that already very kind of robust consensus and plus don cleave and all of these layers of security we actually pass these orders through to the

to the to the to the to the smart contract and verify the signatures on chain as well and the final backstop is this bond right where if an operator submits an incorrect batch of data then there's a dispute process that's going to be ongoing in the dial and to essentially like verify if I determine what happened. And if the operator is determined to have behaved maliciously, then a bond is going to be revoked. So all of these layers of security on top of each other are there to minimize the odds of an attack to what we take is near zero.

Is there either the possibility of or the implication that there's a conflict of interest, that there's a large trader on your platform, which is also an operator? Or do you think that those safety measures you talk about are sufficient to defray that. Well, that's one of the reason why, like, we, like, our role here is just to incubate the protocol. So, like, operate the exchange at first. Yeah, perhaps be the first operator or be very

closely related to who, to the first operator. But we're trying to take ourselves out of the picture as soon as we can. So maybe in the early days, if you don't trust us, you can make this assumption. Perhaps you're going to trade on a different exchange. But if, if, but at the same time, very quickly, we're going to have a very diverse array of operator that are located across the world. And then at that point, it's going to be a matter of, you know, sort of trusting the implementation, which we can kind of prove the characteristics, right? Like, you're going to be able to look at our code, look at how we're implementing it. So, like, kind of trust the security attributes that we're deploying. And also trust the fact that new operators have to

be allowed by the Dow and that the tokens are distributed enough so that when there's a vote, it's actually a real democratic vote and not a vote from like one central entity. And yeah, and that's that's pretty much how we're going to tackle that issue of removing ourselves from the equation. And also to your point, Nick, like it does this ecosystem that Fred outlines solves for the, you know, it mitigates the risk of those type of front running attacks that are generally worrisome in certainly the Dex context, but even in the C-5 context as well. So, Adi, earlier you mentioned some ways that Dex's are, you know, in your opinion, superior to centralized exchanges. Honestly, for the most part, pretty compelling to me in particular,

the non-custodial nature, obviously. And I think, you know, there's been a dawning awakening that, A lot of people said like 2018 would be the year of Dexas, but 2020 has really been the year that Dexas have actually taken off materially. Yeah. And we've even seen token launches on Dexas. People seem to be tolerating the slippage on these AMM systems. Yep. For whatever reason.

Yep. So is it a possible case that Dexes are kind of worse as better, at least the current crop, and that the ones that exist now actually, you know, pass some. threshold of acceptability and liquidity such that they would become entrenched? Is that kind of a possibility in your mind? Yeah. The question of why the interest is really peaking now or certainly starting to is really interesting. And it's something that I get a lot from obviously building in this space, certainly over the past couple weeks. I think a lot of it is around timing. I'm not sure necessarily that a lot of the core fundamentals, you know, like non-custody of

funds that you mentioned, permission lend systems. I don't know that a lot of these principles that are pretty core to the ethos of blockchain tech that they've changed all too much, but projects that maybe spawned during the early days of ETH, for example, I think they were probably just a little too early to some extent. And you're also seeing just the natural progression of financial systems play out, albeit at a very hyper-accelerated pace in the decentralized space. So like Bitcoin trading, for example, you know, started off on centralized spot exchanges like Mount Gawks or Coinbase back in the early 2010s and allowed people to buy and sell Bitcoin pretty directly custody by the exchange. That ended up leading to BitFinex and other margin trading lending

platforms. And then the ICO craze in 2018, after that died down, volumes and liquidity dried up. And so there was a mad rush towards derivatives platforms, right, led by perps on venues like Bitmex and FTX, along with other dated futures and options on places like Deribit. So we're now seeing, that similar story starting to pick up steam here in defy. Obviously, we've started off with things like Ether Delta and other spot platforms like these zero X based relayers that then gave way to the margin trading and some naive first past implementations that you were mentioning of perps, options and other derivatives, which are definitely getting more attention now.

And so I think it's just that natural progression is really starting to play out and come to the limelight as well. And I think furthering that something like comp, I think specifically drove a lot of interest here, compound that is. You know, it's been around as a product for quite some time. But certainly the liquidity incentive to borrowers and suppliers definitely ushered in a new wave of interest and demand. And on top of that, it helps that, you know, some of these defy tokens, whether it's compound's native token or balancer or Khyber, they're now being listed on various exchanges, even thrown into perps or. indexes on something like FTX, which of course have large and broad audiences. So yeah, it's really interesting to see these kind of first, first pass or even second.

I think, you know, Uniswap is now coming to their second version and seeing what that looks like over the next couple months and a couple of years. It seems like as far as your protocol goes, a really big piece of it. And this is probably the one where there is the most ambiguity now, because it's just a matter of experimentation would be the Dow and seeing how that works and how you're able to distribute control of some of these more sensitive operations. And I guess a lot of that comes down to the actual distribution method. So what are you planning in terms of distributing these units of control to users?

Yeah, that's a great question. So we have divided our token up in two different camps, if you, will. So there's kind of a pre-mine bucket, which is given to the team to private investors. These are accredited institutions or accredited angel investors. And so we have, and also towards other kind of marketing efforts or affiliate programming, things like that. And then we have what's, I think now you've probably heard in vogue, liquidity mining, farming, et cetera. So the other half of the tokens are used for that and being issued to users that are, that are using the platform.

And so it's a really fascinating topic. In some sense, liquidity mining isn't a totally new concept, right? Like F-coin, for example, I think in the eyes of many, kicked off this process back in 2018, and it totally spiraled out of control. I think near instantly. The exchange had all sorts of dishonest volumes that I think actually exceeded even what would be in Binance combined,

which is wild to think about. But they were issuing, you know, a crazy amount of F-coin. And also, I think, sharing like 80% of exchange profits as a dividend. And so all of that led to the floor just being pulled out. And CoinFlex is another example that also had this two-year token issuance with a fixed daily emission. And so that also, we studied their model quite extensively. And while volumes boomed in the early going, fairly quickly, I think you experienced diminishing returns. So the concept was there and we were excited by it.

And the question is, how do you incentivize participants in a way that organically develops the platform, but yet, you know, brings on especially the early, early users. And yeah, so we can certainly talk more about the model here. Yeah, I guess it's a really interesting optimization problem because ultimately you want the ultimate owners of the kind of unit of control, the token that determines voting in the Dow, you want the owners to have a close mapping with actual users of the protocol, traders and so on. But it's kind of largely permissionless given that you're doing the liquidity mining method. And so it's the slope of distribution which affects the ultimate outcome.

If you make it too aggressive, you're going to get a huge rush of people that just want to do a flip and want to get access to these new tokens, and you end up with potentially an F-coin kind of scenario. Nick, that's 100% right, and that's exactly the way we thought about it also. So for us, you know, this is something that we've spent, maybe the most amount of time really thinking about. We've worked really closely with Olaf, Neeraad,

you know, the rest of the poly chain team in the early going, really discussing this at length. And, you know, the first step, as you rightfully pointed out, is realizing immediately that a liquidity mining program does not solve for an inherently deficient product, right? Like, liquidity mining isn't magically going to turn a terrible product into the best things in slice bread. That's not going to happen.

In fact, I could ruin a product like we've seen with some of the examples that I mentioned earlier. And so this philosophical realization outlined from the get-go is paramount. And any other interpretation or belief short of that, I think results in the same demise as coin flex fcoin and i think to a certain extent what you're seeing in today's and and likely tomorrow's flavor of incentive programs as well so there's a lot of intricacies and interactions that we'll get into in a very elaborate research post in the coming weeks but at a high level you know you're exactly right like the emissions schedule and the slope of that is really important here

so you want it to be dated long enough but also dynamically changing that incentivizes global outperformance of volume. So like at the platform level, incentivizing growth, not necessarily on the micro, individual trader aspect of things. And so that way you don't have this diminishing returns. And then the token also has,

needs to have a lot of utility on the platform. So by way of governance, by way of fee reductions, other staking opportunities. So that fundamental utility of the token on the platform, coupled with a more sane emission curve as we just discussed, that incentivizes participation, but in a manner that economically disincentivizes wash trading, this flipping behavior that you were talking about. And so really what

you're left with is a platform that organically grows and has the longevity to carry the success of the bootstrapping liquidity mining program beyond that. It seems like a lot of the design considerations are informed by this kind of second mover advantage, whereby you've you've watched both in terms of the decks construction and newer things like liquidity mining for token distribution. You've seen some of these experiments play out and taken learnings. And so hopefully you can fit the whole thing together in a coherent hall. That's definitely right.

And this is particularly the case in the crypto markets. Like if there's an exploit, people 100% will exploit it, right? That happens. If something can be game, they will. And so it's really important to study those that have, you know, pioneered certain aspects before us and really see what works, what didn't work and what was in between and make sure that we come up with the most robust solution that, like I said, really prioritizes the end user, the trader, the customer, and make sure that they have a safe, rewarding, but also, you know, performant and usable experience. This is super, super important. Yeah, the other thing to consider also is our philosophical approach to the product that we're trying to deliver here. We've seen a lot of very, very strong team, often coming from academic backgrounds, that had a particular mission.

That mission usually is pure decentralization. It's like a very purest approach towards decentralization, finding new ideas, new concept to test out in the wild, like automated market maker, and having a certain level of success on that. So we respect these teams a lot because they help the entire space. They help the state of research. They help determining what's possible and what's not possible. Our approach, however, is a lot more pragmatic. So our mission is to essentially convert C-Fi users into defy users. And in order to do that, we have to make pragmatic choices to be decentralized enough

to be a legitimate centralized exchange in the sense that we're censorship-resistant, We're not custodial and all of these attributes of decentralization. But at the same time, we need to be able to offer a UX just as good or even better as centralized exchanges to bring those users over. So sometimes it's just that it's not necessarily that other projects do things wrong and we do it right. It's just sometimes that we have other mission. And our mission is very practical. It's not a protocol type of mission, it's an application their mission of building the best possible decks to convert new users over to the space.

What, in terms of the actual products that you expect to trade on the exchange, what are you anticipating? Or is it actually going to be up to the users of the Dow, for instance? Yeah, great, great question. So on day one, we're going to have perpetuals on single name cryptos like BTC and ETH. It's important to note that the default collateral that we are supporting on day and we're excited to, you know, explore ways to support multiple collateral types. But the default is USDT. In the same spirit as what Fred was mentioning earlier, like, if you think about

stable coins that can be used, defy has generally been married to die or USDC, which is not even, I guess, necessarily in the same context as die. But still, there's kind of how people perceive it, but people generally are stuck to die and USDAC, but the reality is that USDT by far is where a lot of the demand is for stablecoins. You can see the vast growth that's had certainly through Q1 and Q2 of this year, and particularly in the Asia region, which is where the majority of volumes come out. So the default collateral type is USDT, and like I said, we'll have perps on these single-name cryptos, and the Dow will ultimately vote on other products beyond that.

But we're excited that like our platform is really conducive towards supporting really innovative products like index products, for example. So we're excited to start dabbling into that space, but this is ultimately voted upon by the Dow. Yeah. And ultimately, just a quick note on that, we're planning to have at least one index product from like available from day one, probably an index of major cryptos. And we're going to expand on that. And we think we have good formula to make index products, especially. around the rebalancing schedule that are very attractive compared to what's currently available in the market. And also, in the early going to the exchange, we're going to start supporting multiple margin deposit currencies.

So we're starting with USDT, but very soon after, you're going to be able to deposit other currencies for your margin deposits. So even though, for example, the contracts, the perpetual contracts might be, I don't know, eat USD, you're going to be able to deposit. other currency as margin deposit and open position with that, with that purposeual contract. So it's not going to be limited to USAT for very long. So you're letting the ideological shackles fall and opting for these more pragmatic collateral types, kind of in keeping with your, your pragmatist approach here. Exactly, right. Yep.

Yep. And same with product offering also, right? Like people want access to these other products, like having access to index products, for example is a really powerful, you know, trading strategy, right? And so yeah, exactly in line with our pragmatic approach. Kind of the inversion of the bitmax model where you can collateralize with Bitcoin and then take a position on ETH, on that perp. In this case, you collateralize initially with USDT and you can express if you on Bitcoin without actually touching the spot asset. Exactly right. Yep. Yep.

Would the protocol support more exotic things like options? Or do you think that the swaps are really where the action is going to be at? Yeah. So it's also, you know, not to be a dead horse here, but in the interest of pragmatic, like go-to-market strategy, the demand in the crypto markets, like you just see how volumes have panned out, certainly in C-Fi, the interest is with these perps. Now, that's not how it necessarily will always be. So we're excited to start with perpetuals and cater to the needs, the immediate needs of where the appetite is.

But pretty soon after, our protocol is very conducive towards supporting all sorts of product types and asset types, including options and other exotic derivatives. I don't know if you guys want to, you know, reveal your plans here yet. But in terms of index construction, are you thinking cap-weighted or some sort of more alternative way to create an index? Yeah, we are certainly going to put out more information as to exactly how we're constructing these indices, but certainly we've studied things like the FTX model, for example, and studied how they've constructed their indices. And so a lot of the novelty here for us is in the rebalancing strategy and the transparency that results from that rebalancing,

as opposed to kind of the opakness or vague terms that places like FTX, for example, kind of put out there. So yeah, that's kind of how we're approaching it. And then in terms of what actually goes into the index also and how it's constructed, a lot of that also, this almost sounds like a cop-out answer, but it's actually the truth. A lot of it also happens to be how the Dow ultimately votes and what it decides should go into something like this. But I think some of the characteristics of our price feed, which is sort of an off-chain price feed, as opposed to a non-chain price feed, that we see in a lot of other Defi products make it much easier for us to have these sorts of index product, among other benefits of having an on-chain price feed in terms of favorable liquidity, favorable liquidations,

growing the insurance fund. Yeah, and there's no spot kind of realization of rebalancing that happens here. So it's pretty synthetic, if you will. So it certainly opens a lot of doors for us and makes things more efficient. So yeah, briefly on the on-chain oracles, talk me through the issues there and the risks that we've seen in the last couple months with those. Yeah, so what we've seen at the high level is that anything that requires a transaction to be posted on-chain can be manipulated by someone who just has enough resources to do so. And that manipulation is not necessarily changing the values that the price feed are submitting, but more preventing the price feed, say, from submitting a transaction into the next block,

creating a delay. And you could see how just creating a delay for posting the price update is an issue because when an account needs to be liquidated, we want the liquidation to happen instantly because otherwise if it takes a while, then the price posted. by the price feed and the actual price of the underlying might have a large spread, which means that the liquidation can be unfavorable to the exchange. So we have this insurance fund that we're growing similar to how BidMex grows their insurance fund, which is basically using a spread between a liquidation and a bankruptcy price.

So by being able to instantly liquidate orders against our own order book, then there is no delay between when the price feed submits data and the price of the underlying at the current moment. It's done instantly. So we think that this is going to save us a huge attack vector, essentially, of people trying to game the on-chain price feed to prevent these liquidations from happening immediately. Yeah, and we saw that on the 12th, and this is a recurring theme. No big surprise there. Yeah, if you look at the maker there, act that happened, I believe in March, it was also beyond the price feed, right? Because the way that they're doing their quote-unquote liquidations, it was to use a non-chain

auction. And if you were able to delay like bids posted to that auction, then you could get, you could buy those CDPs at a much lower price. And that's also with price feed and liquidations are tightly coupled in our exchange. And we can make sure that this doesn't happen. again, just by being able to take orders in real time from our order book and use that as part of the liquidation instead of some external auction process that happens on chain. So on that theme of adversarial thinking, it looks like you're going to have a dependency, at least initially on USDT. What would be the consequences on the exchange if there was another kind of run on the bank with USDT? and it trades down at maybe, you know, let's say 10% discount to the peg. What, if any, are the consequences and how do you kind of remediate a situation like that?

I guess the answer is that USDT is going to be our margin deposit currency for a limited period of time. If USDT totally collapsed during that time, then, you know, I suppose it would be an issue similar to all of the other exchanges that uses that use us d as a collateral and what we we have reasons to believe that yeah so basically what we need to do to mitigate this is just to add other margin deposit currency as soon as we can so that users have a choice in that matter we are confident in us d t is limited but we believe that it is uh to be able to use it in the early going. Yeah, and certainly you can also have things like the Dow, for example,

and certainly this takes a little bit longer of a process, but the Dow could also step in and perhaps both the blacklist, certain collateral types and things like that in a really, in a situation like that. But yeah, I think the multiple collateral types that will soon support create a kind of buffet of choices that people can choose from to pick it that real, I guess, Black Swan event. been. When you think about the potential risks here, obviously you're doing something pretty new here in terms of trying to really decentralize all of the functions of an exchange to the extent it's

possible whilst keeping experience good for traders. What are you most worried about? Like, is it regulatory outcomes or would it be just bugs? What is it that really concerns you about the system? Yeah, that's a great question. I think, um, To some extent, all of the above, and we as outlined, like, our approaches here are really to mitigate the risks of any of those to the best that we can. But all of them are very valid, right? So I think one that's interesting to me in particular is, or to both of those rather, is the idea of the insurance fund and how liquid the exchange is. So insurance funds are very important, right? they underpin kind of the overall solvency of the exchange as a backstop towards adverse liquidation

scenarios like we were talking about. And insofar as these are capitalized with, you know, in really liquid environments, that would prevent tapping into an insurance fund. And Dexas historically have struggled with liquidity, right? And so you can imagine this like chicken or egg where basically you struggle with liquidity. Well, now you're going to be hitting the insurance fund hard and regularly. And so we take the insurance fund component very seriously. And so to that end, you know, in addition to something like traditionally how insurance funds populate with just liquidation spreads, we're also taking all of the fees. So something that traditionally exchanges profit on, we're taking all of the fees and putting that towards the insurance fund as well. And so doing

something like that and coupled with what we're also calling insurance mining, and we're going to put out more research on this soon as well, and the idea here being that users can stake directly into the insurance fund and receive token emission by way of doing that and taking any quote-unquote risk, especially in the early going for depositing and backing the insurance fund. These are all measures that we are taking towards capitalizing this, because the risk of auto-de-leverage scenarios, that's a horrible U.S. for exchanges, whether it's centralized or decentralized, right? Like, you want to mitigate these clawback events. And so long way of saying, I think liquidity, bootstrapping, and specifically around not draining the insurance fund is something

that is concerning. And we're putting a lot of resources into thinking about how to mitigate that likelihood. So with the insurance mining, so to speak, you would allow individuals, or entities to effectively sell insurance and collect a premium potentially in exchange for taking on the risk of there being really significant amount of slippage or issues with settling these contracts? Exactly right. So like, especially in the early going, right? Like on day one, the insurance fund isn't capitalized by way of liquidation spreads or fees,

right? Because the exchange just hasn't started yet. And so the idea is, you know, kicking off a process where people can start doing that and start staking into the insurance fund. In addition to that, even after the exchange launch, people can deposit into it. And as you said, receive a premium for that almost loan into the fund is something that is a really strong proposition for traders. And presumably as the exchange performs better and better, the likelihood of their deposit being drained or losing that decreases. Well, you guys are certainly introducing some really original concepts into the world of Dex's.

I will be following you progress eagerly, and I'm excited to see where this goes. In terms of having our listeners follow your work, what would you recommend? Where are the best places to follow you? Yeah, that's great. So we have a Twitter account, DDX underscore official. Our website is DervidX.com. You can sign up via there. And then we have a Discord community that's out on the website and on the Twitter channel as well.

And a lot of the discussion is happening there, as well as the Telegram channel, which is where a lot of the trading community ends up participating in. So a bunch of different channels would recommend users participate there. And operating as a Dow, this is very community driven. And so we really look forward to engaging with you, Nick, and the rest of the community from here on out. Well, Fred and Audi, this has been really fast. Thank you so much for coming on. Thank you, sir.

Thanks for my. Thank you, sir.