"V3 is Winding Back a Bit of the Uniswap Revolution; It's Going to be Very Influential:" Dan Robinson

In this week’s episode we speak with Dan Robinson, a research partner at Paradigm, one of the most active VC funds in DeFi and crypto. One of the fund’s investments is Uniswap, where Dan also had a central role in the design of v3, the latest version of DeFi’s most popular decentralized exchange. Dan talks about why Uniswap was so exciting as an investment. He said at the time, many blockchain applications still involved some centralized parties and custodied funds. But Uniswap was really hands-off; A self-contained system with no admin keys.

But the issue was that it wasn’t very capital efficient, meaning funds in liquidity pools for any given trading pair would go mostly unused. And that’s what they set out to solve with V3. The solution Dan helped find was concentrated liquidity, or the idea of concentrating funds at the price range where an asset is traded the most.

Dan talks about his big ambitions for Uniswap -- he believes that most liquidity will move to Uniswap v3 and that the new concentrated liquidity design will have an influence on the whole DEX space as more projects start to optimize for capital efficiency. It will also roll back the revolution Uniswap itself prompted, and move AMM (automated market maker) design into an order-book hybrid.

We also talk about miner extractable value and how he sees projects are evolving from those trying to democratize MEV, to those trying to stop it from ever happening. We talk about his views on governance and how he believes protocols should try their best to minimize decision-making. His main area of focus continues to be on AMMs. In the long term, he believes they will have an impact on investment similar to what exchange-traded funds did to stocks, allowing anyone, from big institutions to small individual investors, to earn a passive income by providing liquidity, something that only large market makers could do before.


🎙Listen to the interview in this week’s podcast episode here:


This week we’re sending the full podcast transcript to all newsletter subscribers (free and paid) to make up for the fact that the sound cuts off at times during the interview. Paid subscribers get the full transcript every week.

Share


🙌 Together with: 

  • Balancer, one of the leading DeFi automated market makers (AMM) for multiple tokens. Dive into their pools at https://balancer.finance/!

  • Kraken, consistently rated the best and most secure cryptocurrency exchange, which can get you from fiat to DeFi

  • Aave, an open-source and non-custodial liquidity protocol where users can earn interest on deposits and borrow assets. 

  • Kyber DMM (Dynamic Market Maker)a next-gen AMM designed to maximize the use of capital by enabling high capital efficiency and dynamic fees to optimize returns for liquidity providers.


Dan Robinson: So when I first got into crypto, I was in law school, and then an unhappy lawyer for a couple of years, so I took the long way around to enter the industry. Once I decided I didn't want to practice law anymore, I left to become a programmer and joined a crypto startup called Chain, which was acquired by Stellar a few years later. And then I joined Paradigm two and a half years ago. I'd been friends with one of the cofounders, Matt Huang, for a very long time. So when he and Fred Ehrsam from Coinbase started Paradigm, I was one of the early hires and helped build out this strange thing we've got at our research team. I should say, I think Paradigm is a crypto asset investment firm. But anything I say on this podcast is my view and not those of the firm. And of course, nothing that I say is investing advice. 

Camila Russo: Of course. Okay, so interesting background there. Lawyer, developer, hacker and researcher. What specifically do you think drew you into crypto when you were studying law?

DR: I was kind of fascinated by Bitcoin when I first saw it, and got into it a little bit. But really, it was when the Ethereum White Paper came out that I got really excited about crypto. Because Bitcoin created this concept of programmable money, sort of programmable at the protocol layer, but you couldn't do all that much with Bitcoin. Still can't really do all that much in terms of programming with Bitcoin. Or rather, there's a lot you can do, but it mostly involves clever workarounds to do really simple stuff.

Back then, after Ethereum launched, there still wasn't very much to do on it, but you could just talk about it and see that everyone's talking about what it would become. I thought that was really fascinating, so I got deep into that kind of Ethereum online research community.”

And so, when Ethereum came out, I think I got very excited because it’s just fun to play around with. I was between this law firm job that I hated, just sort of exploring and trying it out. Back then, after Ethereum launched, there still wasn't very much to do on it, but you could just talk about it and see that everyone's talking about what it would become. I thought that was really fascinating, so I got deep into that kind of Ethereum online research community. You cover a lot of this in your book, of course. 

But back then, yeah, mostly just hanging out on Reddit, and then eventually, ETH research when that forum was created. I ended up steeped in this lore and also connected to a very interesting and fascinating Ethereum research community. So that's how I got to meet Karl Floersch, who was at the Ethereum Foundation, now at Optimism. Through him, I met Hayden Adams, founder of Uniswap, Vitalik, and the rest of this great researcher-identifying community within Ethereum. So I really enjoyed that. 

And then generally, within crypto, there's just fascinating stuff going on and fun people who spend a lot of time thinking about it. 

CR: Very cool. Okay, so you were part of this Ethereum research community before starting at Paradigm and you were able to meet Hayden, founder of Uniswap there. And then when you joined Paradigm, had the funds already invested in Uniswap or did that come after you joined?

DR: We invested in Uniswap sometime soon after. It was the first investment that I worked on right after we joined. And yeah, back then Uniswap had already been launched, and it was growing in popularity among the Ethereum community. And the growth since then, of course, has been extraordinary. 

Uniswap Investment Thesis

CR: Yeah. So I'm not sure if you can talk about this, but what was your thesis behind the Uniswap investment? 

DR: Yeah. So I think what I found very appealing about Uniswap was that it was an application that seemed not to really make any compromises in terms of decentralization or security, relative to the base layer. At the time, even the base layer was not necessarily that hardened or secure. But it was this platform where its code was supposed to be sort of unstoppable, and noncustodial. 

But a lot of the applications that we were seeing on it still involved some centralized parties, still involved decisions being made by small groups. And in many cases, I think, funds were still being custodied in one way or another. I think what's really interesting about Uniswap, and this is unique to that particular algorithm, is that it really is especially hands off. Like, it's a perpetual motion machine where once it's there, it doesn't need to be maintained other than the base Ethereum protocol still running. 

But it's just this self-contained system and there's no admin keys to it. I found that extremely fascinating. Because it's one of the few systems in crypto, especially at the time, that really struck me as being that decentralized, really immutable or invincible. That was very appealing. 

I think what's really interesting about Uniswap, and this is unique to that particular algorithm, is that it really is especially hands off. Like, it's a perpetual motion machine where once it's there, it doesn't need to be maintained other than the base Ethereum protocol still running.”

Another was the permissionless of it, not just in the sense of decentralization, but in the sense that it allowed any assets to be traded rather than some particular set. It was the first DEX at all to support arbitrary tokens, ERC-20 tokens. That was just a huge leap forward in that there's been tens of thousands of tokens listed on Uniswap over the years, and a lot of weird experiments using it. You don't have to go get a token added to Uniswap. You just do it, and now, suddenly, you have the ability to do market making with these tokens. 

“So I think these ideals, this immutability, this permissionlessness were really at the heart of Uniswap. What excited me about Ethereum also excited me about Uniswap, and still really is core to Uniswap’s values today.”

CR: Yeah, totally agree. How instrumental do you think Uniswap has been to DeFi as a whole?

DR: I think there's some very clear, direct influences. Most notoriously, direct forks like Sushiswap. I think it's had some effect on the general philosophy of DeFi. And people actually bend over backwards a little more now to try to support permissionless, and try to have not just decentralized governance, but ideally governance minimization, sort of no over-governance that isn’t needed. 

It's in this particular niche where there happens to be a very elegant algorithm that works across all these different assets. So we're still looking for things like that, that can be applied to really transport that philosophy elsewhere.”

It hasn't been as influential in some of those areas as I might like. But it's unique that Uniswap is able to do that. It's hard to have a lending protocol that could support arbitrary assets. Like, it'd be hard to have an insurance protocol that could support something like that. It's in this particular niche where there happens to be a very elegant algorithm that works across all these different assets. So we're still looking for things like that, that can be applied to really transport that philosophy elsewhere.

CR: Yeah, that would be so interesting to see, like, a Uniswap version of a lending protocol. I don't think we've seen that yet, something as hands-off.

DR: Yeah. Sushiswap and a protocol called Kashi. I haven't looked closely into it, but it's a pairwise learning protocol. And what's interesting is, Sushiswap, they haven't really done... The only product that’s actually really successful has been the DEX that they've cloned wholesale from Uniswap. But they have been doing these other interesting things. And I would love for them to really focus and lean into some of these other areas where they're trying to be innovative. I think one area is in lending. 

We've seen a few other protocols out there that are trying to bring this pairwise lending approach to market. I think it's very challenging. It's a lot harder than it even was for Uniswap, at least Uniswap v1, in what it did for DEXs. 

Uniswap v3

CR: Sure. I want to talk about Uniswap v3, which is the new version of Uniswap that launched recently in March. It was so long awaited, and it really changed the way the whole protocol was structured. I understand you had a key role there in designing the liquidity provision side of v3?

DR: Yes. So v3, I was talking about how the Uniswap algorithm works for all assets. But of course, it doesn't work ideally for all assets. And there are some asset pairs, most famously stablecoin pairs, where Uniswap is not a very good fit, like most of the assets in it end up wasted because the price used for these assets is almost always supposed to be around $1.

So that was one of the core problems we were trying to solve within. And this was one that, after we started working on Uniswap v3, Curve came out and solved this problem specifically. But with v2, we were trying to solve a more general problem, which is that even in pairs of normal assets, volatile assets with USDC, most of the liquidity in that pool is never touched. Most of the assets in any of the USDC pools are never touched. 

And so long after I started, I started looking at Uniswap, and I thought I understood it pretty well. I had a conversation with Martin Köppelmann from Gnosis, and he pointed out to me something that interest was very un-intuitive to me about Uniswap’s capital efficiency, which is that 25% of the liquidity in any Uniswap pool is reserved and not touched unless the relative price of the two assets moves by a factor of 16 in either direction. And that doesn't happen very often, for most assets. Even with the USDC, that's happened a couple times in the past few years.

“The big thing was, we didn't want to give up that permissionlessness or that governance minimization. We didn't want to be choosing ranges for people, or saying when you add an asset that governance would have to decide where liquidity was catered along the curve.”

So for most pools, you can almost double the capital efficiency of a pool just by dropping the tokens that would be used only at really extreme prices. That observation led us to develop this feature of concentrated liquidity, which is the feature at the heart of Uniswap v3, which allows liquidity providers to provide in any concentrated range. 

Now, the big thing was, we didn't want to give up that permissionlessness or that governance minimization. We didn't want to be choosing ranges for people, or saying when you add an asset that governance would have to decide where liquidity was catered along the curve. And so it required a much bigger engineering lift to design this protocol so that any liquidity provider could provide it any range, but do it in a way that was still capital efficient. So that's what we wanted to achieve building stuff for Uniswap v3.

CR: Okay, so to catch people up: The main problem was that when you talk about capital efficiency, it's this idea of in Uniswap v1, v2 liquidity providers, they had to provide ETH on whatever token pair they weren't providing liquidity for, and that liquidity was spread across a price curve, which didn't account for where the most commonly traded price was. So you'd end up with these far ends of the curve where the liquidity provided was never touched. So now in Uniswap v3, the idea is that you have this concentrated liquidity that's concentrated on the price pairs that are the most traded on, so you don't have unused capital at both ends of the curve. 

And it's interesting, because I was looking at stats right now for volume and total value locked on Uniswap v2, which is still running, and Uniswap v3. And there’s still more value locked on Uniswap v2. But there's more volume traded on Uniswap v3. So that speaks to that capital efficiency. You're able to have more volume on a DEX, even with less capital inside of the protocol. That already shows you it's achieving its goal, at least in part.

DR: That's absolutely right. And that was sort of the key goal with v3, to not need $10 billion in TVL in order to provide liquidity. I think TVL has been something of a vanity metric. Honestly, for Uniswap, for a while we sort of equated higher TVL with higher liquidity, because that's how Uniswap v1 and v2 work. You need more TVL liquidity proportional TVL. In v3, it isn't. And I think the uncoupling really makes it a lot more capital efficient. So, yeah, that was a big goal.

TVL has been something of a vanity metric. Honestly, for Uniswap, for a while we sort of equated higher TVL with higher liquidity, because that's how Uniswap v1 and v2 work. You need more TVL liquidity proportional TVL. In v3, it isn't. And I think the uncoupling really makes it a lot more capital efficient.”

The other thing is that I think it opens up this big design space for people designing custom curves. By combining a bunch of different v3 positions at different prices, you can kind of simulate different curves. So that's what a lot of my research recently has been. If you have a design for a custom market maker, because I like to design AMMs, how can you implement this on top of Uniswap v3 so that it gets aggregated with everybody else's liquidity? That's been one of my recent research projects. 

Designing AMMs

CR: Interesting. Okay, so designing for different curves would mean designing AMMs for just specific types of tokens or trades in different ways? So an example would be, if Curve didn't exist, maybe you would create an AMM that was specifically designed for stablecoins, right, and maybe there are other types of tokens that trade in very specific ways that require different types of curves, is that the idea? 

DR: That's right. So Curve is one, and that's a relatively simple one. You can have a strategy that just provides very concentrated liquidity around the price of one, and then maybe a little liquidity at other prices like Curve does. And then that AMM would simulate Curve. But for the user, they would just see your AMM, but it would all be added to v3. Then you can do more complicated things.

So Balancer, for example, allows these different weights of pools. And in fact, you can simulate Balancer pools shown  in a recent blog post of mine, called Uniswap v3 of the Universal AMM. You can simulate a Balancer pool by adding liquidity to different ticks on Uniswap, which actually ends up looking like an exponential curve in the liquidity space. So it's a fun derivation there. But ultimately, I'm very curious to see what other people build with it. And so I'm working on infrastructure to help be able to do that. 

“I'm very curious to see what other people build with it. And so I'm working on infrastructure to help be able to do that.”

CR: So, to replicate a Balancer AMM, could you also make it so that people can provide different types of tokens in a pool?

DR: What do you mean? 

CR: Like, because on Uniswap you need ETH as you have two pairs, but Balancer has pools of many different tokens, so can you still do that on top of Uniswap?

DR: You know, that's a great question. And when I'm talking about simulating Balancer on Uniswap, there are two asset balancer pool rates, like 80-20 pool. That's a great question. I'm not sure how to simulate a three asset pool on top of v3, you may not be able to. Personally, I think two is a pretty magic number in terms of trading pairs. There's a reason that most exchanges quote pairs, and you have trading pairs rather than trading triplets. But I do think there are valid reasons to have pools with more than two assets, and it just hasn't been a focus of Uniswap.

Personally, I think two is a pretty magic number in terms of trading pairs. There's a reason that most exchanges quote pairs, and you have trading pairs rather than trading triplets.”

Influencing the Space

CR: Okay. So Uniswap really impacted the DEX space since its launch in 2018. Since Uniswap launched, it really has been like AMM dominating the DeFi space. Now with this new type of liquidity provision, it's like concentrated liquidity. Do you think the same thing will happen to the DEX space, and that everyone will move to this new model?

DR: I think everyone will move to Uniswap v3. That's my hope. And I think it's a superset of Uniswap v2. You can provide liquidity across the entire price range if you want on Uniswap v3. So it's strictly better, apart from some gas tradeoffs. I do think, ultimately, liquidity should move off Uniswap v2, and the other Uniswap v2 is out there, like Sushiswap, and I expect that to go to v3. 

I think everyone will move to Uniswap v3. That's my hope. And I think it's a superset of Uniswap v2. You can provide liquidity across the entire price range if you want on Uniswap v3. So it's strictly better, apart from some gas tradeoffs.”

I do think it'll also have more of an influence on the rest of the DEX space. And I do think we'll see people trying to optimize for capital efficiency in different ways. We've seen with Curve’s recently released v2 white paper where they're trying to target this issue of efficiency for arbitrary pairs, doing it in a pretty different way from Uniswap v3. But I'm curious to see how that experiment works out. I do think we're going to start seeing more of that. 

It's also worth noting that in some ways, v3 is winding back a bit of the Uniswap revolution. Where before Uniswap, it was commonly believed that the best way to do training was on these order books, because we have all this configurability. And then Uniswap was like, well, what if there was just one strategy, and you just pulled all your money in it, and there's no order book so you have to provide orders across the entire range? 

V3 looks a lot more like a hybrid between an order book and that strategy, where you can efficiently provide the full range using that strategy. But you could also provide it in a very narrow range, or by combining different positions, make it look more like your arbitrary trading strategy. So ultimately, I think Uniswap v3 is going to be very influential on how crypto and DeFi works. But I think it is the next evolution where AMMs are sort of like cost and product AMMs, sort of taking us as far as they could go. And we needed to make some concessions too, so it looks a little more like the flexibility of an order book.

CR: That's so interesting. And I wanted to ask you about this. Because some of the  initial feedback that I've seen from users of Uniswap is that it isn't as user-friendly as v2. It's just more complicated. It requires more upkeep, like, you have to manually change your ticks. And so, do you think is it possible that this new version drives away more novice  users and makes liquidity provision for more sophisticated traders?

DR: First, I think like I said, you can still provide liquidity to v3 in exactly the same way as v2 if you're a passive liquid provider. You can just provide it across the entire range. Ultimately, I think v2 doesn't stand a chance against v3, because passive liquid providers on v3 can't compete with the active ones, and people on v2 forget about it because they're not even getting the volume. So they're just going to get iced out. 

I think it's an inevitability in that sense. Right now, Uniswap v3 has really fabulous liquidity. My understanding is a lot of it comes from more professional, active market makers. Ultimately, the AMM exists for the benefit of traders and swappers. Ultimately, what matters is really having deep liquidity and to be the best place to swap these assets. 

“Ultimately, the AMM exists for the benefit of traders and swappers. Ultimately, what matters is really having deep liquidity and to be the best place to swap these assets.”

The reason that Uniswap was great was that it made it easy for liquidity providers, but I think this compromise here makes it so much easier for advanced equity advisors who are willing to maintain this actively, that just has this huge, up to 1,000x or more capital efficiency benefit.

We have started to see a lot of innovation on top of Uniswap v3 to make it easier for passive liquidity providers to not have to do their own rebalancing, but to still provide concentrated liquidity. And I think we're seeing a lot of innovation in that space. It's a hard problem. I'm not sure. I think we're about to see what people develop experimentally to do this, but I'm very curious to see that.

CR: Yeah, it'll be interesting. So you're saying that ideally, there would be some innovation on top of Uniswap v3 that would allow passive liquidity providers, and that maybe they'd get less fees than active liquidity providers. But at least now, you'll have the choice of being a passive liquidity provider, and okay, you'll get less fees. But then now, more sophisticated traders and market makers will find DeFi more appealing, and maybe the whole ecosystem will benefit from having more liquidity from these more professional traders?

DR: Right. And ultimately, it is aggregated with the pros. You're getting this pro rata share of the total volume. And I think that's very powerful. It is very different from how traditional markets work, where, as a passive market maker is not a professional, you can't really get any order flow, or make any profits from it because it's all very professionalized. And so you still get this pro rata share of all the volume. That's some caveats. But yeah, I think in general, Uniswap v3 makes it a lot fairer than a plain order book would.

LP Positions as NFTs

CR: Super interesting. And then one aspect that we haven't talked about is the NFT aspect. So now because liquidity providers are providing liquidity to a specific price, or a place in the price curve, you get NFTs in exchange for your liquidity provision instead of ERC-20 tokens. So what are some of the implications of this, of getting NFTs back? And what are some interesting things that you're already seeing, maybe in secondary markets trading of these LP NFTs?

DR: Yeah, so I think this was a necessary compromise that we had to make, despite the concentrated liquidity feature, because now not all liquidity is created equal. I think it was worth it. And in fact, I don't think, relative to what v2 was when liquidity was a token, I'm not sure it changed all that much. And you do have this nice benefit of having these pretty NFTs with autogenerated art for your liquidity position. 

I think it requires some more activity to build stuff on top of it. And so some examples of things people have done with Uniswap liquidity, v2 liquidity tokens, is create lending markets where you can borrow against liquidity, or have liquidity mining where you can incentivize people to provide this by having them stake their liquidity tokens in order to get some other token. You can still do those things, I believe, with Uniswap v3 liquidity positions. It just takes some more work,some more effort to go into the protocol built on top of that. 

“...so some examples of things people have done with Uniswap liquidity, v2 liquidity tokens, is create lending markets where you can borrow against liquidity, or have liquidity mining where you can incentivize people to provide this by having them stake their liquidity tokens in order to get some other token. You can still do those things, I believe, with Uniswap v3 liquidity positions. It just takes some more work...”

So I think it's a lot more flexibility with a little more effort, and I do think it took a while before we started to see liquidity tokens get used in creative ways. I think it'll take a while before we start to see these NFTs used, but I'm confident that people are going to find really interesting applications for them.

CR: Right, we'd have to see a more developed intersection between NFTs and DeFi to begin with for people to be able to use NFT LP tokens.

DR: Oh, yeah. And generally, I don't think there's any way right now to borrow against NFTs as collateral, like digital art NFTs, and that would certainly also be something that you could do if you've built that infrastructure. So yes.

Dark Forest

CR: Yeah. No, it's mind-blowing, the turns that finance is taking with all this stuff. Okay. And then going to another topic, I wanted to ask you about your post, the Dark Forest, which I think is kind of becoming an Ethereum classic at this point. But if you could explain about this concept of The Dark Forest and miner extractable value, and why it's important...

DR: Sure. So this post is about a story that happened to me last August, to me and my colleague, Georgios who co-authored the post. I was in this position where I could effectively white hat hack a contract. I wouldn't even call it a hack. It really was just like a call of a function on a contract, and just picking up some money that somebody had accidentally put there, then I could give it back to them. 

And I had this voice in the back of my mind telling me actually, wait a minute, don't call that function yet, because I'd heard a story from Phil Daian about this problem with that. So Phil is a cofounder of Flashbots which is a Paradigm portfolio company with a focus on MEV and I'll talk about them in a second. But Phil had been researching this thing, miner extractable value for a while and miner extractable value is something of a misnomer because it isn't always extracted by miners, although in the long run it ultimately we think is.

Miner extractable value is any opportunity that is created by the Ethereum transaction state or by transactions that can then be exploited and extracted by front running bots or miners on Ethereum. And Flashbots, a Paradigm portfolio company, is designed to basically help make it easier for miners to extract this extractable value in a non-wasteful and sustainable way. So what Phil had told me about was a particular kind of front running bot that he called a generalized front running bot. 

And what this bot would do is we just look at all transactions, and see, can I do a copycat of that? Can I just copy and make money from it, like replacing the address with my own? And if they can, then they front run you. And so I'd heard about this, I'd never seen one in the wild. And it wasn't really all that common, only a few people had seen these or heard about them, and I was just lucky to have heard about it. So I tried to construct this elaborate way to get around it, ultimately failed, and did get front run by one of these bots in a matter of seconds. 

So I put out this post about it, which I thought was this huge failure of mine. I was a bit sheepish about putting out this post. But the response we got was great. I mean, it traveled further than the Ethereum community, because I think it was sort of a fun technothriller…

CR: It was so well written, it was a fun read.

DR: Right. Thank you. And then I think within Ethereum, this specific concept of the generalized front running bots, and the Dark Forest metaphor which was from my favorite science fiction book, The Dark Forest, further elevated this topic of MEV, which has been a very big topic especially in the past year in Ethereum.

Solving MEV

CR: Yeah, I think it was very helpful, because I think most people using Ethereum are affected by these front running bots, but I don't think many people are aware of it. And it's this really helpful metaphor of thinking about the Ethereum mempool or unconfirmed transactions would be the right way to describe it, right? So all these transactions in Ethereum are waiting to get confirmed by miners, but while they're sitting there, they are basically vulnerable to these monsters or silent killers of the night who’re coming to steal any value that's left, or that's there that can be stolen. 

So, there's been a lot of innovation in this space, from projects who are trying to illuminate the Forest. I think there's this acceptance that there's not much you can do with the value that's being left there by users. But at least I think there's these projects that are trying to make it more fair so that anyone can have the opportunity to get that value. Is that right? What's the innovation that you’re seeing there? 

DR: That's right. I think we were in a period for a bit where this was dominated, like the miner extractable value extraction, by these front running bots racing each other with these gas auctions. And it's a very inefficient way to do this and it was driving up gas prices. I'm not sure if Flashbotscan take all the credit for gas prices falling. But I do think it's made the extraction of MEV somewhat less of an arms race between these front running bots. 

Ultimately, I do think we're going to need a solution to it where the miner extractable value doesn't get extracted at all. And whether that involves changes to application architecture, whether it involves some kind of blinding of transactions, we’re also on the forefront of that research to really figure out how to make this not just efficient to extract, but really MEV-proof the entire system.

“I do think we're going to need a solution to it where the miner extractable value doesn't get extracted at all.”

CR: Oh, interesting. I didn't know that was possible, or there was research being done for that. That's cool.

DR: Yeah, there's some ways we thought about it. Can we use cryptography for this? Can we use trusted execution environments like secure hardware for it? I think it's still relatively early on. There is another project actually, I think on Cosmos, that uses verifiable delay functions, this fancy cryptography, in order to make transactions encrypted until they're included and then they get decrypted by the protocol. So a very neat potential way to try to make MEV harder to extract and may give users a little more privacy until their transactions are included.

“There is another project actually, I think on Cosmos, that uses verifiable delay functions, this fancy cryptography, in order to make transactions encrypted until they're included and then they get decrypted by the protocol. So a very neat potential way to try to make MEV harder to extract…”

Governance Minimization

CR: Okay, super interesting. And then the other piece of research or writing that I wanted to ask you about is something that you've already mentioned, minimized governance. And so you've been a big supporter of minimized governance. It's what first drew you to or one of the reasons you wanted to invest in Uniswap, right? How has that thinking evolved over time? 

Because governance has become such a big part of DeFi. Every project is forming a DAO and is trying to decentralize its management by giving ownership to its users and having them make all the decisions via token votes. So it seems like the space is turning towards more governance, rather than less as everyone is trying to decentralize, and so that makes them hand over governance to users. So how do you think DeFi can continue to decentralize, but still be governance minimized? It seems like two opposite things.

DR: Right. Well, I think what we're seeing in DeFi isn’t an increase or decrease in governance, we're seeing something more of basically a move from one kind of governance, which is governance by an admin key, which a lot of projects have early on. And in some cases, that's the appropriate way for early stuff to be governed, to governance by a decentralized organization. 

And so, this is a change in how the decisions are made. But ultimately you've still got some kind of process that is making a decision that affects all the users of this application. And I think when you're trying to design a protocol, if there's a way to do that where they're just isn’t a decision at all, I think that is sort of the most hardened way to try to design a protocol. It's maybe the most likely to be resistant to attacks. 

“I think when you're trying to design a protocol, if there's a way to do that where they're just isn’t a decision at all, I think that is sort of the most hardened way to try to design a protocol. It's maybe the most likely to be resistant to attacks.”

And I think we've only begun to see what a potential governance attack could look like. I think as we start to see DAOs, you know, there are many DAOs out there with billion dollar treasuries, I think we're going to start to see sneaky attacks on governance that really could sort of put some of these systems in jeopardy. 

“...I think we're going to start to see sneaky attacks on governance that really could sort of put some of these systems in jeopardy.”

And you know, so we're going to start seeing an era of maybe governance extractable value. Early on, miners weren’t extracting any of the miner extractable value, and it was a niche thing to do. And now they're all extracting it. I think ultimately, as the market matures, I worry that we're going to see something similar with governance. And so, in my view, I'm searching really hard for system designs that don't involve these kinds of decisions, that don't depend on either a single party, or on some kind of like a majority token voting type process, but can still produce a usable product. 

And so again, I think Uniswap, the constant product market maker and AMMs in general, just happen to be a really nice, clean instance of that, but I think we're starting to see, and it's harder to maybe do this for other applications, but we're starting to see more people doing it. And ultimately, I think those protocols could just outcompete the others in the same way that I think any decentralized protocol will outcompete centralized ones for a particular category of things.

CR: Super interesting. Okay. So the point there is that what you'd like to see and what becomes more unstoppable are protocols that just don't require that much decision-making. So you don't need some complex governance system that's token-based, and requires voting or anything like that. It's just the system works how it's supposed to. And that's it.

DR: Yeah. And I think, look, as systems get more complex, I do think there's some areas that just necessarily require governance. And so Uniswap v2 and v3 have this small amount of governance that the UNI token controls. But the surface area of it is as limited as possible. It can't, for example, steal all the assets in those pools. 

And I think that's very valuable. Because if you put too much pressure on these mechanisms, if you have a $10 billion bounty to subvert this governance system, with the prize being like you get to steal all the assets in all the Uniswap pools, I think that makes governance less good at doing the few things that maybe it has to do. And so it doesn't necessarily mean, like don't have a governance system. It means don't put too much pressure on it by trusting it with too much stuff. Ultimately, treat it as a pretty scarce resource and something to be used very carefully. 

“Because if you put too much pressure on these mechanisms, if you have a $10 billion bounty to subvert this governance system, with the prize being like you get to steal all the assets in all the Uniswap pools, I think that makes governance less good at doing the few things that maybe it has to do.”

CR: So have a few key decisions that a governance has to make and that's it? Like don't rely on it for every action on the protocol?

DR: Yeah. 

CR: Because otherwise, what would protocol tokens be used for? To be clear, governance tokens have a use, it's just it should be limited, right?

DR: Yeah, I think it's yeah, for anything where a market mechanism doesn't work, or where a hard coded parameter doesn't work. Again I think with a lot of these protocols, you need more parameters. I think one example where I think it shouldn't be used unless absolutely necessary is limiting what tokens, for example, are supported. I think if a Uniswap governance vote was required every time someone wanted to list a pair of tokens on Uniswap, it would never have gotten off the ground, and today it would be it'd be really hobbled by that because just the permissionless ability put these things on top of it is part of what makes Uniswap great as we were talking about before.

“I think if a Uniswap governance vote was required every time someone wanted to list a pair of tokens on Uniswap, it would never have gotten off the ground”

Algorithmic Stablecoins

CR: Okay, so another topic. Alright. Okay, so a hot topic for this week has been algo stablecoins, algorithmic stablecoins. So stablecoins that aren't backed by US dollar based assets or by over-collateralized crypto or anything like that, but they have their own internal mechanism that makes the token maintain its peg.

Yeah, I don't know if I explained it very well. But basically, they're not backed by collateral. They're backed by their own internal mechanism. And this one project Iron Finance had this spectacular implosion. We covered it very well, I think at The Defiant. But it's been an ongoing thing, because it does seem like a holy grail of crypto to have this stablecoin that also doesn't rely on external, fiat currency to maintain its peg to the dollar or to other stable assets. So it’s been this big goal to achieve, but very hard to do. And we've seen one project after another crash and burn as they’re trying to do this. So your thoughts on this, do you think this can actually be done? Some people call them Ponzis, I don't know, what do you think?

DR: Yeah. So, I've been following it a little. There's this original paper on seigniorage shares that came out a long time ago, before Ethereum, I think. And then Basis was a protocol designed to work like this. 

I think so far I'm skeptical, but curious of the general category. I think we do have decentralized stablecoins right now. And the design that really has that been, I think, dominant today for those is crypto collateralized stablecoins, so like Maker, Rai these are both projects that Paradigm has invested in. Effectively they’re stablecoins, they're not backed by fiat. There's no money in a bank account somewhere. You know, there's a real crypto asset backing it. And there's a mechanism for liquidation where if the value of that asset goes down, then some particular person gets liquidated and the system remains solvent.

And so these systems, I think they still have sometimes in the mechanisms something that resembles actually the algorithmic stablecoin ones, like Maker uses its MKR token as kind of a lender of last resort. But I think it doesn't depend too much on that mechanism. And I think there's just been a lot of success with that mechanism, that is probably where I see the future being and maybe finding more efficient versions of those. I think the jury's out on whether we're really going below that level really could ever be stable in the long run, because yeah, we have seen a lot of failures there.

That said, if one of these, because with a lot of them, there's a short life cycle. If I see one of them that lasts a few years, again, you start to actually get some Lindy effect there, maybe believe that it could keep going. But so far, it's still very early on that. No, I'm very wary of that category working.

“If I see one of them that lasts a few years, again, you start to actually get some Lindy effect there, maybe believe that it could keep going. But so far, it's still very early on that. No, I'm very wary of that category working.”

Latest Research

CR: Okay. Yeah, fair enough. Okay, and then to start wrapping up, interested in kind of your latest research pieces focus, like what are you working on right now? 

DR: So a lot of my recent research has been Uniswap v3 and Uniswap v3 adjacent. And in the general, meta research category of thinking, like, how do we think about AMM design? I've been working with portfolio company Opyn on an automated market maker for options, like what, can we find one that works? And also what are we trying to maximize here? What's the approach? What's the system for trying to design a new AMM? And right now, we've kind of just floating around in the dark. 

I worked on an AMM that was designed for bonds with a portfolio company called Yield, that we've been incubating and we came up with a design there. But even for that, it's not that clear really how to make that as an asset, like what's the best AMM for it, and how can these things be designed and explained. And so figuring out, really developing AMM science as a science, is one sort of meta area I’ve been looking at.

“Really developing AMM science as a science, is one sort of meta area I’ve been looking at.”

I think, MEV certainly. I think in general, I've been working closely with our portfolio projects just on the things that are interesting to them. Some of them are confidential for now, but I’m very excited about some of the stuff that's coming out.

The Future of AMMs

CR: Okay. So on AMM, what's your long term take? Or where do you think this space is heading? Is it going to replace order book based, centralized exchanges? Will they always live alongside each other? What's your view there?

DR: I think there's elements of AMMs that should eventually filter themselves into the way that all exchanges work. And the pro rata execution thing, where if you're just sort of passively providing liquidity on Uniswap, your trades are getting executed, it doesn't depend on when you actually are right. It's just proportional to give it to all the orders that are in the liquidity that's in the pool. 

Most exchanges don't work that way. And I think there's this first in first out thing, and then exchanges have more central limit order book centralized exchanges. There's kind of a lot more gamesmanship around market making for that. I do think there might be just some, and this is an area where I'm somewhat out of depth, but I sort of have this almost religious faith that there's something there to the idea that someone should be able to just go in, define a strategy that is like supported in the base layer of the order book system, and just benefit from basically providing passive liquidity. 

“I do think there might be just some, and this is an area where I'm somewhat out of depth, but I sort of have this almost religious faith that there's something there to the idea that someone should be able to just go in, define a strategy that is like supported in the base layer of the order book system, and just benefit from basically providing passive liquidity.” 

And the analogy I use too, is in the 70s, when Jack Bogle started Vanguard, it looks like a crazy thing to do just buy all the stocks, just market cap weighted, just all the big stocks, and just not really try to even evaluate them individually, just buy everything. And it turned out that the efficiency of that, and the massive amount of capital that could actually be deployed in that way, just completely transformed how investment works. Today, I think there's some possibility that adding liquidity has the same effect on what today is a very competitive and high margin market making world.

CR: That's so interesting. Okay, so maybe what happened to active versus passive fund management, which I think recently passive fund management and index based investing surpassed active fund management for the first time, you think that might happen to the market making and liquidity provision, and so in the same way that ETFs allowed for retail investors to easily access stocks, maybe these kind of AMMs, or DEXs will allow regular investors to also become market makers.

DR: Yeah. That's the basic idea. And I think these market makers, it's a very profitable business, but they don't deploy that much capital to it. They're relatively capital efficient. I think it's possible that if you have this massive amount of retail liquidity there, that could basically upend that business model. This is very speculative. You know, I'm not an expert in traditional market microstructure by any means. 

“And I think these market makers, it's a very profitable business, but they don't deploy that much capital to it. They're relatively capital efficient. I think it's possible that if you have this massive amount of retail liquidity there, that could basically upend that business model.”

I do also think that Uniswap v3 in some ways, maybe it seems in the short run to undermine this thesis, because right now, you do have to be more active than a Uniswap v2 provider. But again, you still have this pro rata execution. You still have these automated strategies that are built into the core layer, you don't have to be going there and placing orders and running co-locating next to the exchange servers. It’s quite democratized relative to that: ignoring MEV for now, because MEV is not necessarily as democratic yet as it should be.

CR: Right. Very cool, super interesting. Dan, this has been fascinating. I really appreciate your time. Super excited to see this go live, but thanks so much. I could go on talking to you for a long time. You have such interesting takes. So thanks again.

DR: Thanks so much. Thanks for having me on. Big fan of The Defiant, of course. 


Share

The Defiant is a daily newsletter focusing on decentralized finance, a new financial system that’s being built on top of open blockchains. The space is evolving at breakneck speed and revolutionizing tech and money. Spread the word and share!