Gyroscope – The All-Weather Stablecoin

Sunny: Welcome guys to the show. Glad to have you on. My friend Leland has spoken and he talks a lot about you guys and he was like, you got to get these guys on the show. They’re amazing. I’m like okay, let’s do it. So excited to have you guys on finally.

Lewis: Oh, thanks yeah. Thanks Sunny. It’s a real pleasure to be here and thanks also to Leland for introducing us.

Sunny: Before we dive into Gyroscope. Why don’t you tell me a little bit about how you guys got involved with crypto? Is this your first project, or do you have, what were you doing before?

Lewis: Yeah, sure. Well, so myself, the journey started on one fairly decisive day in October, 2014 I think it was, when Vitalik was doing his Ethereum roadshow and he came to Cambridge University and he gave this talk to what was a relatively under attended event I would say at the time.

It was just incredible looking back. He talked about smart contracts and all of these concepts that we now know very well and understand. I was very intrigued, but just sort of couldn’t really wrap my head around what he was proposing at the time.

But that was really the start of my interest. Since then, I learned a lot. I went on to do a master’s in economics. While I was doing that. I continued to really research the space and focused on many projects as they developed.

Lewis: I did a master’s in Economics and worked briefly in the city as an Economist, but my interest in crypto was still very strong. I knew that once I caught the bug, I couldn’t really go back. I got to a point where I saw there was this PhD position advertised at Imperial College, London.

This was in 2018 or so when I started this PhD. So my PhD has been focused entirely on DeFi. When I started it, it was really in the deep depths of a bear market. It felt like quite a contrarian bet going into it, I suppose at the time.

And published a number of papers on this area, on risks and decentralized finance. Then at some point, I met Ariah and we embarked on this stablecoin journey.

Sunny: Ariah, how about you?

Mundt: Yeah, my background is in math and after undergrad, I did a bit of work in financial software for a couple of years. And decided to go into a PhD program after that, because I was working on some side projects around modeling complex financial systems and decided that was the more interesting area compared to the work that I was doing.

Then went to Cornell and part way through the PhD that around like 2017, this is when Maker was coming out. I got very interested in these more complicated DeFi things being built, but before it was really called DeFi and merged half of my research into that area.

So several stablecoin papers and DeFi papers have come out of that since then. This project emerged from a paper that Lewis and I were working on in summer 2020, and then we jump started this project at the IC3 bootcamp, following that paper.

Sunny: Oh nice. I love those IC3 bootcamps. I went to the one in 2017, and that was like I just met a lot of people who have been like close friends in the space from those. Is that where you guys met as well? Or how do you guys like to start work, decide, meet, and then decide to start working together on this project?

Mundt: We met the previous summer. 2019, I was spending some time in London where my advisor is sometimes based, and we met at a PhD meetup here, I think.

Sunny: And Lewis, you mentioned you were doing economics. You’re practicing Economics. Was it like monetary economics or was it like something else?

Lewis: My main area in economics at that time was like econometrics and micro-econometrics. Yeah, really concerned with finding causal impacts and trying to understand exactly how certain types of economic systems work.

In my work as an economist, it was, yeah. One of the main areas of focus was on litigation type work and also had a lot of exposure to regulators, financial regulators. It was that thing.

Sunny: Do you think any of it really carries over to what you’re doing now?

Lewis: There are two parts to that. Of course, the formal training in economics, well I should probably say so. Actually, right back in the very beginning of this journey for me, I did a degree in philosophy, politics and economics, and it was during a time of I just after the financial crisis and went into this.

In this attended many talks and there were many lectures where everyone was giving these talks about the misbehavior of banks. This is of course something that has really been at the center of Bitcoin itself, where I’m one of the first block headers.

There’s this statement about the chancellor declaring the bailout of the banks or something like this. But yeah, there are two parts of my main background in economics of course that have carried over immensely currency models.

Lewis: And this thing is really the bread and butter of degree courses and master’s courses in economics. Then on the professional side, the focus on engagement with regulators has also really shaped how I see this industry and the role that we should be playing in this industry as well through whatever reason I’ve engaged with a number of regulators in different areas, not just financial services actually.

That one of the main things that I quite passionately believe is that regulation isn’t something necessarily to be feared. In fact, it’s something that for, well-functioning markets actually helps enable these markets. It really can play a very important role.

Of course the challenge is to make sure that it’s playing the right role and that it’s an industry that you can grow together like industry and regulation is the things that they co-evolve. That’s one of the main beliefs I took away from my professional experiences as an economist.

Sunny: Nice. So then, how did this lead you guys to build a stablecoin? Like there’s so many different interesting things going on in DeFi from like Dexes to lending to pure Ponzis, but, why stablecoin? Why did that, why was that the thing you guys decided to do, okay, this is what we want to go in and fix?

Mundt: Well, actually, a stablecoin is something that pops out of what we have in mind building, but there’s like a number of other things that go along the way as well. Basically, we see a need for robust infrastructure just like connecting various things in various different DeFi protocols, making composability, but with risk control, an idea as well.

One of those things is a stablecoin. Another thing is what we consider like a robust Dex structure. Something where if there are issues with a certain pair that you can still route through the Dex to do swaps that you might want to make without the problem with something like a central hub asset or something.

Sunny: I see. Interesting. We have to dive into how that works. Yeah. Why don’t we just start off with like this, the brief summary of what the Gyroscope protocol is, and then we can dive in from there.

Lewis: Sure. The gyroscope stablecoin is an autonomous stablecoin. It’s one that we can primarily think of as consisting of two components.

Sunny: When you say autonomous, what does autonomous mean?

Lewis: It runs like by itself without necessitating constant human intervention. Just to add to that, I suppose, otherwise it will be ambiguous. Not to say there’s no role for governance, but just that it’s not a second by second or block by block process.

The stablecoin, you can think of it as comprising two main elements. The first element is a reserve that is fundamentally as diversified as possible. We try to diversify all risks that we can think of in DeFi. If you can think of the risk and think of some way to diversify away that risk, we will attempt to do that in the reserve.

It’s constructed so that it contains multiple assets, that if there are certain types of failure modes, the first of course is price risk, or other stablecoins going through de-pegging events or crashes or this thing.

This is one type of risk, but of course there’s also regulatory risk is another major one, particularly in the stablecoin space.

Governance risk, plenty of risk factors. So what we try to do with the reserve is separate assets into different compartments. That these risks are diversified to the greatest extent that we can.

That’s really the first thing about this reserve. The second thing is that we have an innovative approach towards constructing primary and secondary markets. Perhaps, yeah. Ariah maybe wants to go through this one.

Mundt: Yeah, sure. Here, first worth highlighting, what is a primary market? What is a secondary market? Because sometimes the naming can make it a little counterintuitive, but this is coming from like ETFs and traditional finance.

Basically the primary market is if you’re minting or redeeming a stablecoin or in the case of ETFs, if you’re minting or redeeming a share for the underlying assets in the ETF. Then secondary markets means that you’re trading shares, or trading stablecoins that have already been minted and trading it with someone else.

So actually the problem in the naming here that gets confusing is that the primary activity, as in like most trading activity usually happens on secondary markets, not primary markets. So that can be a little confusing.

But basically, what we have in mind is bringing this structure into the DeFi space with the stablecoin and basically designing the structure of this primary market in the right way that helps to reinforce liquidity in the stablecoin, but doesn’t take charge of like where most trading activity is supposed to happen.

It’s just supporting the secondary markets and the secondary markets are then taking the information from the primary markets. In particular you have pricing balance coming from the primary market then, and then you can construct these secondary markets as AMMS that concentrate liquidity within specific ranges.

Sunny: Cool. Okay. Let’s maybe start then with treasury first or the reserves. Just start off with, what collateral do you guys accept currently? The goal is to maximize the, or minimize correlation risk. What is the current collateral that’s being used?

Lewis: Yeah, so primarily stable coins, but not exclusively. Yeah, you said the goal is to minimize correlation risk. That is true in one sense, but it’s not an explicit objective, because it’s not just about minimizing correlation in returns or prices or something like that. We do want to control for other risk factors.

Mundt: It’s more about tail risk, correlation in tail risks.

Sunny: You mentioned a couple of different ones like Oracle regulatory. Sorry. You mentioned regulatory and governance. What about Oracles? Do you think, I feel like within the DeFi space, one of the biggest things is this heavy reliance on very few sets of Oracles especially like often chain links. How do you deal with this and nudge the protocols that you’re integrating with towards more de-correlated Oracle risks?

Lewis: There are a number of things. The risks posed by Oracles and we’re well aware of, and we’ve seen many bugs have occurred due to problems when over reliance on single Oracles. In the first instance, we will go with things that are tried and tested for the Oracle approach, because also with a system like ours, it’s relatively complex in terms of code base.

There will be this bootstrapping phase where we have to roll things out and make sure that we sort of, don’t try to do too much innovation all at once, but actually we have put quite a bit of thought into ways to control this Oracle risk.

In particular, trying to find ways to bound the behavior of Oracles and see if we can set parameters around what we would consider reasonable input feeds. This is certainly on our list of things to do here.

Sunny: You mentioned the stablecoin, so do you have a list like which stablecoin particularly you accept right now or you plan to at launch?

Mundt: Right. It’s not live right now of course. We haven’t come out with the specific portfolio that it’ll start with. The space can evolve significantly and even just like a couple of months here, before the actual launch.

We do also want some flexibility there, but we see at a high level that the portfolio is just split between the various different risks. One of the top risk distinctions is between custodial assets and non-custodial assets.

With the custodial assets having a lot of counterparty risks and regulatory risks more so than the non-custodial ones. But then the non-custodial ones having more Oracle risks, more governance risks, or at least on-chain governance risks a little bit different from centralized governance risks.

As well as just like the pricing mechanism risks what we were seeing in Maker back on black Thursday, what we were seeing in like the crash of all of these algo stablecoins recently.

At a very high level, you would think of the portfolio as like first separating between these custodial and non-custodial risks and then segregating the risks to the extent possible. The simplest way to visualize is the hierarchical ordering of risks. It may not be exactly that, but that’s easy to visualize.

Sunny: Do you guys have a mechanism then also to target a specific ratio of assets in the reserves? Like we can go ahead and say that, hey, okay. Let’s say it accepts USDC, USDT, and DAI as the three things.

And we accept all three, but then what happens if our reserves just end up becoming 99% USDC because that’s all people want to deposit. Do you have some mechanism of making sure to incentivize balancing these out?

Mundt: The idea is that their desired portfolio weights. And these weights do change over time naturally, but just as the price of those assets change. You can think of it like how cap weighting might change the weights in an ETF.

It happens in a passive way as those assets are repriced. Then at some point in the future, governance might come in and say, actually we think the new desired portfolio weights are slightly different. Then there’s like a rebalancing event to that.

How the primary markets work essentially is there’s these desired portfolio weights coming from how it was initially set up to segregate risks. There’s some ability to go outside of that, but not too much. Eventually you pay for going too far outside of that.

So basically if you get to the point where you’re getting too far from those weights, then it’ll be more efficient to instead of just coming in. Say, it’s overweight a little bit in USDC.

Instead of coming into mint mara Gyro dollars with USDC, it’s probably going to be more efficient to come in and mint with DAI instead. But maybe you start with USDC, but you swap to DAI somewhere else and it’s more efficient to do that routing.

Sunny: Got it. Okay. Great. That makes sense. That’s like from your blog posts, you talk about like the three layers of defense in maintaining stability. The first one is just these like reserves. And like in a world where over collateralized or at collateral, great the world’s easy.

If you want to get out, you just swap it for $1 of the other thing. Obviously the fun part of stablecoin is how to design for the other world where we are going into a risky scenario. So this is where the primary, the PAMs and the SAMs come into play. Okay.

Would it be fair to say that the primary, so if I’m getting this right, the primary automated market maker is, would it be fair to call it more like a bonding curve that allows you to mint and burn new Gyro dollars or like collateral tokens that you deposit?

Then the secondary automated market maker is something like a Uniswap or something where the Gyro dollars are trading on what we traditionally think of as an AMM?

Mundt: Yeah, you could say that. Basically, it’s the bonding curve with the specific shape that we’ve encoded and the shape changes depending on the health of the system is the important thing. The idea there is that it’s a sustainable curve, always taking into account the health of the system.

Then there are two curves, actually, one for redemption, one for minting. This creates a spread in the primary market and this spread, if the secondary market is behaving smartly, we’ll take that spread into account in deciding how to price things.

Sunny: And is the secondary market kept in line with the primary market by the protocol itself, or does it just depend on external arbitrageurs to arbitrage the secondary market into alignment?

Mundt: Basically, if the secondary market is ever trading outside of the bounds from the primary market, then right, there’s an arbitrage opportunity and somebody will be running a keeper that brings it back in line is the idea.

Sunny: Okay. But the protocol itself is not actually ever like, it’s not like Fay or something where the protocol itself is market-making on the secondary market.

Mundt: Well, Fay is an interesting case here where their secondary market and primary market is a bit confused into one. And so we are just making it explicit. What should be the primary market structure? How should that influence the secondary market? Then it’s a much cleaner and much more fundamentals driven approach for how to build it as is our idea.

Sunny: Got it. Okay. What’s the benefit of, do you not like to end up losing value to arbitrage instead of letting the protocol do the market-making on the secondary market? Why did you choose not to do something similar to Fay? How would you compare the pros and cons of the two models?

Mundt: Basically, the primary market is market-making, but in its own AMM. Basically, actually how it will work out if you’re coming to the stablecoin market, as a user, looking to buy some Gyro dollars, you will first implement this on a balancer.

You’ll go to a Balancer exchange. The Balancer exchange has a smart order routing mechanism, both this Pam and the Sam will be built into the Balancer V2 system. Actually, this smart order routing algorithm will take into account how much you’re trying to swap and the resulting gas fees estimates.

It’ll route your order in the most efficient place. Usually, that should be the secondary market. If it’s a bit out of alignment, you may actually be going straight through the primary market.

If for some reason, your order is too small, maybe the gas fees on the primary market are a bit higher than on the secondary markets. Maybe you still go through the secondary market and then there’s a rebalancing opportunity from an arbitrager.

Sunny: Got it. Okay. Then let’s talk about what is the exact mechanism of how this helps maintain stability when the reserves go under collateralized? Let’s say the value of the reserves USDT declares tomorrow, psych guys, we actually had no reserves the entire time, and the value just goes to zero.

Now let’s say your value of your stable, the reserves goes to like 80% of what the outstanding Gyro dollars are, what is the mechanism here? What happens next?

Lewis: One thing to immediately say here is that in this situation, this is already very extreme given our design. We would like at the point that it’s possible to be at the collateralization ratio of 80% or so.

This means that the first defensive line, which is this diversified reserve, means that this has already been well breached, to continue with the analogy.

Sunny: What is the maximum that you’d be willing to let something like USDT become in your reserves?

Lewis: As a proportion? Yeah. It’s too early to say at the moment, but this reserve itself is intended to also be yield accruing. It’s not just that we can only go, we have like the say, a straight value of the capital and any drops will harm us.

We’ve got this additional buffer as well built into the reserve. When we first started talking about this, we were using this term like rainy day fund. Having some capital in case of a rainy day.

This, just to say that in the situation that you’re describing, where we’re already and hopefully what is a quiet part of the tale risk for us. But then sure, let’s say we are in the situation where

Sunny: Just really quick, where does the yield come from? Are you taking the collateral that’s put in and then throwing it into yearn or something?

Lewis: Yeah, so we will be simulating what safe levels to do with different parts of the capital. The most important thing that we will not compromise on is putting the reserve in any dangerous or vulnerable position.

This is a feature that, again, we expect to be rolling out very slowly and possibly not in the very first version, but yeah. It will be through integration with other protocols where it’s possible to be earning yield and with Balancer, this sort of thing.

Mundt: The idea is basically that the portfolio’s supposed to segregate various risks, but some assets have very similar risks and you can deploy those assets together in certain pools to earn a yield without bringing in much more risks to the portfolio. So the aim of the portfolio is a good risk adjusted, deployment of assets.

Sunny: Cool. Okay. Then let’s go back to the, alright. We’re at 80%. It is the doomsday scenario. What now? What do?

Lewis: Yeah, just wanted to highlight the extent to which we hope this is unusual, given the design, but then given we’re in this situation. What we fundamentally do is enable agency-like coordination on the possibility of the stablecoin returning to peg.

We leave open in design the possibility of this equilibrium of the peg like re-pegging occurring. One of the mechanisms that we use here is baked into the primary market AMM itself. This is where we will essentially, in times of heavy outflows from the protocol, the redemption rates will be made less favorable.

We’ll start to sort of, in some sense, impose penalties on redemption, but we would never stop redemption. That’s also an important point to make. In the worst case, it will always be possible to redeem it, just the net asset value of the reserve.

We will, we’re trying to engender these like dynamics in the market, that means that we only would get down to this point very slowly. In doing so enable agents to coordinate around this belief that the currency will return to the peg.

Mundt: I can fill in there if you want. But I had mentioned before that this primary market bonding curve shape basically is dynamic with the health of the system. The idea there is that if you do have large shocks to the reserve the setting we’re describing here, you want to open up the possibility that a peg can be maintained because there’s enough like natural economic usage, economic demand for this as a currency.

And be able to support that up to a point. Then after that point, there’s like a circuit breaker effect. The redemption price decays down to the always sustainable level basically.

This is to basically deter banks and deter speculative attacks and just ensure that the system actually survives and has a chance to recover. And supposing it’s, that gives it time. Then all the other mechanisms can basically help the system recover is the idea.

Sunny: Git it. Cool. This goes into another thing, which is this natural demand where I feel like in stablecoins, you can come up with whatever algorithmic trickery you want. At the end of the day, stablecoins to work really need to have real organic demand and usage and adoption. Right?

I don’t know, in my opinion, I feel like so far I’ve only seen two non-custodial stablecoins that I’ve actually managed to, seem to have gotten like legitimate organic demand, or that’s probably like DAI and Terra.

What would be your plan to, how do you guys plan to drive legitimate demand for this stablecoin, for Gyro dollar so that when you get into a scenario, there is this expectation and reason and demand for it to go back to $1?

Lewis: That’s really answering that question. That really brings up many different topics. Firstly, yeah. Firstly, if you have the option of having an asset that is as low risk as you can go in DeFi, that already seems to be offering improvement over existing non-custodial stablecoins where the risk is concentrated.

This already is itself just a massive improvement, having a reserve back stablecoin like this, where the reserve is specifically diversified to the greatest extent that you can do so. Then we have a number of core ideas that we intend to go with which will enable us to bootstrap usage.

But before talking about it, is to jump the gun a bit because something that’s very important to us is, for a stablecoin legitimacy and the stablecoin is a super important facet of it. And this is something that’s like saying a fundamental value of this project is to ensure that we achieve rapid decentralization from the beginning.

We already have a very active community on Discord and elsewhere. It’s like helping us to formulate the ideas that we want to be going with and shaping a stablecoin fundamentally as a public good for DeFi and for the DeFi community.

And saying that, because that’s a real caveat to the fact that where we do have these ideas to bootstrap economic usage, which is the most important type of usage that will help boost demand for us, or one of the most important, this is all very much subject to community input and so on.

With that caveat out the way, one of the ideas that we are interested in at the moment is possibility of a UBI type program where the some amount of tokens could be distributed to community members and on some frequency in the future to be determined and referring here to stablecoin as opposed to any other token.

The idea being that we can somehow try to encourage organic growth in how the stablecoin is used and in particular, in places where it really is of fundamental value to two different market players, to be able to use and have access to a stable currency.

In unstable economies and places where inflation rates are very high and it’s difficult to get access to a stable currency, these are really for us, this is the level of values. This is a very important area that we would hope to encourage.

Mundt: Another important area here is a kind of more, almost B2B area of basically demand from DAO treasuries for stable-ish assets to keep the treasury.

Sunny: One thing, that would be, let’s say I’m a user that has a lot of USDC and I am sufficiently confident in USDC is like stability. By swapping my USDC for gyro dollars. Then the protocol is taking those USDC and putting them in like yearn or whatever, and keeping the yield as part of the reserves.

How do you drive them? How are you going to drive demand for people to actually do this? Like giving up the return that they could be getting on the USDC and instead use it to contribute to this public good which is like growing the reserves?

Mundt: This is where it makes sense to talk about a distinction of the use cases. Most crypto use cases today are fairly speculative in nature, as opposed to you’re saying these are like real economic adoption cases.

And realistically those speculative cases, although like if that’s the only thing happening it’s a little bit self-reinforcing of course. At some point you have to have actual economic adoption of something for the whole chain of things to make sense.

Realistically, today the first adopters and first use cases are going to be in speculative uses. Then once you’ve, our aim is basically to make the first speculative uses, make the system more liquid and more able to function and provide various services that the people would want.

Then like once that’s established as a baseline, try to bootstrap economic usage on top of that. And so the first use case there was our building out of the gate, is this idea of, you can compare it to a money market fund replacement. It’s quite different in structure from a money market.

If it’s the same sort of aim basically you want a fairly stable asset with the aims of getting some good risk adjusted returns. The idea here is that this, you would mint the first Gyro dollars, because you are going to provide liquidity across these different Sams, the secondary market AMMS that we’ve set up.

The secondary market AMMS are basically forming a network within the Balancer vault and connecting trades, making more trades like more efficiently possible within balancer. If you’re a user of balancer, and this is all set up, you’re going to come to balancer, you’re going to use their smart order routing algorithm.

You’re going to be routed more than the most efficient path. Probably without even knowing it, you’re going to be going through two of these Sams or something. Basically, compare this with Curves, and like their Three Pool for instance.

The problem is if you’re providing liquidity to this Three Pool, you’re composing all of the risks of all of the assets in that, and all of the places those assets that are being deployed.

For instance, well in the Three Pool, you’re certainly taking on USDT, USDC and DAI all of their risks. If anything breaks, the whole pool basically goes to zero. In some of the other pools, you also layer in Compound and Aave on top of that.

This means that it’s not actually like a great money market fund strategy, because you’re actually getting fairly bad risk exposure, but we’re setting up this network of Sams in a way where if you provide liquidity across like a number of these different Sams, your risk exposure is actually quite contained. The same idea as this, all-weather reserve that we’re building.

Sunny: Yeah, the difference here is that in Curve, they are allowing people to trade against the pool. Using this network of Sams is sort of, you’re saying that this is a way of replicating that functionality that curve provides?

Mundt: Yeah, exactly. You can also make sense. Basically, let’s go into what Sam is I guess. A Sam is basically a specialized pool where you’re trading Gyro dollars against another asset.

One Sam would be like Gyro dollars – USDC, another Sam would be Gyro dollars – DAI, another Sam going to a BUSD maybe, and you can even have a Sam that goes to Eth, which is a bit harder in Curve.

Now they have their tri-pole, but basically it’s, the Sams are working because there’s this primary market structure. As long as there’s a primary market for this other paired asset, you can do basically the same ideas in constructing the strategy of this pool.

Then you would basically be the idea that you would compose these trades in these different Sams. Say, you’re starting with USDC and you want it to swap to be USD. Maybe there’s a pool where you could directly do that.

Mundt: Likely it also includes a bunch of other risky assets in it. If you’re doing it through the Gyroscope, a robust Dex idea, you would compose two Sams, you would compose a swap to Gyro dollars and then compose a swap in the other Sam from Gyro dollars back to the USD.

This has the benefit that if you actually have some real market turbulence, and one of these other assets actually breaks, Dex still functions, whereas the Curve pool is gone.

Sunny: I’m sorry, just so if I recall the Sams exist on, the primary market maker exists as a bonding curve, but the Sams exist on external AMMS like Balancer or Uniswap and stuff.

Mundt: They’ll both be integrated into Balancer V2. It’s just that the Pam is something where if you’re a liquidity provider like an individual, you can’t provide liquidity to the Pam. That’s only the protocol itself.

It’s going to be a Balancer pool where, not anyone can provide liquidity, it’s for the protocol itself, but it’s built into Balancer so that you can easily compose trades and easily route through the right pools. Then if you’re an individual liquidity provider, you would be considering these different samples.

Sunny: I see. And the protocol is not the one market making in those samples?

Mundt: No. Basically, the protocol provides these pricing bounds through the Pam. These pricing bounds influenced how the Sams should be shaped and either you’re coming into Balancer to do a trade. You’re being routed to the most efficient venue.

If it’s a large enough swap, it might be the Pam if the Sam is out of alignment, but it’s usually the Sam. But if you’re too small of a trade, maybe you still go through the Sams and then there’s an arbitrage opportunity for a keeper to come along later.

Sunny: Why do you believe that the Same will become highly liquid? If your goal is to make this be this common, one of the most popular liquidity routers, are there incentives for people to add liquidity to these samples?

Mundt: Well, the precise form of incentives is certainly TBD. The long-term idea is that as with any LP position to be sustainable long-term, there should be an organic trading demand to use these pools. So that’s the aim basically is to make these profitable LP positions that people will actually want to use because it’s providing a useful service.

In this case immediately connecting new trade possibilities in Balancer. Eventually, also like hopefully there’s an organic demand for usage of Gyro dollars itself and natural demand to seize the Sams, just because they allow you to use Gyro dollars and move from assets into Gyro dollars or out of Gyro dollars.

Why it’s going to be liquid at that start, comes down to how we’ve designed the system and the reserve also. Basically that’s why, it makes sense to start with a reserve that is mostly other stablecoins.

Maybe there’s some other assets in the reserve at the beginning to the extent that it makes sense, but because the system is just starting out, it’s a lot more sensitive to price risks. Unless something really bad happens really soon is going to be guaranteed basically like a hundred percent reserved or slightly more.

As it earns yield, then the pricing bounds from the Pam are going to be quite tight and you don’t actually even have to have that much liquidity in these Sams to have good liquidity and have good trades possible.

Sunny: You guys also talked about a couple other forms of stabilization that you’re looking into leverage loans and stuff. Can you talk a bit about that?

Mundt: Yeah, sure. The best comparison here is to draw a comparison with Maker. And Maker on Black Thursday encountered this de-leveraging spiral, the short squeeze effect on DAI that brought the DAI price up to over a dollar ten.

This had the effect on faults that the people getting leveraged and in Maker by minting DAI that in a crisis where they actually have to de-leverage, they have to buy back DAI on the market at this very large premium, and actually they ended up losing more money in this de-leveraging process than they thought they would have because it’s supposed to be a dollar asset.

But instead there’s a short squeeze effect on DAI. All of a sudden to de-leverage the same amount. There’s a like 12% premium. And Maker since then, and this has been one of the big motivations for the PSM in Maker, has introduced this PSM that basically ties the DAI price to one USDC.

This is basically their form of introducing a primary market, I would say. Their primary market is just adopting USDC’s primary market. Where now, if you’re in this crisis and you’re a vault, you can come in with a, if you have dollars, you go to circle, you get a new USDC, you bring it in, swap it to DAI, and de-leverage your position.

You don’t have to worry about these de-leveraging spirals, but now you’re importing USDC risk to a very large extent into Maker itself.

That’s motivated to Gyroscope actually, because this mechanism that we’re building is basically like if you were to try to do something like the PSM, which is making this primary market structure in as decentralized as like a risk-adjusted way as possible, that’s our idea for gyroscope basically.

So gyroscope is making a better PSM, which means we can also integrate with something like Maker to make a leveraged loans backing for part of the supply. So the idea would then be that part of the Gyro dollar supply would be backed by leveraged loans, just like Maker.

Part of it would be backed by the primary market mechanism of this PSM replacement so to say. And then you have basically similar security to Maker, but maybe a little bit better because you have this stronger, more robust primary market mechanism.

Sunny: There was also a proposal I saw of considering getting Maker to use Gyro dollars as part of their PSM. Is that orthogonal to this, or are these like related in some way?

Mundt: It’s the same idea. Basically, if we can work together with Maker to just make one thing, that would be ideal.

Sunny: Right. Got it. Is that in progress or like what’s the status of that?

Mundt: I don’t think that there was much like immediate interest just because it’s like so far in the future, we’ll see when things get rolled out and what the interest is at that point.

Sunny: Got it. Okay.

Mundt: It’s an idea that I’m quite excited about.

Lewis: We’re also like as a general point of principle, very excited about integrations with other protocols and believe very strongly in like cooperation in this space. Yeah, I personally feel, Maker was one of the key projects back in 2017 or so that really piqued my interest in this area.

We would, we’d love to integrate with them yeah. Back when they devised this CDP mechanics and their naming scheme, all of these things it was this was some real ingenuity and big fans.

Sunny: Let’s talk a bit about your governance. You guys also have been doing a lot of work on making your governance protocol for Gyro dollars, really resilient in a lot of ways.

Tell us a bit about the two big key words that are bolded in the dock are conditional cash flows and optimistic approvals. You want to tell us a bit about these?

Mundt: We’ve done a fair amount of research on the issues about decentralized governance in particular, when you have pseudo anonymous people and you, the goal is not to have the, not to rely on the backing of the legal system to work out something if there’s a problem.

And several things that come out of this basically is that there are actually issues that are very parallel to some things in corporate finance. It’s just amplified because you don’t have the protections of the legal system guaranteed or guaranteed to some extent, which is some probability.

These two issues that are along the terms of, along the lines of short-term decision-making so things that are maybe good for a protocol revenues that can go to governors short-term, but at the expense of the long-term health of the system. Then also basically, like governance attacks or rug polls on the system that might be attempted by a governance community.

That’s influenced us to design these two mechanisms that you mentioned. The biggest one, what I consider the most interesting and innovative one really is this optimistic approval mechanism. What it means is that on a structural level at least, you have some party that is delegated to be able to make decisions.

You have another party that’s this guardian role in the system, and it has this optional veto right that it can exercise if it doesn’t agree with the decisions made by this designated party. This can fill in two important places in a DeFi protocol and not just Gyroscope actually like it applied in DeFi more generally.

One of them is streamlining governance, which we’ve seen as a very important issue these days, because of basically voter apathy and just too many things to vote on, too many things to decide. Basically, you can use this optimistic approval mechanism to delegate to a specific group of a subgroup of governors.

These governors can make decisions and post these decisions. And basically, then there’s a time lock period during which the whole governance community can come in and say, actually, maybe I don’t like this decision.

If enough of us feel that strongly about it, then we can exercise this optional veto to stop it from actually being implemented.

In that way, usually this veto right so you’d expect if you made the right decision about who’s delegated, that it’s rarely exercised, but it’s really only exercised if there’s like very bad decisions being made.

This is why it can also be applied to prevent rug pulls actually, because you can make the guardians of this actually be the real users of the system like the stablecoin holders, for instance. Let’s suppose that the governance has proposed a change that would effectuate like a rug pull.

Now, there’s a time lock period. As long as there’s some Gyro dollar holders like paying attention to this, and then raising the alarm that, hey, this is a rug pull, then the idea is that there’s this opportunity during this time lock for enough Gyro dollar holders to come together, exercise this veto right and block the rug pull action from actually happening.

This just changes the dynamics of this governance game theory, basically, where now if you’re a governor and you’re thinking about doing a rug pull, you should be skeptical there’s actually going to pass. That there’s like a significant probability it’s going to be caught.

There’s this optional veto that’s going to happen. Actually, you’re just going to hurt the value of your position, because now everyone is like, oh, maybe I don’t want to use the system because look what the governor’s trying to do. So I’m going to exit and use a different system.

Sunny: That’ll make sense. It’s similar to the role that guardians play in a lot of governance protocols, but it’s like the idea of, hey, we can give this guardian power to the stablecoin holders I guess. What about, the part that I was actually more interested in was the conditional payment.

Because I thought that was really cool. So let me see if I understand this correctly. What happens is that there’s like some cash flows. First of all, where are the cash flows coming from? Those are coming from the yield strategies?

Mundt: Basically, there’s a number of places that cash flows could come from in the future. It really depends on how the community decides to evolve this system, of course. Just to point out a few, there are some tasks to govern.

One of them is the slow update of the reserve portfolio design as the DeFi space evolves. At some point they may be compensated for that by some of the excess yield from that portfolio. Other things are like designing the right, designing and implementing the right Sam structures and interaction with Pam.

At some point you could have a fee switch like in Uniswap where part of the fee revenue from the Sams might go back into the system by default, but maybe could also be used to incentivize governors in the right direction.

That’s where these cash flows could be coming from hypothetically at some point. The idea behind the conditional cash flows is really that these cash flows that are coming into the system, where should they go? And we’re saying by default, they should go to be an extra buffer to the reserve portfolio.

Then if the system stays healthy, like far into the future from when these cash flows actually came in, that means that the governors are doing, it’s an indication at least that the governors are doing a good job. Then at that point in the future, condition on the system remaining in this healthy state, then some of the cash flows should be unlocked to incentivize governance.

It’s really about expectations of future cash flows, conditional on the system remaining healthy, which then sets up this incentive dynamic that, hey, we actually have to be very serious about the system, designing it as best as possible so that it actually remains healthy.

Sunny: And so it would go to the current governors or the governors at the time when it was? So what I’m imagining is there’s like some sort of, let’s say a six month delay on like pay-outs. It’s, okay, we look at the protocol six months from now.

Then we decide, okay, it actually the governors from six months ago actually did a good job. Then we pay them out now, is that what it’s like? Or is it, or do we treat the governance token holders which we haven’t even gotten into the governance token, but, do we treat those token holders as relatively stable over time?

Mundt: The precise form of this can vary, and this is something that hasn’t been built yet. This is going to be one of the next pieces that comes out,  and also, it doesn’t matter to some degree. Like basically, if markets are efficient, which may not be the case right now in crypto, of course, but let’s suppose they are.

Then if you are selling out of your governance token position at some point, the buyer of that should take into account the probability that the pay-outs would actually happen from these conditional cash flows.

So you should get a fair price if it’s a fair market, but that’s really a long-term idea, of course, that these systems are actually like efficient markets.

Sunny: You guys have this live incentivized testnet going on right now. What is the, how does it work? And I saw that there’s different phases and it’s a bit gamified. What’s the goal of this? Is this to get people familiar with the UX of the protocol or is it to test out certain economic assumptions or what’s ultimately the end goal of this testnet?

Lewis: There are a few goals, really, the first is again about growing an active community of users from the beginning. It’s also been, so perhaps I should just, yeah maybe give a bit of detail about what the game actually is.

It’s very, at least visually it’s based on a flight simulator, but unfortunately there is little flying that you can actually do this. If you attempted to build a flight simulator on Solidity, I’m not sure what would happen, but I don’t suggest trying.

So that there is no flying, but the idea is that you go through these like two and a half levels we have at the moment where you can learn about the core mechanics of the protocol and get some familiarity with the how the minting and redeeming operations work in the primary market and these sorts of things.

By going through that, we hope that one of the first things is that people are able to learn about what we’re building and learn the language that we think fits us best. It’s also for us an interesting experiment and seeing what feedback we get and seeing what resonates with the community.

Also there are a few mechanisms in this very early testnet version that we wanted to get tested a bit in a live environment to see how they work and see if there are edge cases that emerge through time that we haven’t thought of up front.

It’s been a way for all sides in this to learn in the beginning. Yeah, it’s been quite a fun ride as well with this game. So yeah.

Sunny: Do you think you’ve found interesting economic insights, even though these are using valueless testnet tokens?

Lewis: From the testnet itself, I would say this is a question that’s like to be, the full answer to that is to be revealed later, because what we’ve mainly done so far is tested like the most basic mechanics in the system, but there are more sophisticated elements that we’re interested in testing out and seeing what insights we can get from that.

We’ve certainly taken many other very big insights into how to manage sudden influxes of users. We had an occurrence where we suddenly had just an absolutely insane amount of traffic, six weeks ago or something like this and this was a really real experience.

We have, so one of the things that we had to deal with, it’s a bit of a detailed remark, I suppose, but some of the faucets for the coven were a bit shaky. So we built our own faucet mechanisms and we had this massive influx of people wanting to try this.

We had to do on-the-fly, a debugging of what was going on with our faucet and really scale this quickly. We’ve certainly learned quite a bit from the process of having this in a testnet environment.

Mundt: One of the other main names of the testnet, is really to provide a setting where we can help walk users through how the system works. So this whole idea that I was explaining before about, what’s a Pam, what’s a Sam, what do they do, and how do you use them.

They’re opaque concepts until you actually like to walk through it and use them. You can actually do that on the testnet. That’s helped a lot in terms of the community, understanding how these mechanisms work at least on a high level.

Sunny: Nice. What’s the best way for people to start getting involved? Step one is getting involved with this testnet? You guys mentioned there’s a new phase that’s coming out the day that we’re recording this right now?