At a high level, Gauntlet believes that contributors are important to the growth and success of a leading money markets protocol like Compound. GFX has been a great contributor to the community and we are excited to see this proposal. We could be supportive of the proposal, but suggest a few changes to scope below.
We believe that GFX Labs can drive an impact on handling the operational overhead of protocol development. For example, upgrade testing and refactoring the developer code base are areas where GFX can make a highly valuable contribution to Compound.
At the same time, it’s really difficult to weigh in on a proposal with:
- Such a broad, heterogeneous scope
- No information on the intended methodology for these various initiatives
We suggest the following changes:
-
Remove items that duplicate work by other contributors. For instance, Gauntlet has already committed to managing the reserve factors and has published a detailed methodology here. OpenZeppelin and Gauntlet have already agreed to evaluate risk (market and smart contract) for new assets - but it would be great for GFX to do the dev work to create these proposals and set up oracles, etc.
-
Provide basic information on what success looks like for each initiative
- For instance - COMP reward management. It’s hard to support this in the current state as there’s no info on what this would entail.
-
Reduce the work to two key programs (a and b below) and allow the community to vote on them separately, as they are two very different services in nature. You could group the 15 work items into a couple of buckets:
a) Protocol development work
b) COMP reward management
c) Laundry list of parameter changes and management
GFX has strong experience with a), and it’s high priority for the protocol to address b). However, the addition of c) adds risk by increasing scope and it’s not clear why GFX might be successful with this work, as @allthecolors and other community members have noted. Even if someone had deep experience here, we’d hope they would provide more info on methodology and what infrastructure they have built to support the ongoing management of IR curves and similar quantitatively driven optimizations.
-
Provide more info on how GFX will work with the community to ensure they are building the right solutions for the protocol. For instance, if you want to list new assets, there might not be a need for supply caps at all. How will you prioritize these initiatives over the next year? How will you work with other community developers as well?
| Initiative | Bucket | Next Step | |
|---|---|---|---|
| 1 | End COMP rewards. There is ~$1B of capital in Compound providing little value to the protocol beyond printing a higher TVL at the cost of $100m in COMP a year. While the original intention of COMP incentives was to disperse COMP to its users, the incentives have brought in mercenaries that claim and sell COMP. | COMP Reward Management | Provide more info & methodology |
| 2 | Refactor COMP incentives to be used to bootstrap liquidity in new markets and scale back as liquidity enters. | COMP Reward Management | Provide more info & methodology |
| 3 | Seek liquidity from other protocols like MakerDAO via their Direct Deposit program. GFX Labs has successfully pushed forward a Dai Direct Deposit Module for Compound 3 that will conservatively bring to the protocol a 100 million DAI credit line that is insensitive to liquidity mining rewards. GFX Labs plans to propose a similar program for USDC and USDP, though they are more complex from MakerDAO’s perspective and require more time and collaboration to prepare. Compound currently spends $38m in COMP incentives on the DAI market and another $38m on the USDC market. Sourcing liquidity from MakerDao can lead to substantially reducing stablecoin incentives. | COMP Reward Management | N/A, this seems pretty straightforward |
| 4 | Add supply caps. While the protocol currently has borrow caps to lower governance hijacking risk, the protocol can similarly add supply caps to limit the amount of an asset that can be deposited to the protocol. Similar to MakerDAO’s debt ceilings, supply caps provide governance with an additional tool to manage asset risk. Adding a supply cap mitigates the protocol’s concern of an infinite-mint attack and empowers the protocol to onboard assets it might not otherwise support. | Development Work | Cut. Happy to discuss further but this just does not appear to be an additive risk lever |
| 5 | Add more assets. The protocol makes money by supporting markets that are in demand. With the addition of supply caps, the protocol can support more assets, since it can better manage risk. | Development Work | N/A, Sounds great |
| 6 | Improve the oracle system to support more assets. The existing oracle system only supports assets with a Uniswap v2 market. GFX Labs has been working with Chainlink to develop UniswapAnchorView (UAV) v3 contract, which switches the protocol’s anchor to Uniswap v3. While this does support more markets, it is still somewhat limiting. Adding Balancer and other markets as available anchors will grow the range of assets Compound can safely support. In addition, the protocol should develop tooling to support other popular assets such as LP tokens and token derivatives. There are a number of large (by MCAP) assets the protocol could support that the market wants to use as collateral, such as wSTETH, that could significantly increase protocol revenues. | Development Work | N/A, Sounds great |
| 7 | Deploy Compound v2 to L2s and sidechains. Interacting with the protocol is expensive on mainnet and can crowd out users who do not have the scale of capital to justify the gas expenses. Deploying Compound on Polygon, Optimism, and other networks would grow the protocol beyond Ethereum and put competitors at bay. | Development Work | N/A, Sounds great |
| 8 | Transition off Legacy CTokens. The ETH, USDC, ZRX, & BAT markets are all running on a non-upgradable legacy ctoken contract. While WBTC has already migrated, the other markets have not. The protocol is currently missing out on additional revenue by not being able to configure those markets with the improved interest rate model, set a kink in the interest curve for ETH, ZRX, & BAT, and protocol liquidation fee. Generally, this change is easy and only has upside for the protocol. | Development Work | N/A, Sounds great |
| 9 | Interest rate curves. Every interest rate curve utilized is the original interest curve chosen when the cToken was first configured (except DAI, which was updated on July 28th, 2020). This is a critical piece of Compound’s borrow/lend market, and yet the protocol has done little to research alternative curves or test the interest rate curves already utilized. | Misc Parameter changes and Management | CUT |
| 10 | Deprecate old markets. Compound v2 is nearing 1000 days old. The crypto landscape has changed significantly over the last few years, and just as Compound needs to adapt to what participants want today, governance also needs to offload what isn’t required or safe to support. Depreciating REP & SAI, closing down legacy markets, and managing risk on older markets such as BAT & ZRX is important to maintaining the protocol. | Misc Parameter changes and Management | CUT |
| 11 | Update the liquidation system. The efficiency of the close factor and liquidation system as a whole hasn’t been researched meaningfully or iterated upon. As a critical component of the protocol, it needs more dedicated resources. | Development Work | Provide more info & methodology |
| 12 | Upgrade testing and developer code base. While Compound’s codebase and testing was state of the art at its launch, it has since fallen behind. Without a clear, incentivized, and responsible party to improve and maintain the tools, they have become hard to work with and outdated, which has hurt development efforts. | Development Work | N/A, Sounds great |
| 13 | Clean up the existing protocol. The Comptroller needs to be cleaned up and possibly entirely refactored. Similar to the prior point, the codebase was innovative at launch but has since fallen behind without incentivizing someone to improve the system. | Development Work | N/A, Sounds great |
| 14 | Separate revenue generation and risk management. Reserve factors should be optimized to balance protocol revenue and liquidity. | Misc Parameter changes and Management | CUT |
| 15 | Rebalance reserves. The stock of reserves should be adjusted periodically to minimize risk. Compound is not in the business of taking positions and should seek to offload the surplus of volatile assets for stablecoins or even COMP. | Misc Parameter changes and Management | CUT |
We’re looking forward to working with GFX and the greater Compound community to land a scope and move forward here.
Separately, using KPIs to drive a performance bonus sounds like a great way to align costs to the protocol with value delivered to it. However, as @TylerEther and others have pointed out, the proposed bonus structure is not tied to the work GFX is doing any more than it is tied to the overall market recovering. Hopefully by creating a clearer, more focused scope with defined success metrics, another KPI to incentivize success will become an appealing option. The fact that the COMP price is an input to the bonus formula (and the bonus is paid in COMP!) begs the question - how is this bonus structure better at incentive alignment than just paying the team in COMP that vest over a long period of time?