Trustless is deeply similar to both RAI and LUSD.
Trustless is related to RAI because both systems have a controller which attempts to correct excessive supply/demand by disincentivizing that side of the market. The HUE controller pays borrowers if HUE is overpeg, while the RAI controller pays borrowers if there is more demand to long RAI than short. The fundamental difference here is price vs quantity. RAI pays borrowers in price, while HUE pays borrowers in quantity.
The problem is that a pure quantity based mechanism (like Trustless) requires a source for where that quantity comes from. Trustless’ first attempt at solving this problem comes from the protocol reserve. The problem is that the protocol reserve can run dry. Now if we expect a positive rate regime to be common this is less of a problem, as positive rates fill up the reserve. I don’t expect a positive rate regime to be common. Why is this the case? Opportunity cost. The Ethereum staking rate means that anyone longing ETH should get at least ~4-5% of yield. If this is the case, there is only one option for a pure ETH only system. Negative rates that match ~4-5% of yield. Forever. RAI has run into this problem and has negative rates of around ~4-5% once you consider the risk of liquidation, governance etc…
One potential opposing point to my argument is that ETH staking is not risk free. This is true. But the “lived experience” of systems like RAI and LUSD show us that the demand for decentralized stables is way higher than the supply if you only have ETH as collateral. Onboarding Lido’s stETH or Rocket Pool’s rETH as collateral would greatly help Trustless prevent a negative rate regime.
This is something I have also been wondering about. Allowing for multi-collateral ETH staking derivatives in a borrowing protocol with a stablecoin is essentially what Gravity Protocol is planning to do.
In my mind, the question about whether or not to pursue this comes down to a philosophical tension between the Trustless ethos vs. what is necessary to be competitive (as Miles said). Risk is the other important consideration.
Introducing staked ETH as collateral might cause some to question Trustless’ mission to “build a truly decentralized lending protocol that prioritizes trustlessness and community ownership above all else”. However, if the opportunity cost of disallowing Ethereum LSDs (liquid staking derivatives) represents a high enough existential threat to Trustless’ ability to remain competitive, then perhaps it is necessary for Trustless to allow it. I think it would be extremely important to have risk assessment done by an outside firm, perform market research of potential users, and wait until after the Shanghai upgrade before proceeding with this though.
If Trustless is considering this, it would be more aligned with the ethos to allow LSDs from “decentralized” staking protocols. Rocket Pool is an example of that, with the upcoming Swell Network and Stakewise V3 also claiming to be decentralized.
It seems that what Miles and Lucas are suggesting is to stake or lend the Trustless-staked ETH somehow. I’d be interested to know what ideas they have for this.
Just a note, Rocketpool relies on the Oracle DAO (a group of trusted people) to provide beacon chain data to the smart contracts, so it’s not really that censorship resistant or trustless. Still leaps and bounds ahead of all of its competitors though
Vitalik himself recently posted in the Reflexer forum, sharing ideas for how to deal with the challenge of RAI CDPs struggling to compete with ETH staking as a source of yield.
tl;dr - If I’ve understood correctly, Vitalik proposes Reflexer to “create a copy of Lido where users who create CDPs can stake those funds”. This means Reflexer would effectively become a liquid staking derivative protocol where “the stablecoin that the scheme maintains is the (stabilized) liquid staking derivative”, instead of ETH-pegged LSD derivatives that the other major providers offer (stETH, rETH, ankrETH, etc.).
Becoming a de-facto liquid staking provider seems like quite an ambitious pivot to me. For example, how would the validators part work? Lido contracts out the running of their validators to various organizations (29 node operators for Ethereum). That would be a major undertaking for a newcomer project and could raise questions of centralization. Unless Vitalik had some other method in mind?
Also worth noting that while Vitalik doesn’t seem to like the idea of using existing staked ETH derivatives as collateral (based on the Ethereum-level systemic risks it poses), there are CDP projects currently planning on doing just that. There is Gravity Protocol which I mentioned in an earlier post, multi-collateral RAI forks such as TAI, zero-liquidation “hyperstaking” protocols like Myso Finance, and more.
Regarding alternatives to allowing stETH as collateral, Ameen said this in the Reflexer discord:
That strategy raises questions for me about rehypothecation, yield competitiveness, and liquidations, but it is interesting.
I summary, I think there are a few ideas here that each have their own tradeoffs but could be added to Miles’ list of ways to make Trustless more competitive (with modifications of course, e.g. for Uni V3 TWAP oracles).