Just wanted to comment that @fubuloubu and @dripdrop are very much in line with what i think about revamping tokenomics.
Agree also that giving away YFI to holders without any work seems bad alignment of incentives, there are several mechanisms that have been mentioned that we need to improve decentralization on and this needs incentives/penalizations. (see below)
Problems that need solving in our mechanisms that we want to explore in v3:
-
Adding a strategy to a vault without the need for msig operation (current governance). If we change this to be governance by voting/timelock, How can we stop this from being gamed? a strategy that has not been reviewed by the internal review security process could have a huge loss or be hacked or even be a rug pull itself? do we allow anyone to propose a strategy or only whitelisted addresses, who decides who is trusted? do we have a delegate voting where security experts have the sign off and are delegated by holders, do we discuss like MAker risk before adding a strategy like they would a backing asset? keep in mind the problem here is that current system allows for fast turnaround of strategies to prod, any system involving voting and discussion may take weeks where a strategy will no longer be profitable or even discussing in public a strategy may make it no longer profitable since its zero sum with competitors (although the future seems to favor long term strategies more than weekend affairs of LM nowadays)
-
Currently there are good incentives to be a strategist but none around risk evaluation of code and security reviews before attaching a strat, this has led to have a bottleneck around needing security experts to vouch for strats, maybe outsourcing this to token holders with a penalization/staking aspect may be an option (insurance also sounds like another option). Incentivize DAOs like yacademy to sign off or stake to review strategies that are secured (risk framework + reputation system?)
-
Vault allocations. There should be an optimal allocation that can be applied to a vault to get the best APY under current on chain conditions, somebody needs to do the analysis off chain and calculate and propose the change on chain. This is currently not well incentivized and needs to be more decentralized if possible, managament of vaults is a tricky business, and current system needs additional on chain data to improve. Some strats cannot go at the front of the queue because they risk liquidation or loss, others can and so on. I think future versions of the vaults need to express these parameters on chain to make decisions, still not sure if a delegate expert team voted by holders is the way to go here, again these decisions are daily, this is not Maker adding a token that can take 3 months discussion, these decisions need to be balanced quickly and in a decentralized way, not an easy problem to solve without risking potential loss of funds or security exploits.
-
Solve the whale problem. The vault needs to track best APR out of optimal allocations to avoid a whale favoring bad allocations, say Compound decides to vote for allocation of strategy funds in their system although AAVE has higher APR%, these changes need to be evaluated onchain and penalized or blocked somehow.
-
Emergency conditions. Timelock and governance mechanisms may be part of some of the above functions but i believe emergency state needs to be without a timelock, worst case scenario we should limit scope of emergency functions but still can act in a good time to salvage a bad situation from becoming worse.
All of the above problems seem more fitting to be rewarded YFI for good management and burning YFI if an actor acts in the wrong way during these operations.
If this operations are decentralized more, it makes sense for protocols upstream to hold YFI and be more involved in governance/management to configure the vaults for optimal yield, yearn becomes the decentralized automated yield function others can participate on.
P.S we are also exploring off chain trusted computations like STARKS where we can simulate and prove the calcs to send to vault state and validate a computation was made off chain with on chain code which is handy for vault allocations, but not sure this can be implemented on next version since its bleeding edge tech, but something that we are looking to discuss/explore in the future as part of decentralizing management of yearn vaults.