November 12, 2021: Nuclear Fallout 🤯☢️
Aftermath of the Emergency DAO’s Judgment
The world watched in awe as the Curve Emergency DAO swiftly took action to resolve Mochi’s feint in the Convex Wars.
This was the first major test of the Curve Emergency DAO. In order to keep Curve decentralized and uncorrupted, the Emergency DAO serves as a 5/9 multisig of which no Curve employee has a key. Among other benefits, this protects Curve from the dreaded wrench attack.
The existence of an Emergency DAO also allows for swift action within a relatively limited scope of protective actions, with most functionality reserved for the general DAO. This particular action (killing the gauge) could still be undone by the DAO if the community felt the Emergency DAO acted improperly in swiftly passing judgment on the situation.
Now that the emergency has been resolved, it’s time for the hand-wringing. Proper functioning? Overreach?
To get our biases out in the open, we feel the Emergency DAO made the correct move. So we’ll spend a bit more time trying to explore the opposing viewpoint, because I’m old enough to remember when honestly considering both sides of a debate was considered a journalistic virtue.
To start with though, we’ll just present several of the views that came down in favor of Curve’s action without much comment.
We therefore first consider the POV of the AZtztk, founder of MOCHI. It’s difficult though, because he spent little time discussing the present incident, mostly defending his role in previous incidents. AZ had little to say in comment on this particular incident, mostly:
Additionally, his argument mostly boiled down to “this isn’t a rug”:
“No rug” may be a tough sell to the $100MM in people who lost everything in the $USDM pool. Although when you join a group, it’s tough to escape the power of the narrative. A lot of people in the $MOCHI Discord are pushing the blame to $CRV, $CVX, Tetra… embracing the frame in which $MOCHI and AZ are the victims. So it goes. It’s not possible to convince somebody who’s mind is already made up.
What I can offer up is a thought experiment. Imagine you find yourself in the position of leading an honest DeFi protocol (legally impossible for a US citizen, so I’m forced to retreat into imagination land), and your protocol suffer a major depeg. What you do? I know exactly how I would act. I would be impossible to silence. I would be lashing out at anybody and everybody in any channel I could find. This would be a full-on war.
In contrast, if you flip the scenario, imagine you’re running a DeFi protocol and you just rugged your users out of millions. How do you react? Maybe suddenly you don’t want too much attention. Maybe you issue a few CYA statements, try to shift the blame, but you pretty much want to get out of the spotlight.
Maybe my impulses are wrong. Maybe it’s impossible to know what happened here. Maybe there’s legal stuff behind the scenes forcing some sort of reply. It’s just a thought experiment. Still worth considering how you would act if you were working in the best interests of your users.
To be honest, a more satisfying statement than a tepid “not a rug” would be: “You bet we rugged, can you blame us? We’d do anything to get that sweet, sweet $CVX. Would you rather a piece of the flywheel or our garbage token?”
At any rate, what are the other arguments against the action? A lot of arguments raise concern about “the precedent.”
When you furniture needs to serve as your conscience, you either need to get yourself to church or get yourself a new couch.
Speaking as somebody who’s not terribly smart, though, I confess I’m having trouble understanding the argument. Curve has architected a system in which external stakeholders can pull the emergency brakes when a 5/9 majority feel the protocol is going in an unwanted direction. Wouldn’t the more dangerous precedent be if most external stakeholders judged that Curve was under a direct threat and refused to take action?
Another great thought experiment:
At least 1 vote of the Emergency DAO is held by Convex, so in this case we’d need 5/8 instead of 5/9 to oust the threat, but it’s nice that there’s a process that accommodates this. Were I a member of the Emergency DAO (that would itself be an emergency!), my thought process might be that Convex was a greater risk before Votium. The successful decentralization of Convex’s voting power has mitigated this risk considerably. If $MOCHI had credible plans to decentralize their $CVX stake, this would be a consideration in my vote.
Other issues raised:
I have a question for people who voted for the $USDM gauge. If you knew $MOCHI would abuse its minting powers to rug the $USDM pool and obtain a large stake of $CVX, would this you want to have changed your vote?
Or if Convex turned evil, should Curve work to excise Convex from its platform? Or would one argue Convex’s role in the platform should stay because a vote previously occurred based on different information. Circumstances change, and a good DeFi platform will be resilient enough to adapt to such changes.
It’s possible that somebody’s interests fell to the tyranny of the majority, but I have to ask: What is the better system for determining which pools are eligible for rewards? Curve has put in place a system where the DAO has the final say - $MOCHI could easily have their rewards reinstated by a simple vote of the DAO. Is it preferable to change this to a 33% vote? Anybody who asks?
Curve needs to have some process, and the protocol is shaped downstream of these processes. If anybody wanted to articulate a better process, I’m sure the DAO would be receptive. While I hear concern about the process but no specific improvements, it appears to me that the process cannot be improved upon.
Several people talked about the wisdom of accepting or allowing bribery.
It’s worth noting that none of the bribery actually takes place within Curve. The various bribery all happens off-site. Although Curve is surely aware of its existence, it’s not technically under its dominion.
For the life of me, I cannot possibly imagine a means of actually preventing such activity. Possibly Curve could try to aggressively police bribery, and propose votes to disincentivize protocols discovered engaging in bribery.
But what would this really accomplish? Undoubtedly, bribery would still happen, but it would inevitably be pushed into the shadows. I personally prefer the system where the bribery takes place in the light of day, entirely on-chain and redounding to the benefit of regular users. The alternative is smoke-filled DMs in a throwback to the legacy meatspace governance mechanisms of yore.
Pretty good… here’s my best impression of the SEC:
No. 11 Texas A&M will head to Oxford to take on No. 15 Ole Miss in a critical SEC West matchup that could provide plenty of fireworks, and No. 17 Auburn will look to rebound against a Mississippi State team that fell to Arkansas last week.
nyuk nyuk
I’d mentioned my bias is in favor of the Emergency DAO. In this case, though, I don’t actually see any argument in favor of “protecting” any particular group — after all, everybody in $USDM got utterly hosed. Where I concurred with this being an emergency was the potential for a cascading effect that could threaten Curve’s entire existence.
It sounds a bit hyperbolic, but play it out a few weeks. $MOCHI is now armed with $46MM extra worth of Convex. The $USDM pool already had the greatest share of rewards by a significant margin, attracting ever greater liquidity. This necessarily creates a destructive feedback loop that can result in $MOCHI cornering the market.
Even if the existential risk to Curve is… 2%, what’s the harm in calling a time-out to reassess things?
It all does raise one issue though. My biggest concern in all of this is that there is a plausible attack vector against Curve. It involves over-leveraging and wrecking other protocols, but it’s still an attack vector nonetheless. It’s somewhat summarized in this thread.
Why did $USDM ever get a rewards gauge? Because there’s little incentive to vote against a rewards gauge, ever. It’s very high up in the process that this creeps in. Once the rewards gauge is in, the attack vector is tougher to shut off without the emergency DAO. The $USDM vote happened without a lot of people paying much attention. If somebody wanted to fix the process, considering process improvements around these votes would be the place to look.
Or maybe this can be solved with increasing recursion.
Read Disclaimers! Author is a CRV/CVX maxi, and thankfully has no funds in $USDM or $MOCHI.
If you like this, please subscribe to support PAC DAO Crypto Activism.