Terra Classic's Looming Mexican Standoff
The Terra Rebels designate August 12 as the date of a head-on collision with Terraform Labs
At Terra Classic block 8,890,141 in Terra blockchain-time, or sometime on August 12th in IRL time, the Terra Rebels and the active validators who support them (45% of the total was the last I heard) will propose the Rebels’ Terra Version 0.5.21 Terra Classic upgrade, or “v21” as the Rebels call it. Their consensus will conflict with the legacy consensus of TFL, which has been running the network without any tech support or other sign of life for over 70 days. Validators who do not explicitly support the Rebels’ upgrade will, by default, propose the old consensus.
At that point, a Reservoir Dogs-style standoff over the governance of the Terra Classic chain will begin.
Breaking the Chain
In the Tendermint consensus process, blocks are typically produced every 6 seconds or so. Every time a block is produced, all validators on the active validator set vote on the proposed code. They must all propose identical code for the block to be added to the chain.
If a validator(s) proposes code that doesn’t match the others’, the network halts, immediately compares all the proposed code blocks, and begins a ruthless process to create a new network consensus.
If two-thirds of all staked L1 tokens (LUNC) propose identical code, that code is recognized by the network as the consensus majority. The minority, with less than 1/3 of the votes, is overridden and sent to validator jail. If the defeated minority continues proposing “erroneous” code after being sent to validator jail, their staked L1 tokens are gradually slashed to zero, they become powerless, and new validators appear to take the terminated validators’ place. The whole process could take an hour or less.
However, if the majority cannot muster a two-thirds majority and the minority is greater than one-third, the chain halts until one side racks up a two-thirds consensus.
So, once the Terra Rebels’ contrarian code is uploaded to the network, the chain will halt, and a very noisy battle between the missionaries (Terra Rebels) and the mercenaries (supporters of the legacy code, who will be widely seen as TFL/Kwon stooges) over the future of Terra Classic will erupt.
The votes
There are currently 315 million staked LUNC out of 6.9 trillion (.004%). Further delegation was disabled by Kwon/TFL’s own decision in the name of preserving network security during an emergency. Kwon / TFL / the Luna Foundation Guard (LFG) control 221 million of those votes (70%), which they bought and staked immediately before halting the network. The LFG’s takeover and halt of the network was itself a “governance attack” on the network.
Since then, the LFG has been dormant despite a tsunami of requests for tech support, governance handoffs, and other basic maintenance decisions. Kwon publicly stated that he has “no control or any special privileges [and] the [original poster] should coordinate with validators for a network upgrade.”
Kwon / LFG are certainly bending over backwards to create the appearance of not controlling the Luna Foundation Guard or the Terra Classic network, presumably for legal reasons. If the LFG were to exercise its votes and interfere with any attempted network upgrade, it would be a lethal nail in Kwon’s legal coffin, and a clear-cut case of lying.
Thus, the Rebels sensibly presume that the TFL-controlled Luna Foundation Guard won’t overtly interfere in any governance change. Furthermore, an LFG veto would prove that Do Kwon is a bandit, destroy the reputations of any TFL employees involved, and spur a wave of public defections by TFL current/former employees to the Rebels and South Korean law enforcement.
Encouragingly, TFL has engaged in communications with the Terra Rebels at multiple levels; TFL employees are leaking them information to varying degrees unofficially, in addition to very measured official communications. TFL has sped up preparations for an airdrop of the LFG Wallet’s $65M of remaining off-chain assets (appx $50M AVAX, $9M BNB, $6M BTC) to USTC holders, seemingly assuming some kind of transfer of control.
Soon after Kwon’s post, Orbital Command, a validator widely seen in Kwon’s back pocket, posted an interesting analysis of claimable network assets which was very consistent with the Rebels’ own internal assessment of claimable network assets. Orbital Command noted that of the 9.8 billion USTC in existence, over 4 billion is controlled by community- or TFL-affiliate entities (i.e., rightful community property for future network recapitalization). The $65M of airdropped AVAX/BNB/BTC would thus be spread across the remaining ((5.8B/9.8B) * $370M USTC marketcap) = $220M of non-TFL-owned USTC market cap.
All of these signals amount to a growing accommodation on TFL’s part of a potential changing of the guard, which the Rebels found encouraging.
But when it comes to concrete actions, at all levels, TFL has put up an iron curtain of radio silence. Nothing happens with any governance proposals. No tech support is offered for any of the 20+ projects that have signed up to join Terra Classic. Kwon has never replied to Vegas, one of the Rebels leaders on public social media, when Vegas has directly addressed him.
Kwon’s ambitions to launch a new algorithmic stablecoin are well-known, and many Terra Rebels see in Kwon’s “lack of control” a self-serving, malign neglect of the community Kwon left behind. Why would Kwon help a protocol if he’s planning on launching or backing a competing product?
After 2 months of being stonewalled, the Terra Rebels have assumed the worst, and are going to go to the mat to see which consensus will prevail. (Which, by the way, was exactly what happened when TFL halted validator delegation in the first place: TFL didn’t get a 2/3 majority for the vote, and broke the chain until they peeled off enough validators to support the decision.)
The Rebels have publicized version 0.5.21’s final code, barring last-minute adjustments, for implementing Terra Classic proposals 3568 & 4095, which implement the 1.2% transaction tax and restore validator delegation rights respectively. (The Rebels do not support the burn-tax proposal, but wrote the code to implement it out of Proof-of-Stake democratic principle, since it passed governance with 90% of the vote.)
Their code comes with several other precautionary provisos:
it restricts delegation to the active validator set only for a period of 60 days to prevent potential governance attacks.
IBC would not be re-enabled for v21.
No validator can be delegated more than 25% of all staked LUNC at any time.
The transaction tax, because of a technical parameter change requirement, would require an additional governance action in order to be implemented.
The code has passed all internal testing and third-party reviews with flying colors, and raised no substantive issues. It’s currently undergoing third-party security audit as well as independent validator code reviews.
Intense lobbying of validators, and current and former TFL employees, continues. The Rebels take comfort in a third “neutral, but increasingly anti-TFL” validator faction, which has basically signaled to TFL that unless they find something seriously wrong with the Rebels’ proposal, it’s time to hand over the car keys:
In the meantime, Terra Rebels and TFL are on course for a head-on collision on August 12th. It should be interesting.