Decentralization Isn’t Everything. It’s the Only Thing
People in crypto often behave like the pretense of decentralization is equivalent to the presence of it. Investors and users readily flock to any project that launches a blockchain and issues a token with little consideration for the features that matter, like the token model or governance. Not only is this foolish, it’s dangerous. Every project will eventually go through a painful adolescence when it will be tested. The tech, incentive model and community will be pushed to the brink. Only the ones that are actually decentralized and credibly neutral will survive.
Not for any ideological reason, but because neutrality is an evolutionary advantage, the ultimate form of anti-fragility. Vectors of centralization are like the exhaust ports of the Death Star. On a long enough timeline, insurgents will find them, exploit them, and blow the damn thing up.
Bitcoin has survived this long because there are no such vectors. There are points of concentration like mining and custody, but the players within them are fluid. The last cycle’s leaders are often the current ones also-rans.
Despite BTC’s size and importance, there is no hard power and little soft power in the ecosystem — as the recent debate over inscriptions showed. Just like during the block-size war of 2017, the only thing the insurgents could do was leave. Tellingly, the projects created by the defectors 6 years ago all blew up. They had too many points of centralization, including their self-anointed leadership. Bitcoin was lucky that Satoshi disappeared.
Ethereum flirted with danger during the crisis following the DAO hack. The controversial decision to hit “undo” revealed that Vitalik and a few others had a lot of soft power. But the community was fortunate to have a leader who didn’t want it and has worked hard to diminish his role. Too much power in the hands of a single person is deadly in crypto, a phenomenon we’ve seen time and again.
Technical design is another dangerous source of centralization. Part of the elegance of Nakamoto Consensus is the way it pits miners, nodes, and coin owners against each other. It’s a system of checks and balances that defaults to nothing changing unless there’s near universal agreement that something must. That rigidity has its downsides — Bitcoin is hard to upgrade — but proves the system is actually decentralized. Proof of Stake (PoS) is inherently more flexible but has certain pitfalls, such as:
- Tokens have to be pre-mined so the protocol can have stakers at launch
- Investors and validators are no longer separate entities
- Active governance can be embedded into consensus
- Liquid staking and restaking are likely to be popular
- Defection is difficult
Unlike Proof of Work, PoS systems are more versatile because they offer both carrots and sticks. This might enhance the cryptoeconomic security of a blockchain, but it leads to more vectors of centralization, more exhaust vents. Validators can now be slashed, but they can also be bribed or otherwise corrupted. Miners are always harder to bribe because there is no way to know who will find the next block.
Ethereum was able to avoid some of these pitfalls by launching as a PoW chain. Despite its ICO and pre-mine, half the coins currently in circulation are from mining. Not only did this approach lead to broader token distribution, it diffused power.
If Ethereum had launched with the same token model of virtually every other L1 then the debate about bailing out The DAO wouldn’t have been a debate at all, a few entities would have decided the outcome. The lack of mining would have meant no botched hard fork and no Ethereum Classic, either. This would have set a terrible precedent and made other bailouts — like the one requested for the Parity bug — likely. PoS is more malleable than PoW, for better and for worse.
Today, we hope that Ethereum is decentralized enough, and power within the ecosystem is diffuse enough. But the forces of centralization never rest, and there are new threats to worry about, like liquid staking and restaking. Culture has a heavy bearing on how these risks will play out. The teams behind Lido and EigenLayer seem to take decentralization seriously, perhaps because Ethereum culture wouldn’t tolerate otherwise.
Most other chains don’t have such a culture and don’t really want one. Their founders gloss over important issues like node diversity, client diversity, and token distribution. They raise too much money without realizing that cash in the bank is also an exhaust vent. They give too many tokens to investors, then let those investors stake locked tokens to avoid dilution — as if they want to prevent diffuse ownership
Most new projects also launch with a trio of entities who pretend to be independent. There’s a DAO that’s supposed to be accountable to token holders, but thanks to the pre-mine, these are mostly insiders and VCs. There’s a foundation that is supposed to manage growth but had its leaders picked by the Labs. The Lab gets its own token allocation but also raises money against its equity, raising thorny questions about value accrual and governance. Anyone who invests in the Lab’s equity gets a say over its tokens.
None of these projects pay enough attention to the nuts and bolts of what it means to be decentralized. Then again, if developers and users don’t care, why should the founders? So long as numbers go up, everyone is happy.
But all speculative things eventually come to an end, and crypto isn’t good at being just a casino. Speculation can help bootstrap adoption in the short term but can’t be the main attraction forever. Once the hype fades and the governance devolves into chaos and the exhaust ports get discovered the whole thing goes to hell. Then people will wonder: was it just a get rich quick scheme all along?
There’s a lot to be learned about the how of launching a successful project, but over a decade later, we have a long list of what not to do.
Decentralization isn’t everything, it’s the only thing.
