Special thanks to Karl Floersch, Dan Robinson and Tina Zhen for feedback and review. See also Notes on Blockchain Governance, Governance, Part 2: Plutocracy Is Still Bad, On Collusion and Coordination, Good and Bad for earlier thinking on similar topics.
One of the important trends in the blockchain space over the past year is the transition from focusing on decentralized finance (DeFi) to also thinking about decentralized governance (DeGov). While the 2020 is often widely, and with much justification, hailed as a year of DeFi, over the year since then the growing complexity and capability of DeFi projects that make up this trend has led to growing interest in decentralized governance to handle that complexity. There are examples inside of Ethereum: YFI, Compound, Synthetix, UNI, Gitcoin and others have all launched, or even started with, some kind of DAO. But it's also true outside of Ethereum, with arguments over infrastructure funding proposals in Bitcoin Cash, infrastructure funding votes in Zcash, and much more.
The rising popularity of formalized decentralized governance of some form is undeniable, and there are important reasons why people are interested in it. But it is also important to keep in mind the risks of such schemes, as the recent hostile takeover of Steem and subsequent mass exodus to Hive makes clear. I would further argue that these trends are unavoidable. Decentralized governance in some contexts is both necessary and dangerous, for reasons that I will get into in this post. How can we get the benefits of DeGov while minimizing the risks? I will argue for one key part of the answer: we need to move beyond coin voting as it exists in its present form.
Ever since the Declaration of Independence of Cyberspace in 1996, there has been a key unresolved contradiction in what can be called cypherpunk ideology. On the one hand, cypherpunk values are all about using cryptography to minimize coercion, and maximize the efficiency and reach of the main non-coercive coordination mechanism available at the time: private property and markets. On the other hand, the economic logic of private property and markets is optimized for activities that can be "decomposed" into repeated one-to-one interactions, and the infosphere, where art, documentation, science and code are produced and consumed through irreducibly one-to-many interactions, is the exact opposite of that.
There are two key problems inherent to such an environment that need to be solved:
Early blockchain projects largely ignored both of these challenges, pretending that the only public good that mattered was network security, which could be achieved with a single algorithm set in stone forever and paid for with fixed proof of work rewards. This state of affairs in funding was possible at first because of extreme Bitcoin price rises from 2010-13, then the one-time ICO boom from 2014-17, and again from the simultaneous second crypto bubble of 2014-17, all of which made the ecosystem wealthy enough to temporarily paper over the large market inefficiencies. Long-term governance of public resources was similarly ignored: Bitcoin took the path of extreme minimization, focusing on providing a fixed-supply currency and ensuring support for layer-2 payment systems like Lightning and nothing else, Ethereum continued developing mostly harmoniously (with one major exception) because of the strong legitimacy of its pre-existing roadmap (basically: "proof of stake and sharding"), and sophisticated application-layer projects that required anything more did not yet exist.
But now, increasingly, that luck is running out, and challenges of coordinating protocol maintenance and upgrades and funding documentation, research and development while avoiding the risks of centralization are at the forefront.
It is worth stepping back and seeing the absurdity of the present situation. Daily mining issuance rewards from Ethereum are about 13500 ETH, or about $40m, per day. Transaction fees are similarly high; the non-EIP-1559-burned portion continues to be around 1,500 ETH (~$4.5m) per day. So there are many billions of dollars per year going to fund network security. Now, what is the budget of the Ethereum Foundation? About $30-60 million per year. There are non-EF actors (eg. Consensys) contributing to development, but they are not much larger. The situation in Bitcoin is similar, with perhaps even less funding going into non-security public goods.
Here is the situation in a chart:
Within the Ethereum ecosystem, one can make a case that this disparity does not matter too much; tens of millions of dollars per year is "enough" to do the needed R&D and adding more funds does not necessarily improve things, and so the risks to the platform's credible neutrality from instituting in-protocol developer funding exceed the benefits. But in many smaller ecosystems, both ecosystems within Ethereum and entirely separate blockchains like BCH and Zcash, the same debate is brewing, and at those smaller scales the imbalance makes a big difference.
Enter DAOs. A project that launches as a "pure" DAO from day 1 can achieve a combination of two properties that were previously impossible to combine: (i) sufficiency of developer funding, and (ii) credible neutrality of funding (the much-coveted "fair launch"). Instead of developer funding coming from a hardcoded list of receiving addresses, the decisions can be made by the DAO itself.
Of course, it's difficult to make a launch perfectly fair, and unfairness from information asymmetry can often be worse than unfairness from explicit premines (was Bitcoin really a fair launch considering how few people had a chance to even hear about it by the time 1/4 of the supply had already been handed out by the end of 2010?). But even still, in-protocol compensation for non-security public goods from day one seems like a potentially significant step forward toward getting sufficient and more credibly neutral developer funding.
In addition to public goods funding, the other equally important problem requiring governance is protocol maintenance and upgrades. While I advocate trying to minimize all non-automated parameter adjustment (see the "limited governance" section below) and I am a fan of RAI's "un-governance" strategy, there are times where governance is unavoidable. Price oracle inputs must come from somewhere, and occasionally that somewhere needs to change. Until a protocol "ossifies" into its final form, improvements have to be coordinated somehow. Sometimes, a protocol's community might think that they are ready to ossify, but then the world throws a curveball that requires a complete and controversial restructuring. What happens if the US dollar collapses, and RAI has to scramble to create and maintain their own decentralized CPI index for their stablecoin to remain stable and relevant? Here too, DeGov is necessary, and so avoiding it outright is not a viable solution.