The Efficient DAO Paradox

Traditional corporations benefit from a strict hierarchy with well-defined responsibilities. The corporate structure is the product of hundreds of years of evolution, and does a great job delegating decision-making power to those that are best qualified—those that have risen to the role through a meritocracy and are given the trust and latitude to execute and take risks as they deem appropriate.

And while the CTO or CMO may be given direction by the board or CEO, they have freedom to take risks and follow their intuition, and as a consequence, are held accountable for the consequences of their decisions. A company that required a board vote on everything would likely be too risk-averse or slow moving to enjoy a competitive advantage. And an enterprise in which all decisions were voted on by shareholders would likely enjoy a swift death.

And then we have crypto…

I, and likely many in crypto, view the decentralized voting of DAOs to be inherently inefficient. Every token holder votes on how the DAO’s money should be spent, on how big the grant program should be, on who should be hired or fired, etc.

Others seem to be wearing rose-tinted glasses and believe that decentralized voting will lead to some sort of beneficial societal or cultural reform—that having everyone vote on every feature is a feature, not a bug. To me, this mindset is often reminiscent of a fondness of socialism — but that’s beside the point.

The reality is that neither of these extremes are choices that we have in crypto.

Gaining and maintaining control of resources is hard. Extremely hard.

Paraphrasing Google results, inherited money is lost within a generation 70% of the time. By three generations, 90% of wealth has been lost (to other, better, resource allocators).

In a free market, capital flows towards those that demonstrate the most effective use of it — those that invest it the most wisely. Maximizing this tendency results in the greatest societal benefit as sound investments pay off with new wealth creation.

In closed-source tech, or traditional corporations, the decision-making payoff spectrum looks something like this:

Having individuals who have total responsibility for how resources are allocated in their corporate function, coupled with total accountability for the consequences of those decisions, works much better than having diffuse decision making power in which everyone gets a say… and therefore, nobody has accountability.

Society has evolved toward centralized resource allocation in closed-source technology companies for good reason. Resource allocators compete in a meritocracy to invest both human talent/time and capital as efficiently as possible, and are rewarded or demoted depending on results.

I believe the spectrum is a bit different with open-source technology—with a community of users eager to help shape or evangelize the product.

How can these projects most effectively decide how to use their funds, i.e. on development, growth, research, grants, or investments?

With token ownership in particular, a purely centralized model fails to surface ideas from the periphery or hive mind, and risks losing touch with its engaged base of user-investors.

So perhaps the spectrum looks a bit more like this, with the most efficient (green) somewhere between the two extremes:

An example of a centralized system with delegates is the Arbitrum DAO on Tally. Here you can see that many of the top delegates are individuals like Eric Wall and firms like Wintermute and Gauntlet that have heavily invested in the Arbitrum ecosystem. They have skin-in-the-game and are in unique positions to provide input.

Ultimately the final execution of ideas is still controlled by a multisig, but the signers are expected to honor their soft commitment to execute according to the will of token holders.

I might argue that veto authority by a centralized entity would be better than a soft commitment to honor vote outcomes, but unfortunately we are forced into ignoring this possibility because of legal risk.

Regulatory risk makes it extremely difficult for a crypto project to maintain meritocratic centralized control over resource allocation. In other words, even though the left two points in the above graphic may be preferable from an efficiency perspective, regulatory pressure makes these infeasible.

Even pseudonymous reputation systems, full stack distributed infrastructure, and so on cannot completely solve this problem as long as there is fear of regulatory retribution or abuse. While it’s possible that centralized control could be possible via a small subset of individuals from relatively crypto-friendly jurisdictions, the space needs a global meritocracy, rather than limiting the pool of talent to certain geographies.

So we are stuck, reluctantly having to build on the right side of this spectrum.

Thus, regulatory pressure forcing us to decentralize decision making comes at the cost of suboptimal investments (i.e. moving from left to right).

And so, the question then becomes: “How might we optimize within the inefficient extreme of this spectrum (the orange dotted box below) as restricted by regulatory pressure?”

In summary, despite governance often being pushed as “centralized bad, decentralized (someday) better”, the traditional corporate structure has proven its efficiency in maximizing output. While it has been temporarily hampered in crypto by an antiquated legal system, it’s important to keep in mind that we are attempting to maximize output in an ‘artificially constrained’ environment.

The models we’re building don’t need to replace corporate structure, a product of social evolution. They just need to be efficient enough for now. I’m bullish on improving the flow of information from the hive mind of users and investors to the top of a centralized hierarchy, and I think it’s in this domain where the crypto industry can realistically lead the way on improving the traditional corporate structure in the long-term.

And within our current constraints, many ideas have been tried, adopted, rejected, and proposed on how to potentially push us further left on this spectrum, from splitting ownership in 3-entity systems (the builder / DAO / foundation model) to ways to filter out noise and promote the delegation of decision making.

A great and recent article I enjoyed was Grant Theft DAO: Grant programs are broken, can we fix them?

Am I spot on or losing the plot? Let me know what you think.