What Types Of Organizations Should Be Autonomous?
How to Think About Autonomy
This piece shares some recent thinking about blockchain-mediated organizations. It is an extension of ongoing conversations with Steven McKie reflected in prior pieces on DAOs, blockchain-specific governance, and the emergent properties of blockchain governance as we further explore the best means for coordinating agents globally at scale.
People who are interested in blockchains as organizers of human activity often invoke the concept of the DAO, or Decentralized Autonomous Organization. But I increasingly wonder whether this concept — which emerged quite early in our attempts to understand and articulate the possibilities of blockchains — is confused.
It is time to bring it in for a round of scrutiny, to ensure that it is not limiting or biasing our thinking.
After all, it is quite odd to think of an organization as “autonomous.” We can picture autonomous machines, like wind-up toys. But what would an autonomous organization be? Would its autonomy imply independence from the will of its human constituents? And if so, why would anyone join it?
In fact, all organizations resemble wind-up toys in that they are comprised of component parts — except unlike wind up toys, organizations are made of people. And unlike springs and gears, people are, speaking colloquially, “autonomous”. They don’t behave according to simple, predictable physical rules. It is therefore odd that to suggest that something made out of people, like an organization, could be autonomous.
In other words, it would seem that either DAOs subvert the will of humans, or the will of humans subverts the autonomy of DAOs.
People familiar with the space might find this line of inquiry rather absurd. For example, the fact that a DAO has posted an incentive for me to do something does not compromise my autonomy. And conversely, the fact that I might cast an unpredictable vote in some DAO’s decision process does not compromise the hidebound-ness of the code that defined the DAO.
But all of this casts light on what DAOs really are. They are not human-will-subverting Molochs, but rather static systems that people use to modify complex incentives for one another. Their “autonomy” simply denotes the fact that another crowd of people — the miners maintaining the blockchain — continually predictably executes the code that mediates these incentive-modification processes. The miners are like the spring in the wind-up toy.
Yet, if DAOs are really just incentive-modifying systems, perhaps we should think and talk about them differently. There are two reasons for this. First, the current framing is strange and off-putting. (I.e., it tends to suggest that a person must check her own autonomy at the door if she wants to participate.) Second, and more interestingly, the framing biases and distorts the thinking of DAO creators themselves. Namely, the term DAO seems to suggest that DAOs should have their own institutional agendas, rather than neutrally serving the agendas of participants as collaboration tools. It is this second point that I would like to emphasize here.
Why “DAO”? Focus On the Function, Not the Ideals
It turns out that there are really two superfluous assumptions built into the notion of a DAO. One is that it’s autonomous, and another is that it’s an organization.
As I alluded above, the first assumption can be punctured by picturing a DAO-like entity that makes its decisions by democratic vote. Such a thing is not really any more autonomous than a democratic polity whose constitution is hard to change. And democratic polities, when they work, are much better understood as being responsive rather than autonomous.
This assumption is very much worth puncturing. After all, don’t we prefer things that are more reflexive and respond to human needs to things that don’t (or things that, like wind-up toys, serve only the will of whomever set them in motion)? While it is possible to build DAO-like entities that resemble wind-up toys, it is much more attractive to build responsive ones that can augment accordingly based upon the current needs or the organization.
The second assumption, that DAO-like entities are organizations, can be punctured by simply noticing that they need not behave like organizations at all. Organizations usually act to preserve themselves — they don’t want to go out of business. Furthermore, they almost always serve their owners or members preferentially to all other humans. DAO-like entities in fact need not have these characteristics. They can be one-off collaborations that die harmlessly when they have served their purpose. And even more interestingly, they can strive to be public good providers, instead of private good providers.
This last point is absolutely key, in my mind. It is my main motivation for encouraging everyone to rethink the use of the term DAO. Because when I picture a DAO, I don’t picture the best version of what these entities can become. I would like to see persistent decentralized tools that, for example, help communities fund public goods through mechanisms like LR/QF. But these shouldn’t be designed with the concept of “autonomy” in mind. They should be designed to be responsive to peoples’ values, not to impose new institutional values. Furthermore, we need not think of them as “entities” with an inside and an outside — with members and nonmembers — as the very idea of an organization implies.
This may seem like semantics. But words are important — they can restrict, expand, and guide our thinking. If we use more accurate and helpful terminology, we just might build more accurate and helpful technology. I therefore invite you to join me in a slow migration away from the term “DAO”, and try to more specifically hone in on what your smart-contract lead organizations and services are attempting to coordinate. This ensures we focus on the problem-solving nature of these tools, and don’t get others lost in our oftentimes confusing jargon, blended with our utopian ideals.
Edited by: Steven McKie
Subscribe to our Newsletter to get this in your inbox:
— — — —
 It is worth remembering that the word autonomy has two parts: auto, meaning self, and nomos, meaning law. An autonomous thing operates under its own rules.
I have in mind, of course, the acronym DAO, for Decentralized Autonomous Organization. Recent reflection has led me to the conclusion that autonomy is the wrong framework through which to understand these kind of organizations. I am far from the first person to realize that the concept of “autonomy” is used somewhat loosely in this industry defined, and want to ensure we focus on and promote the functions that matter and push us forward as a community.