The D in DAOs has been stumping everyone for a long time.

I've been in countless meetings where DAO contributors are throwing around their ideas of what decentralization is in a human-to-human context. You'd be amazed at how differently everyone defines this word that seems to be the underpinning of everything we do. Finding the meaning of decentralization in a DAO context—a non-computer, day-to-day, human-to-human context—has been extremely difficult. In regards to the confusion of what decentralization really means, Chase Chapman wrote on Twitter, "The human layer requires different mental models than the computational layer."

I'd like to begin exploring that human layer.

I reached a moral dilemma while trying to solve a problem of quality control in BanklessDAO, a DAO I've contributed in. I worked with other members to find a solution, and that solution ended up being a smaller group of people in charge of the quality control itself. This essentially cut off a pipeline for new people to take part in one key aspect of the organization. I spun the ideas over and over in my head at night. Is being less "open" and telling certain contributors that they can't participate in something inherently more "centralized"? What does "decentralized" even mean?

I looked for answers in the canon, but found that definitions of decentralization were all about computers, rarely about humans. In an internal DAO proposal to help solve the problem, I wrote, "Decentralized does not mean distributed, structureless, leaderless, flat, fair, equal, or egalitarian. I’m not saying it should or shouldn’t be these things. But, inherently, all the word really means is not centralized. But even then, there are myriad ways something can be not centralized." Some contributors agreed, some didn't. I saw the word "centralized" thrown around a couple times in channels, and I tried to shake it off. And I realized....do any of us (me included) even know what decentralization means when taken out of the context of computer code?

Today, I'll be exploring and sharing ideas around a few areas where DAOs are attempting decentralization: holding and recording votes, making daily decisions, managing and moving money, and the presence of codified or non-codified hierarchy.

How are votes held and recorded?

When asked what a DAO is, most people talk about voting first. For many, the voting aspect of DAOs is what makes DAOs the most "decentralized" compared to legacy organizations. People aren't accustomed to having a say in how their Web2 day job is run, so the voting aspect of DAOs looks like the silver bullet of decentralization.

DAO canon is no different. The most concrete explanation Ethereum.org provides on how DAOs embody decentralization is related to smart contracts and voting. This post says that the use of smart contracts for controlling the treasury "means that DAOs don't need a central authority. Instead the group makes decisions collectively and payments are authorized automatically when votes pass."

The well-placed use of vague terminology and passive voice is something I'm familiar with when researching the DAO space. This vagueness exists for a reason: no one knows the answer to those few key points, and much should be left open to interpretation so we can innovate without bounds. But, I can't help but notice where our varying definitions of decentralization originate when our canon is filled with terms like "the group makes decisions collectively." This statement hints at consensus-based decision-making, something that sounds excellent in theory but becomes endless round-robins of collective bargaining in practice, stopping DAOs in their tracks (more on this below).

But, putting the collective decision-making aside, I'll sum up that the voting and treasury-control layer seems to be the key aspect of decentralization according to the writers of Ethereum's definition of DAOs.

In Ethereum.org's definition, as well as in Vitalik's foundational 2014 piece, "DAOs, DACs, DAs and More," most DAOs today would be more like DOs. Silly me, thinking we were closer to autonomy than to decentralization! It seems that it could be the other way around. Vitalike writes, "The obvious difference between a DO and a DAO, and the one inherent in the language, is the word "autonomous." In a DO the humans are the ones making the decisions, and a DAO is something that, in some fashion, makes decisions for itself." The humans making the decisions sounds about right for most DAOs (or are they DOs?) today.

It's funny that this particular definition of "autonomous," and arguably the first and most influential definition of the word in the sense of DAOs, is the one that keeps me up at night. "Autonomous" and "automated" seem to be almost used interchangeably here. And the part of about the DAO making decisions is what really gets me. Aren't humans always the ones making the decisions? How could a DAO make decisions without humans?  To me, DAOs are humans through and through. So I struggle to reconcile the canon of what decentralization in DAOs says with the reality of what day-to-day decentralization and autonomy feel like.

DAOHaus writer Adam J. Kerpelman uses voting as a base layer for the definition of DAO decentralization, too. He wrote for the HausParty newsletter that, "the ultimate actions of the organization are made autonomous through the use of software, which is why it can end up sounding like a tool and not an organization." If we replace "actions" with "decisions" (a close synonym in this case) then Kerpelman's definition of what makes a DAO a DAO is also based on recording and executing the results of a vote via a smart contract or other unspecified software.

Voting, and the blanket "decision-making" that gets wrapped up with it, is a key aspect of decentralization. It's the most discussed aspect. But it's an aspect that doesn't really hold muster for me. Just holding a vote doesn't mean we're decentralized. It means we're another D-word—a democracy—but not necessarily decentralized. Endless voting leads to voter apathy, a major problem plaguing DAOs today. So, this all leads me to wonder: how are those decisions made, before the vote or entirely separately from any voting mechanism, on the day-to-day?

How are the day-to-day decisions made?

I've read many tweets and articles that include a flavor of this idea: voting is just a way to record decision-making that has already happened elsewhere. I believe we spend too much time talking about the how of voting and recording votes, and almost no time talking about where those decisions are made, which channels they're discussed in, by who they're discussed and made, and how this is done using that elusive term of decentralization. Because in reality, the voting can look decentralized, but the decisions made leading up to the vote and swaying the voters were made in a deeply centralized, hierarchical way. Due to the public, immutable voting layer on top, those centralized and hierarchical decisions are hidden. You'd only know they're centralized if you were truly "on the inside" of the decision.

When faced with a decision that no one feels (or, more often: knows if) they can make unilaterally, DAOs seem to go after consensus first. Make working groups and hash it out in a series of meetings. Admittedly, when I first started exploring the DAO space, I believed that consensus-based decision-making was the answer. We all get a say, so it gets better through our collective brain power, right? I remember my first few meetings where consensus was the goal. I felt like I was back in a college classroom, sharing my ideas and defending them in front of a room of other students. This is real, intellectual debate, I thought. This is the future. But a couple more go-rounds of debating the same topics, followed by reading the work of Frederic Laloux and Aaron Dignan, has spurred me to throw the "all consensus all the time" idea out the window. I've watched ideas get watered down beyond recognition. I've felt the frustration at stalled work and experienced the stomach-sinking helplessness that can come with a fully consensus-based system. I've seen the same ideas get rehashed a few weeks later, just because there was a slightly different group of people in the room who were steering the conversation completely differently from how it went a week before.

But, what's the solution to decentralized decision-making, if we're not using consensus? Most DAOs have more questions than answers right now. I like the idea proposed by Rodney Evans in this episode of the "On the Other Side" podcast. She said, "people who should consent are the ones doing the work, paying for the work, or significantly impacted by the work. If you use those three heuristics to gather the consenters, that might be sufficient in the absence of more clarity." I'm intrigued by this simple rubric for gathering the necessary people to consent when there isn't a more defined structure yet. But where the argument breaks down for me in the context of DAOs is the paying part. If we all have collective ownership in a way that's considered "decentralized," then aren't we all "paying" for the thing? We all own the token, we all vote on proposals that hopefully are used to move the money (more on this below), and we all have "skin in the game" when it comes to money usage in the DAO. And then I'm once again stalled at consensus.

I personally believe that DAO contributors must be empowered to frequently make unilateral decisions within their role so the DAO can make progress within a reasonable time frame. I also believe that contributors can be trusted to make these decisions because they will have gone through a permissioned hiring and onboarding process, similar to MolochDAO's process, in which value alignment is emphasized and self-managing principles are held in utmost importance. (Completely permissionless, unchecked entering and leaving can slow down the important work of DAOs. I will expand on this in a future newsletter.) Those decisions should be constrained within the bounds of value alignment and what is economically feasible for the DAO, but not every decision or initiative should be made using consensus.

As possible solutions, I'm interested in agreement making at Murmur and the advice process pioneered in "Reinventing Organizations." I'm particularly interested in the fact that decentralized decision-making in Teal Organizations is seen in a near-opposite way from many DAOs today. In a Teal Organization, everyone has the authority to make decisions that directly affect their work, point blank. According to the Reinventing Organization Wiki, "Consensus may sound appealing, but it's not always most effective to give everybody veto power. In the advice process, power and responsibility rest with the decision-maker. Ergo, there is no power to block."

Blocked and unblocked power: an interesting idea to ruminate on in regards to decsion-making.

Who manages and moves the money?

In a beautifully written manifesto-style article for Aragon titled "A DAO Defined: The Big Picture," the control of funds is considered a key element of decentralization. The Aragon team posits that we've been living in a historical precedent of centralizing power in the hands of the few, and that "a huge element of this historical precedent is the organization of money, specifically, the control of money protocols by powerful banking and political institutions." In other words, centralized money power creates centralized systems. And the same for the reverse: decentralize the money power to decentralize the system.

When I look at most DAOs, I see multi-signature wallets owned by founders or early contributors that move money around based on votes, not elegantly-built smart contracts. An elegantly-built smart contract is definitely no easy task—requiring all DAOs to have one could severely restrict the number of DAOs that can become DAOs, stunting the growth of the space. But, as a thought experiment, one could argue that those multisigs are nothing more than power concentrated in the hands a few people who are are trusted to move the money based on the results of a vote. I take the thought experiment a step further and wonder how those individuals are selected. Is there a vote, in which majority rules? Are they rotated out after a term limit? Go even further and dig around the DAO's Notion page or other documentation—you might see the same names come up again and again on various multisigs around the DAO. I'm not saying that this is inherently bad—trusted contributors have more power because of their status, and that isn't going to change. And I think centralization is healthy (and quite necessary) in some situations. But if we trend toward a couple multisig signers controlling all the money, we should call it what it is: a centralized system hidden under a cloak of decentralization.

I want to explain further why it's important to look at these multisigs as centralized entities: The indiviudals who sign for these multisigs are particularly important members of the community, because they really do control the keys to the castle. What if one of those contributors rage quits before they can be replaced? What about a less extreme example—not rage quitting, but simply being away-from-keyboard for an extended period of time? And should these individuals' backgrounds be vetted to ensure they're "fit" to control the entire DAO's treasury? Again, more questions than answers right now. A recent scandal regarding the ponzi-scheme-riddled background of an anonymous multisig signer with money-moving powers threw the benefits of these relatively permissionless multisigs into a harsh reality for many. Trustless becomes trusted before you know it.

What would a truly decentralized money-management system look like? Maybe a smart contract that executes on command. Maybe a five of seven multisig with a detailed election and vetting process in which members are subject to term limits. Maybe money is one of the things that should be centralized with a group of trusted founders and core contributors who are most invested in the project's success. Maybe the solution still isn't here yet.

MolochDAO admits that, regardless of its foundational status in the DAO space, it doesn't even know yet. On the Treasury Management page, one can find this statement: "MolochDAO currently lacks an adequate treasury committee or guild tasked with managing the assets accumulated from reciprocal grants to replenish the DAO treasury. Members are encouraged to consider the best way to attend to this need."

I look forward to following along and seeing what the members come up with.

What type of hierarchy exists?

One of the oft-repeated "check-lists" for what makes a DAO a DAO, housed on Ethereum.org, begins with this: a DAO is "usually flat, and fully democratized," while a traditional organization is "usually hierarchical."

On the surface level, this sounds like it checks out for most DAOs. There's no single CEO calling all the shots, and the community votes on many key decisions. But I find the perception of a total lack of hierarchy in DAOs to be false. We can strive to eliminate the hard and fast hierarchies of the traditional corporate world, but other hierarchies are still at work—whether for better or for worse.

Social and soft hierarchies are everywhere in DAOs. Founding members and outspoken, popular individuals direct decision-making through social capital and soft power every day. Newer members look to these individuals as higher on the food chain. I've long felt that, without really clear planning and role delineation, "coordinators" "champions" and "point persons" of groups in DAOs are really just managers under the guise of a more flat-sounding name.  Even individuals holding more governance power can be considered to be part of a hierarchy—they certainly exercise more power over those who hold fewer tokens. The writers of the Reinventing Organizations Wiki discuss the positives of "natural hierarchies" formed based off of skill and merit. But I believe the hierarchies of token holding, tenure, and loudness/frequency of opinion-sharing are the more realistic (natural or unnatural), and sometimes damaging, hierarchies in DAOs and flat organizations.

I also believe that many DAOs, by not codifying clear roles and agreements, are making even more of a hierarchy than they realize. By leaning into chaos, a sort of "accidental" hierarchy blossoms. Social power is a hierarchy in disguise, and possibly a more damaging one because it means the loudest person in the room rises up above quieter, less popular, or minority voices. With a clearer structure of who makes decisions, who's supposed to "weigh-in" on certain things, and who's in charge of what, the negative effects of accidental hierarchies can be avoided.

How do you define decentralization in the context of DAOs?

Decentralization in the context of DAOs is gradually becoming a buzz word that is losing meaning without clear definitions backing it up. While decentralization is a sliding scale, I believe it's a sliding scale that hasn't had the two ends of the scale defined yet.

I'm excited to keep exploring the meaning of decentralization this year. So, how do you define decentralization in a human-to-human context? Reply to this article or send a tweet and tag me, @samanthajmarin, on Twitter. I'm excited to hear your thoughts and continue this conversation!