The shape of DAOs to come
with Andrew Dickson & Ankur Shah Delight
In this post we explore DAOs -- Decentralized Autonomous Organizations -- as well as blockchain projects developing tools and infrastructure to support them. Such organizations and projects offer solutions for coordinating groups of humans in a decentralized way.
The promise of DAOs is the promise of trustless coordination. DAOs attempt to capture parts of the governance structures of real-world organizations in blockchain protocols or smart contracts, thereby encoding decision-making processes into software systems rather than traditional legal structures.
The term “DAO” itself has been understood to mean different things by different researchers, leading to significant disagreement around exactly what qualifies as a DAO. Some have argued that the Bitcoin network was the first DAO, while others claim Dash and Bitshares were the first DAOs, owing to those projects’ more explicit focus on incorporating governance processes into their protocols.
Regardless of which DAO was first, what is certain is that interest in DAO technology increased dramatically in May of 2016 when a blockchain project called “The DAO” was launched on the Ethereum blockchain and raised more than 150 million USD in Ether over a period of several weeks -- the largest crowdsale ever at the time. The goal of The DAO was to act as something like a decentralized venture capital fund where third-party contractors could propose projects, which would then be funded (or not) based on the results of token votes by DAO token holders. Any profits from such investments would later be distributed to token holders in proportion to their share of The DAO.
What could have been a fascinating experiment in decentralized funding and coordination was cut short abruptly on June 17, 2016 when a hacker exploited a security vulnerability in one of The DAO’s smart contracts, draining $70 million dollars worth of Ether from the contract in just a few hours. A heated debate immediately occurred in the Ethereum community around how to respond, eventually resulting in a hard fork in Ethereum, with the goal of returning the funds to the project’s creators.
Despite the fact that funds were restored, the damage was done. By the end of 2016, The DAO was delisted by several major exchanges, including Kraken and Poloniex. But the final blow came on July 2017 when the SEC released a report clearly stating that tokens issued by The DAO were securities under US law.
Since the fall of The DAO in 2017, research on DAOs has included a much greater focus on the security of the underlying technologies, as well as compliance with international securities laws -- particularly US securities laws.
While DAOs can take as many forms as the real-world structures that they aim to mirror, at a conceptual level, there are a number of components that are common to most DAOs.
The most important part of any DAO is the collection of assets that the DAO coordinates. These assets typically reflect its raison d'etre. For instance, in a DAO related to making investments or dispersing rewards, the coordinated assets may be accounts containing cryptocurrency. Alternatively, in a DAO governing a public blockchain project, like Maker DAO, the assets may be “knobs”, or variable values that can be controlled in the project’s smart contracts that further influence an asset. The coordinated assets represent what’s at stake in the DAO and also answers the question of why we should care about its existence.
The second critical piece of any DAO is its membership model. The fact that every DAO involves a number of actors implies that one or more individuals or entities can somehow become members of it. The membership model dictates the rules for how members are admitted.
In many DAOs, including those governing large public blockchains like Tezos or Dash, becoming a member is as simple as buying tokens. In such models, ownership of tokens grants voting rights akin to the shareholder rights granted to holders of stock in public companies. However, other designs (such as the design employed by Moloch DAO, discussed later in this article) look more like flexible multi-sigs where the only way to become a member is to be nominated by an existing member and approved by a majority of existing members in a vote.
Ultimately, the DAO’s membership model governs who its members will be, and the processes by which members will be allowed to join the organization, as well the processes by which they may leave it.
Similar to the membership model, a DAO’s coordination logic determines which members will be allowed to control which assets and under which conditions. While token voting is commonly used to handle coordination logic in DAOs that govern public blockchains, many DAOs also specify key roles that come with certain explicit privileges, similar to the powers available to officers in real-world corporations.
The types of coordination logic available in DAOs are as varied and unlimited as the models of authority and control in real world organizations. In the future, more artifacts like identity and reputation may even have a role to play in determining the assets that members can control, and the logic of how that control is exercised. We can see early signs of the above, in the grant-giving process of Dash where previous reputation in accomplishing work has become a significant factor for receiving a grant.
Moloch is the biblical name of a Canaanite god associated with child sacrifice, through fire or war.
The most interesting recent experiment in DAO technology, Moloch DAO, was launched in March of 2019 by Ameen Soleimani of SpankChain, along with several other contributors. Moloch runs on the Ethereum blockchain and offers an alternative mechanism to more traditional organizations like the Ethereum Foundation for managing grants related to the development of the Ethereum ecosystem.
The beauty of Moloch lies in its simplicity — the entire thing is just 2 smart contracts comprised of 400 lines of code. Ultimately, Moloch supports just two functions:
New Member Proposal — Existing members can propose that a new member be granted shares in the DAO. Those shares may be granted in exchange for a tribute (a quantity of ETH) in the case of new members who are donors. Or, the shares may be granted in exchange for work done in the real world, in the case of new members who are receiving a grant. Proposals are voted on by existing members over a period of seven days, and require a quorum to pass.
Exit (aka “Rage Quit”) — After any new proposal has been passed, there is a grace period during which any member of Moloch may choose to leave the DAO, giving up their shares and receiving ETH in proportion to the shares that they held, before the new proposal goes into effect. This mechanism, colloquially known as a “rage quit”, gives members a chance to walk away with their stake if they don’t agree with how Moloch is being run, an option that was not available in earlier DAO designs. Members may also simply quit the DAO at any other time (presumably without any rage), cashing out their shares in exchange for their proportionate share of ETH..
Today, the 72 members of Moloch collectively manage about 6000 ETH, worth approximately $1.1M USD in November 2019. A list of Moloch’s completed and in-process proposals can be seen on this page. The grants are the proposals that have tribue values of zero ETH, whereas the remaining proposals are proposals to admit new fund-contributing members to the DAO. An example of one of Moloch’s larger grants to date is funding of approximately $40k for a security audit of Ethereum 2.0 by a Sydney Australia based consultancy, Sigma Prime.
Because of its simplicity and actual adoption, with real money at stake, Moloch represents one of the more interesting recent developments in DAO technology. Despite its relatively short track record, Moloch’s design and simplicity has already led to several forks of the DAO.
The DAO hack in June of 2016 shocked the community badly enough that it took nearly a year -- an eternity in those days -- before a new project stepped in to continue the work. That project was Aragon, which raised $25 million in under 15 minutes in July of 2017. It was the second largest blockchain crowdfunding campaign at the time, bested only by “The Dao”.
As a project, Aragon is often misunderstood by those working in blockchain due to the fact that there are in fact a number of different technologies and corporate entities that all share the name “Aragon”. When most people think of Aragon, they probably think of Aragon OS, a collection of “nuts and bolts” smart contract tools, allowing third party projects to create and operate new DAOs on the Ethereum blockchain.
And while Aragon OS does represent some of the most advanced tooling available for operating a DAO today, so far the project has gained somewhat limited adoption. Despite that, we believe that the Aragon DAO, along with the Aragon Association (a real-world non-profit entity that it governs) may, in fact, be the most interesting parts of the project.
The Aragon DAO probably represents the most developed experiment to date in DAOs governing a real-world corporate entity (and its officers) where significant financial value is at stake. Specifically, the Aragon Association -- a Swiss non-profit entity that holds the money raised in the original 2017 crowdfunding campaign -- has committed to invest and commit its assets as directed by the Aragon DAO. This governance of the Aragon Association is executed through a system of on-chain proposals that are decided based on token voting.
While such governance is not fully trustless -- DAO members must depend upon the officers of the corporate entity to implement the DAO’s proposals -- it is still fascinating and so far has allowed the Aragon DAO to coordinate actions in the real world that are far beyond what most DAOs can do purely trustlessly and on-chain today. As an example, consider the recently approved Aragon Governance Proposal (AGP) #106 directing Aragon to start its own blockchain. Such an effort will require the successful coordination of researchers and developers to do work in the real world, very much outside of the trust boundary of the Ethereum blockchain. While concerns about voter participation and adoption remain, Aragon remains a fascinating experiment in DAO technology.
The DAO that governs the Tezos blockchain may well be the most exciting all-around example of a DAO in production today. Membership is open and truly decentralized, determined exclusively through staking by token holders. And unlike many other DAO experiments, The Tezos DAO coordinates an asset with substantial financial value: the Tezos blockchain (valued at $1.1B in December of 2019). Perhaps most importantly, the DAO utilizes an innovative design that enables execution of full-fledged software updates to validator nodes automatically and trustlessly -- a significant improvement over what has been possible with previous governance DAOs.
How does it all work? Well, members of the Tezos DAO are validators (known in Tezos as “Bakers”) who are elected through token-holders staking. In addition to acting as validators in Tezos’ Liquid Proof-of-Stake consensus algorithm, Bakers may also submit and support protocol amendments.
So far, this is fairly similar to other governance DAOs for large public blockchains like Dash and Decred, and even Aragon DAO discussed in the last section. However it’s when we start to look into how Tezos actually executes its protocol upgrades that things get more interesting.
Protocol amendments on Tezos that make it successfully through the 4-phase proposal process (which takes around 100 days) are written onto the ledger as transactions. Each of these amendment transactions contains the actual code that validators need to run to implement the update. Therefore, as soon as an amendment transaction is approved under consensus, all Tezos validators download and run the new code automatically. The difference between Tezos and all earlier governance DAOs is that the results of the Tezos DAO’s votes are immediately and automatically implemented on each validator’s machine. (For additional detail, see this excellent writeup by Nomadic Labs).
To understand why this matters, let’s compare the Tezos DAO with Aragon. In Aragon DAO, any proposal approved by the DAO is effectively a suggestion from the token holders to the Aragon Association. Since the Association controls all project funds, its officers choose how the Association responds to the DAO’s proposals. While the Aragon Association has honored all of the DAO’s proposals so far, there are no guarantees that it will continue to do so. What happens if the Swiss government freezes the Association’s bank account, or brings legal action against the founders? What happens if all of the corporate officers die in a plane crash on the way to Devcon? In scenarios like these, it’s quite possible that Aragon DAO will be left governing nothing at all.
The same cannot be said for protocol amendments in Tezos, and that’s what is most exciting about the project’s governance DAO. Tezos may well have created the first fully-fledged DAO that governs assets of significant value, while also extending the trust boundary around the kinds of compute assets that can be coordinated.
The Shape of DAOs to Come
While in our opinion DAOs hold immense potential to reshape human collaboration, DAO adoption remains limited. Using our learnings from the first part of this essay, in this section we explore five themes that could lead to increased DAO adoption in the near future.
Bureaucracy Minimization is the Killer Use Case
Bureaucracy within an organization or at the interface between an organization and the state is recognized by many as one of the most significant challenges for organizations today. DAOs, by design, provide a programmable layer for organizational functions that has the potential to be much more performant and reliable vs. traditional government or corporate platforms. In addition, the recent investments in governance tooling and research might further accelerate innovation in a space that desperately needs it, before the world catches up with the concept of trust minimization.
One motivating example along those lines comes from the small European nation of Estonia, which launched an e-residency program in 2014 to allow non-citizens to establish “virtual residency” online. Once approved, virtual residents are given smart cards, which they can use to sign documents, create corporations, open bank accounts, and most importantly, pay taxes fully online. Instead of offering a tax-haven, Estonia provides a bureaucracy-haven where virtual residents can manage their business operations online and minimize their in-person interactions with the Estonian state.
If blockchain-based solutions succeed in providing competitive organizational services, but in a trustless environment, then the case for DAOs becomes very appealing.
Digital Nomads Will Be the Early Adopters
Expanding on our previous point, the frictionless and flexible cooperation enabled by DAOs marries well with the rise of digital nomads who choose careers that allow them to shift locations (and cost of living) at will. The lack of geographical ties dovetails perfectly with the decentralized nature of DAOs, where membership requires “presence” only in the form of cryptographically-signed transactions.
Digital nomads are usually self-employed, engaging in client/employer interactions that are often borderless. Today, it would not be uncommon for a software company based in Berlin to hire a developer with Canadian citizenship, based out of Bali. In addition, many digital nomads are eschewing long-term projects in favor of plugging in to a variety of small teams in a highly optimized manner. This penchant for modular participation as well as the desire to digitally and programmatically manage their business interactions make digital nomads a potential captive audience for DAO technology.
As the number of digital nomads and other fully-remote workers continues to rise, and as DAOs find better and better ways to coordinate them, the combination may lead an entirely new type of software firm, with an agility and global reach beyond anything we can imagine today.
The Proliferation of DAO-Directed Legal Entities
In July of 2018, the state of Vermont enacted Act 205, which allows limited liability companies to manage governance, ownership and conduct material operations on a blockchain. Building on this legislation in June of 2019, the blockchain cooperative dOrg launched what they claim is the first DAO-governed LLC. According to dOrg’s public Github repository:
“This research probes a live case-study of a BBLLC-linked DAO. In doing so, we hope to address open research questions regarding the design and viability of DAOs with traditional legal standing.”
While the legal and corporate mechanics of dOrg’s LLC remain somewhat opaque, the DAO’s proposal history and crypto accounts are transparent and publicly available on this page for anyone to see.
What’s most interesting about DAO-governed LLCs is that, in principle at least, they dramatically reduce the requirement for trusting individual humans. For instance, in our earlier discussion of Aragon, we observed that the Aragon DAO still depends on officers at the Aragon Association to faithfully act on its proposals today. The same cannot be said for dOrg’s LLC. If corporate officers refuse to do the bidding of the DAO, the members can vote to remove or replace them and -- at least in places like Vermont -- such votes appear to now have solid legal footing in the real world.
In the meantime, Metacartel Ventures has recently proposed a very promising for-profit DAO that is coupled with a legal entity and has the goal to finance early-stage decentralized applications. If these two experiments succeed as proofs of concept and manage to overcome legal complexities and other administrative hurdles, they may pave the way for far more ambitious legal constructs, dramatically expanding the kinds of real-world entities and assets that a DAO can control.
Coordinating Assets Beyond Money
While most DAOs today coordinate financial assets, we believe that a more viable path towards DAO adoption might come from other types of digital assets.
One of the main innovations that the Tezos community made in governance was the realization that a DAO could be used to push protocol-changing software updates to validators, trustlessly. Now a number of early-stage projects are looking to expand on this idea to allow DAOs to control a diverse set of assets.
The Handshake protocol and other name registries (like ENS) aim to create a decentralized network that would serve as a new root DNS registry for the entire internet. From a trust perspective it works the same way as Tezos. Namely, as long as Handshake is able to convince most users of the internet to accept that their protocol is the true root DNS protocol, then any updates to DNS entries on Handshake will be propagated to millions of machines all over the Internet, by default. Pairing Handhsake’s capabilities with a DAO platform will enable a set of actors to trustlessly control internet real estate (domain names) collectively and trustlessly.
Other projects are even more ambitious, hoping to use these ideas to allow DAOs and decentralized P2P tools to replace many of the centralized SAAS tools that tech startups depend upon today. Radicle (the Berlin-based project responsible for funding this series of blog posts) aims to enable trustless coordination of any type of software project artifacts (e.g. canonical project head, license, repo and issue access control) using smart contracts and multi-sigs, allowing a set of coders to collectively and trustlessly develop open source software that could be further monetized without a legal entity.
While It’s unclear at this stage what kind of assets fit best with the model of DAOs, non-monetary assets might offer a lower risk environment for new DAO experiments, which would significantly accelerate the pace of innovation in new DAO technology.
Reputation Can “Smooth” Trustless Coordination
Finally, we believe that interoperable (between DAO), on-chain reputation schemes will become much more widely used within DAOs, “smoothing” coordination in trustless environments.
Researching DAO-based grant programs, like Dash and Moloch, we observe that in most cases the grants are given prospectively (i.e. before the work has been done) without any contractual guarantees of any kind. While that might appear to create a problem of accountability, in practice it doesn’t, as reputation schemes keep grantees honest.
Reputation schemes also add confidence in decision making within a group of participants that don’t “know” or trust each other, minimizing the need for enforcement mechanisms used by traditional organizations. This insight aligns with our evidence from the Dark Web, where reputation has proven to be remarkably effective at guaranteeing good behavior in marketplaces, even with pseudonymous agents.
As time goes on, we may be surprised to find that DAOs can use reputation systems to coordinate more kinds of work than we would think. Imagine a future where startups organized as DAOs use reputation systems to assemble specialized dev teams in hours rather than months and avoid complex contracts and legal agreements entirely? How much overhead would be saved and what kind of new interactions this setup will enable?