I stand in solidarity with the UAW members striking for better living conditions. I planned to give this talk in AnnaLee Saxenian's Politics of Information class, but I decided to record this talk as a podcast so as not to cross the picket line. I hope this talk with speak to the concerns and interests of UC Berkeley student workers and working people everywhere.
Welcome. Thank you for tuning in, and thank you to Anno for having me back to speak again. I gave a talk in this class last year, and this talk is, in some ways, a response to that one. That talk was called This Internet, on the Ground, which I've linked back to in the transcript. That talk discussed how I thought, and think, the internet should be—in a phrase, more locally controlled, more bindingly tied to the will of the people who use it. This conviction has not let go of me yet. Right after giving that talk, I realized I needed some tools to make that vision happen. So I joined DAO DAO, the subject of this talk.
First: who am I? My name is Nick Merrill. I am a research fellow at the UC Berkeley Center for Long-Term Cybersecurity and a worker-owner of DAO DAO and Juno, both of which I'll tell you more about in this talk.
So let's get to it.
What is DAO DAO?
In short, DAO DAO is a cooperative that builds tools to help people run cooperatives.
What's a cooperative? Cooperatives can take many forms, so definitions can get heady, but a simple example is a worker-owned coffee shop. A regular coffee shop has an owner or owners who hold all of the voting rights over the entity—say an LLC—and take all of the profits—that is, the surplus value workers produce. In a cooperative model, ownership over that entity would be split among the people who work there. You can imagine more or less radical formulations of this model. Workers could own everything and vote to decide how to distribute profits. Workers could gain ownership slowly over their tenure. Workers could have ownership taken away if they stop working. And so on. These are all the sorts of design decisions you think about when you build a cooperative.
But worker-owned businesses aren't the only type of cooperative. Cooperative grocery stores might give ownership shares to customers, allowing customers to vote on what the grocer stocks, what prices are, what to pay to workers. Balancing the interests of competing groups, like workers and customers, is an important part of designing and managing a cooperative. Also, not all customers want to vote on every issue. California voters should feel me on this, with the nonstop ballot props concerning issues I have no expertise on nor ability to judge. The ability to delegate one's vote—to let someone vote on your behalf, and also to revoke that privilege on a vote-by-vote basis—is a critical feature, especially in larger co-ops. Enabling these features—letting people make organizations that have features like these ones—that's what DAO DAO does. We make it easier to build organizations that have the sorts of features cooperatives need.
I'm not going to bore you with market research in this talk, but suffice it to say that starting a co-op today is hard and expensive. If you want to start a proper cooperative business, you can expect to pay several thousand dollars in legal fees, often hundreds of thousands of dollars, depending on how complex your cooperative is and particularly whether it functions across national borders, a subject we'll talk about soon. Also, even if you're starting a small, simple co-op and you're willing to fork out a couple of grand to do it, designing a structure for that co-op is error-prone. Example from my life: a Berkeley-based coffee co-op recently closed its doors. I have a friend that was a worker-owner there. What happened? They had a voting threshold of 100%; everyone needed to agree on every decision. For a small co-op, that works great. Consensus can be powerful. But what happens when you grow? What happens when you have to bring someone new on, and they turn out to be a bit Machiavellian, and manipulate that consensus rule to gridlock decision-making? It sounds obvious in retrospect, but by the time co-op members realize their mistake, it can be too late.
Similarly, say you sell shares to your community to raise money, as a local Oakland grocery co-op did a few years back. As a co-op, you want to know how many outstanding shares you have and what voting percentage they represent. If you have too many outstanding shares, someone can come around, knock on doors and buy up shares, and end up with a controlling stake in your co-op. As someone who runs a co-op, you typically want to know if this is possible, what the so-called "economic cost to control" might be for your business. Our tools help you keep track of things like that. And those are just two examples of factors co-ops need to keep track of. What's voting quorum? If you set it too high, and too many shares are outstanding, you may be unable to vote on anything, which means people don't get paid and nothing gets done. Again, we use design to make tools that make it easier to design cooperatives that work reliably.
In fact—and this is where this talk starts to converge on notions of the internet—I believe that the tools we're building at DAO DAO can do more than run traditional cooperatives. I believe DAO DAO can help run cooperative platforms; alternatives to tech companies. Imagine a version of Uber that's owned by the drivers and passengers. Imagine a version of Google that's owned by the engineers and users. Imagine data unions or data cooperatives who work together to negotiate on a price for their data. We expect DAO DAO to help make all of these cooperative forms possible in a way they aren't today.
But, we start small. We, DAO DAO, are a worker-owned software company. We build tools to manage a worker-owned company like ours. We run our business with those very tools. That's right, we use the tools we build to manage ourselves. DAO DAO runs on DAO DAO.
Now—you're probably asking yourself: how do we make money? After all, we're an open source project; all of our software is freely available under a permissive license. Short answer: share swaps. I'll note briefly we haven't done this yet, but we're just about to, so the mechanism we're about to test is what I'll tell you about here.
Here's the principle. We find a cooperative, ideally one that already has some business, some good activity. We talk to them and see if they're interested in using our tools. Typically, they are, but they need some help getting set up. So, in exchange for helping them get set up, we swap our shares for their shares. Their shares, hopefully, are worth something, and we can sell them to pay our contributors. Our shares give them control over DAO DAO. They earn a say in the tools they use to run their business.
How does DAO DAO work?
So, we're a co-op that builds co-op tooling, and the co-op is run by the co-ops that use it. That's the eventual dream for DAO DAO.
How do we do it? How does DAO DAO work? Before I answer that question… who has heard of blockchains? Who's ready to slam their laptop shut right now at the mention of the word? Who hates, if not blockchains, what they represent, the people who are really into them?
To the people who felt something at the mention of that word: I feel you. I'm going to tell you why we decided to use blockchains for DAO DAO. And, before I do, let me reveal to you: I am a deeply ideological person, but not about blockchains. I believe people should have control over their everyday lives, and I view cooperative ownership—ownership over the place you work, over the businesses you patronize, over the schools you learn at—as an important component of returning that control to people—specifically, to more people than currently enjoy that control, to the people who live at the mercy of institutions and organizations, of all scales, in which they have no say. I could give a fuck about a database technology. I do what works and that's it.
So, let me start with how I want DAO DAO to work.
Here are the properties I care about:
Co-ops shouldn't need to run their own software. They shouldn't need to have anyone on staff with technical expertise.
Co-op members shouldn't be able to "steal from the till." No member should be able to cheat the other members, manipulate accounts, or manipulate the rules by which accounts update.
Moving money should be easy. Co-ops should be able to send and receive money from and to anyone.
Starting a co-op should be as close to free as your jurisdiction allows.
No third party should stand between you and your cooperative. No one should be able to "turn your co-op off," for example.
The rules should be verifiable to all participants.
Say we don't use a blockchain. We use a regular, centralized database, like Postgres, and we build a regular web2 app. If we do that, we have to either host it ourselves or make you host it.
Let's say we host it ourselves. In that world, you don't need to run your own software—great. Also, we can protect you from stealing from the till—as long as you trust us. Moving money: yeah, we can work with payment processors and make that easy, though it would cost us money, and we'd need to charge you. In fact, hosting all this stuff will cost us something, and that means we have to get money from somewhere. That creates a problem for our low-cost imperative. We become a SaaS company, effectively we extract rent, or worse, we become incented to get shares of cooperatives, hoping they'll go up in value later, which makes us a cooperative venture capitalist—not a prospect that jolts me out of bed in the morning with excitement. More importantly, the SaaS version of DAO DAO puts a party (us) between you and your business. We could turn your co-op off, for example, if a court in your country made us. Finally, our hosting software makes it impossible for you to verify that your co-op is working the way you intended.
Now, let's say you host it. This is good in the sense that there's no third party between you and your business… for some value of "you." Someone needs to run this thing, which adds technical complexity and is intrinsically exclusionary of the non-tech-literate. It also incurs costs for the co-op, as they now need to pay for hosting. Their payment processor will probably charge them too, which huts our low-cost imperative. It also allows those co-op members who are tech literate to "steal from the till," The people who run that software have an ability to steal that the other co-op members do not have. And finally, although the tech-literate members of the co-op may be able to verify the way the rules work, those members who don't have access to the deployment. Everyone else cannot verify, no matter how tech-literate they are, that the co-op will adhere to the rules as stated.
That brings us to running a co-op on a blockchain. First, I owe you an explanation: what is a blockchain? Blockchains are public databases that make the most pessimistic possible assumptions about the parties that provision them. Specifically, blockchains provide the capacity to run multiple versions of a shared ledger (effectively, a state machine) across multiple parties, even when some of those parties have an incentive to manipulate the ledger or the rules by which the ledger is updated.
For co-ops that live on a blockchain, we can make it so that co-op members don't need to run their own software. They can interact with a decentralized application from their web browser. Starting a co-op here could be free, and for now is close to free thanks to low network fees on Juno, which we'll discuss momentarily. No third-party stands in your way. And the members can verify the way the rules work by examining smart contracts. Of course, you need some technical expertise to understand those smart contracts, so I conservatively award that property with a yellow on those technical-skills-required exclusionary grounds.
On the money aspect: the blockchain is great at moving around cryptocurrencies. As for taking fiat, that story is a little bit more complicated and a bit outside the scope of this talk, but suffice it to say that the barrier between national, pay-for-your-groceries-with-it currencies on-chain is in constant flux but headed toward increased porousness. The wild west of centralized exchanges is in retreat now, especially after FTX, and better mechanisms for "onboarding from fiat," as they say in the industry, are improving. Also, central banks are eyeing using blockchains for their own national currencies, and I think they'd be wise to do so to meet their own purposes, like control and regulation, which we'll talk about later. For now, I give it a yellow—co-ops deal with money well as long as that money is legible to the chain; and if they don't want to mess with cryptocurrency, our tool still works to manage voting and all the rest.
So, of the options available—us hosting it for you, you hosting it yourself, running the tools on a blockchain—we believe blockchain is simply the best option for us given the properties we want for our co-ops. That's why we build co-ops in the form of Decentralized Autonomous Organizations, or DAOs. We're a DAO that builds DAOs. That's why we're called DAO DAO. (We're also fans of the Tao Te Ching; our logo is a nod to that. I recommend Ursula K Le Guinn's translation).
What are DAOs? DAOs are cooperatively run organizations democratically controlled by their members. Coordinated on a blockchain via smart contracts (on-chain code), the rules of the organization, and the rules by which those rules can change, are cryptographically assured. Optionally, DAOs can use governance tokens to create transferable voting rights, similar to shares in a company. Smart contracts provide automatic accounting; the blockchain prevents members from manipulating those accounts (no stealing from the till); their digital form allows flexible, cross-border coordination (they do not necessarily require a legal team to scale internationally); and, for regulators or tax authorities, the blockchain gives them a publicly inspectable state (they do not require an accounting team to audit).
Juno
So that's how DAO DAO works. By now, I've told you what DAO DAO is, what problems it aims to solve, and roughly how we decided to solve those problems. I've also told you how we expect to make money—what I described as share swaps earlier are, in fact, token swaps: we swap our DAO's governance tokens for theirs.
Our first token swap is set to be with the very platform we're built on top of: a cooperatively owned computing platform called Juno.
Juno is a DAO that provides computing resources for developers who want to build applications. You can think of Juno as a worker-owned alternative to Amazon’s AWS. Where AWS charges you for their server time (think: running a website or a database), Juno provides analogous services. Using Juno as the backend for your application can be considerably cheaper than hosting an application with equivalent functionality on AWS. Our insistence on Rust as a programming language tends to produce safer and more reliable code.
As a bonus, because we run a chain, developers can provide assurances to users about the code they run. Consider: What code runs on Uber’s backend when you call a car? Does it discriminate based on pick-up location or destination? They can’t prove to you that it doesn’t. With Juno, users and developers can see and verify the code they are interacting with. I'm not saying everyone can do that, it means some expert could, and tell the rest of us what they find.
This is not to say that Juno is strictly better than AWS. AWS and Juno have different technological affordances and are well-suited to different kinds of applications. Juno is, I would argue, the best way to develop web3 applications today. It’s also worker-owned; it has received no venture capital backing whatsoever.
After this token swap, DAO DAO will be a co-op that runs on top of a co-op, and that co-op will also run DAO DAO. It's co-ops all the way down.
And this state of affairs brings me to an interesting point, and to the reason I came to you today.
A mesh governance protocol
Do you remember the kids’ game where everyone jumps simultaneously while holding a parachute? If you all jump at once (the teacher says), you’ll all float down together. The lesson of the game is that some initial coordination can create a common good. Hold this image in mind for now.
I've written in more technical detail about why and how of all this in a blogpost titled Interchain, and I'd recommend that post to gain a finer-grained appreciation for what I'm about to say.
In a world in which DAOs exist across multiple chains, the aggregate of partially overlapping chain usage creates a “mesh,” a network of DAOs that share an interest in the security of all the chains in the mesh. This—the notion of organizations that share incentives to secure shared chains—is the essence of mesh security. Patterns of demand across populations of organizations create a mesh of mutual interreliance, sharing the incentive to secure one another economically and through governance. This is the vision of the interchain.
Now, let's imagine DAO DAO in an interchain world. In that world, a world where people are using DAO DAO to build co-ops and coordinate those co-ops’ activity across many chains, DAO DAO becomes a mesh governance protocol—a protocol for coordinating governance over the mesh; over the computational resources upon which that mesh relies. We see hints of that world in a DAO DAO/Juno token swap—the people who build the software own the computational resources on which that software relies, and vice versa.
I mentioned earlier the image of the parachute game. The core lesson of this story—that up-front coordination can yield a self-sustaining reaction—is the essence of my theory of the Interchain: initial, emergent coordination can, like the genesis of life from the primordial soup, give rise to a stable, competitive, antifragile ecosystem. Guided by governance, the tools that provision coordination could too become self-sustaining, themselves coordinating at the highest level to form a self-regulating (in the cybernetic sense) economic union.
Regulation
This observation brings us to a critical term: regulation.
We need regulation. Life depends on regulation. Things that regulate themselves live. Lao Tzu would have something to say about this.
How does our economy regulate itself today? Our economy uses formal systems, preferring and rewarding opacity within processing centers (firms), which leads to unpredictability and confounds effective regulation.
I will not opine whether web3 will or can make a “better” (that is, more just, more ecologically bounded) financial system or economy. I do believe web3 could create a safer and more predictable means of coordinating economic activity, one more organic, emergent, and shock-resistant.
I think some nation-states, though perhaps not all, will end up rolling out central bank digital currencies, or CBDCs, to replace their paper money systems. Today, state-run or fiat currency systems use a mix of electronic records and paper to track economic flows across separate balance sheets. These balance sheets are not currently linked to one another by any single system. Even those systems that do exist, like SWIFT, do not internal firm state beyond the transaction they facilitate. CBDCs could create a publicly transparent record of money in circulation and, with some cryptographic trickery, even provide bounded constraints on banking privacy.
Whether or not you or I want to use a CBDC, we can probably all agree that they would make tax authorities' jobs easier. Those tax authorities could see how much money everyone sent and received. It would also be great for law enforcement, who can more effectively crack down on money laundering. (Remember: immutable databases are not a good way to launder money. I asked a federal pursue and an IRS enforcement agent off the record one time whether they preferred to prosecute blockchain cases or cash cases. They both agreed: they'd take a blockchain case any day of the week. It's hard for me to think of a better system for enabling money laundering or tax evasion than a paper cash system.).
I think states understand these benefits. What I think states do not yet fully appreciate is the potential for better—that is, more fine-grained and more predictable—financial regulation.
By “regulation,” I mean specifically the assurance of a stable and predictable economy subject to certain constraints. Central bank constraints today include the “dual mandate” of high employment and acceptable inflation. Plausible constraints may also include maximum acceptable inequality or ecological limits. A regulator, in this sense, is happy when it finds reasonable assurance that its system will continue existing within the specified constraints.
In the latest crypto crash, we saw de-fi protocols built around cybernetic regulators survive. The institutions that failed were the institutions that did not have cybernetic regulation, like Celsius, Galaxy, or Three Arrows Capital. Those institutions did not have in-built control systems for maintaining their stability, and they lost their stability when the world turned. In contrast, Aave and many other protocols are built to withstand stochastic pressures—and, in the last crash, did. In the realm of stablecoins specifically, Terra’s fed-like model perished while Maker’s more algorithmic model briefly de-pegged and then returned to equilibrium. These are examples of control systems making better regulation—engendering more predictability—within firms.
A mesh protocol that coordinates DAOs could—that is, has the opportunity to—do this sort of cybernetic regulation across firms. They could regulate on a whole-market or ecosystem-wide basis. The result of that capacity could be, first, a more stable, predictable, or at least inspectable economic system in the whole, and, second, a more locally-driven, flexible, and naturally-emerging form of regulation in which the cooperatives cooperate to regulate themselves.
First, on the notion that a DAO-wide economy is amenable to more fine-grained regulation. Bashing on economics has become a bit of a hobby horse for non-economists lately. My attitude, in a phrase, is that today's economics is an example of a societal-level risk management strategy. Today's economics is likely not the apex of all economic thinking. I believe a cybernetic, control systems understanding of regulation clarifies the vision of what an economics is supposed to achieve. In the language of Zargham & Shorish’s General Dynamical Systems, regulation might be successful when it finds a global asymptotically stable equilibrium among the aggregate of all entities in its area of economic concern. (In a phrase: generalized dynamic systems, or GDS is a way of thinking about whether complex systems are likely to achieve their objective functions in a dynamic and often unpredictable environment). I expect these and probably other mathematical and analytical tools will end up reforming economics, enabling it to deal with a more multifaceted and complex world, and do so in a more predictable way—if they have sufficient visibility inside of and across firms. It is this within- and between-firm visibility that an economy of DAOs provides.
Second, on the notion that an economy of DAOs could engender a more locally driven, flexible, and emergent vision of regulation. To describe my vision, one articulated by Murray Bookchin, E.F. Schumacher, and others, I will slightly abuse a quote from Theodore Roszak, tiptoeing around a few words that I believe obfuscate more than they clarify. What these thinkers propose is:
"[a] political economy that distinguishes from orthodox socialism and capitalism by insisting that the scale of organization must be treated as an independent and primary problem. The tradition, while closely affiliated with socialist values, nonetheless prefers mixed to 'pure' economic systems. It is therefore hospitable to many forms of free enterprise and private ownership, provided always that the size of private enterprise is not so large as to divorce ownership from personal involvement which is, of course, now the rule in most of the world's administered capitalisms. Bigness is the nemesis…"
The puzzle becomes: what does regulation look like in such a system? How can a system in which bigness is opposed on all grounds possibly regulate at the scope and level of a whole economy, indeed a whole world? The answer, I propose, is a democratic agreement among cooperatives on certain goals for the economic system—for example, low inequality consistent with both full rights and environmental limits. By confederating around these shared goals, cooperatives can consent to shared regulation, one that itself works verifiably across all cooperatives.
One word, "self-regulating," captures both the cybernetic and anti-coercive meanings of this vision. This vision has transnational aspirations. And a mesh governance protocol, perhaps some future version of DAO DAO, could be the thing that coordinates it.
I see the interchain as the beginnings of that system. It is already a set of transnational cooperatives linked in a loose, confederated association. Once DAO DAO performs a few token swaps across a few chains, the world's first mesh governance protocol will be born. DAO DAO will coordinate and structure activity across cooperatives, themselves scattered across many chains. Through the cooperatives increasing ownership over DAO DAO, those co-ops will, someday, convene as DAO DAO, and find themselves facing only one another. They will be left in a body that serves only to allow them to coordinate amongst each other—to compete, and it's important to compete, but more importantly, to cooperate, agreeing upon the rules under which competition can occur. The composition of that body will determine how its economy works.
Internets
A beautiful dream. But what does it have to do with the internet—or, as I, borrowing from Dourish, call it, this internet?
You'll know from my prior talk, This Internet, on the Ground, I believe this internet should be returned to local control. In particular, I think control over the internet should return to autonomous systems, the atomic units that make up the internet we live with today. These autonomous systems could be cooperatively run, instantiated as DAOs. I describe that vision in the seven-part blog series I link to below. What I discover through that series is that this vision—the internet made locally through cooperatively run autonomous systems—is not just a cooperatively run internet but the basis for cooperatively-run governments. It is a sociotechnical sketch for a new model of small-scale mutualism—a new take on ideas like Murray Bookchin's communalism.
You can find that blog series here.
Thank you for your time, and thank you to Anno for having me again.
Thanks to Zeke Medley, Noah Sasso, and Jake Hartnell for the conversations.
This blog is supported by your delegations. Delegate to me to support my work.
What (and why) is DAO DAO?