Ever since I first heard the term, the "DAO" or Decentralized Autonomous Organization, has captured my imagination in a way few other technological ideas have in my career. When BlueYard hosted our first mini-conference, Decentralized & Encrypted, in Berlin in 2016 the conversation between the participants orbited around the original DAO (called simply "The DAO", which was subsequently and notoriously hacked) and the directions this new economic lifeform might take. There was of course talk of artificial intelligences participating in, controlling, and even becoming DAOs way back then. But we also heard from Primavera De Filippi about her self-propagating art project, Plantoids, a statue which would order parts and reproduce itself. Imagine unstoppable processes running on unstoppable decentralized computers with their own will and their own ability to interact with the world. Imagine self-propagating physical objects, reproducing or even creating new types of objects. Would this lead to a new kind of technological evolution?

The possibilities were mind-boggling.

Since then, DAOs have been a constant focus of discussion, development, and organization in the crypto world. Countless startups have built software to help instantiate and manage DAOs. DAOs have formed around themes as diverse as purchasing a sports team, creating new ways to fund scientific research, and investing in profile picture NFTs.

Yet with all of this activity, the disappointing fact is that most DAOs are nothing more than a treasury and a chat room. And despite the science-fiction-like possibilites we all fantasized about almost 8 years ago, the decentralized autonomous organization has become a decidedly "soft" area of progress in crypto.

What really is a DAO? A decentralized collective of people doing something together? No. We already had that before the term DAO was coined. A DAO is defined by its code. That's what makes DAOs special and new. They exist as code, and humans (and their bots) interact with them in only the ways the code allows.

As of now, almost all DAO code is entirely centered on the intersection of treasury management and voting. Participants contribute to the treasury in some way, usually as a part of the fee to join the organization. Then members can propose actions for the DAO to take and other members vote. In the most precise cases, these votes are to determine how to spend money. And much of the actual spending of that money ends up being performed by humans in a way that isn't ultimately controlled by the DAO's code.

DAOs are decentrlaized but are pretty far from the definition of "autonomous" I would expect as an outsider. DAOs are, on average, still very soft technology. Until DAOs are truly trustless and autonomous there is nothing you can do with a DAO that wouldn't be possible, and probably better done, with a traditional centralized organization.

So if DAOs are soft, what would a hard DAO be? Here are some attributes I'd look for:

• Programmable - A DAO is ideally defined by its code. The DAOs capabilities are expressed in code. A DAO should therefore be programmable.

• Composable - Programmable organizations can and should also be composable. This means they can be chained together, swapped out, and combined in emergent and unexpected ways. The DAO APIs must be created with composability and interoperability in mind. Frameworks and standards (even defacto) will help define the types of capabilities DAOs might have and allow DAOs of similar capability to be interchanged or to plug in trustlessly to other DAOs.

• Automated - The actions the DAO takes should happen automatically without human intervention. If its code (or code it manages) should be upgraded, these upgrades should happen automatically after a successful vote or other trigger condition. If the DAO is to interact with another contract, protocol or external system, these interactions should happen atomically and automatically when the DAO decides to take the actions. When a system isn't automated, it isn't trustless.

• Tamper-proof - For obvious reasons

• Unstoppable - If humans can intervene and stop a DAO from functioning, it isn't actually autonomous. One of the key powerful concepts in crypto networks is they are censorship resistant and unstoppable. A truly decentralized autonomous organization must be similarly unstoppable and untamperable by individual humans (or other agents).

In the same way we talk of narrow vs broad AI, a hard DAO might be narrow or broad. The "harder" a DAO is, the more constrained it is, which means initial hard DAOs are likely to be restricted to a narrow set of actions. This is OK and possibly even desirable in a complex, network of DAOs/agents as it would encourage modularity and encourage systems to do one thing well vs many things poorly. DAOs are programs, after all, and good programs are small, modular, and decoupled.

The transition towards hard DAOs is more than a technical shift; it's an opportunity to revolutionize digital governance and collaboration. With hard DAOs, we could see the emergence of fully automated marketplaces, decentralized venture capital funds operating with unparalleled transparency, or even autonomous networks managing renewable energy distribution. These DAOs, built to be programmable, composable, automated, and secure, could significantly enhance efficiency and transparency.

I asked GPT4 to read this article and give me some more ideas of the kind of things hard DAOs could do. Here they are, gently coaxed but un-edited.

1. Autonomous Traffic and Transit Systems: Hard DAOs could manage and optimize public transportation and traffic systems in real-time, adjusting routes and schedules based on traffic data and passenger demand, potentially integrating with autonomous vehicle networks.

2. Decentralized Energy Grids: These DAOs could autonomously govern and balance energy production, distribution, and consumption across a decentralized grid, facilitating peer-to-peer energy trading and optimizing for renewable sources based on real-time data.

3. Automated Disaster Response: By monitoring environmental data, Hard DAOs could automatically deploy resources and coordinate response efforts in the event of natural disasters, such as earthquakes or floods, ensuring timely and efficient aid delivery.

4. Smart Public Health Systems: Utilizing health and environmental data, these DAOs could manage public health initiatives, automatically deploying resources for vaccination drives or epidemic containment based on predictive modeling and real-time data.

5. Decentralized Urban Planning: Hard DAOs could facilitate participatory urban planning, using simulation models to predict urban growth and infrastructure needs, and autonomously funding and deploying urban development projects based on community voting and needs assessment.

6. Automated Environmental Protection and Conservation: These DAOs could manage conservation areas autonomously, using sensor networks to monitor wildlife and environmental health, and deploying resources for habitat restoration or anti-poaching measures based on real-time data.

7. Decentralized Autonomous Factories: Factories run by DAOs could manage production lines, supply chains, and maintenance schedules autonomously, adjusting operations in real-time based on demand signals, material availability, and machine status.

8. Water Resource Management: Hard DAOs could optimize water distribution and usage in agriculture and urban settings, allocating water resources efficiently based on real-time data on weather, usage patterns, and reservoir levels.

9. Waste Management and Recycling: These DAOs could optimize waste collection routes and schedules based on data inputs, and manage recycling processes, automatically sorting and routing materials for efficient recycling or disposal.

10. Autonomous Real Estate Development and Management: DAOs could manage real estate assets, executing leases, managing maintenance, and even autonomously buying or developing properties based on market conditions and pre-defined investment criteria.

Ever since I first heard the term, the "DAO" or Decentralized Autonomous Organization, has captured my imagination in a way few other technological ideas have in my career. When BlueYard hosted our first mini-conference, Decentralized & Encrypted, in Berlin in 2016 the conversation between the participants orbited around the original DAO (called simply "The DAO", which was subsequently and notoriously hacked) and the directions this new economic lifeform might take. There was of course talk of artificial intelligences participating in, controlling, and even becoming DAOs way back then. But we also heard from Primavera De Filippi about her self-propagating art project, Plantoids, a statue which would order parts and reproduce itself. Imagine unstoppable processes running on unstoppable decentralized computers with their own will and their own ability to interact with the world. Imagine self-propagating physical objects, reproducing or even creating new types of objects. Would this lead to a new kind of technological evolution?

The possibilities were mind-boggling.

Since then, DAOs have been a constant focus of discussion, development, and organization in the crypto world. Countless startups have built software to help instantiate and manage DAOs. DAOs have formed around themes as diverse as purchasing a sports team, creating new ways to fund scientific research, and investing in profile picture NFTs.

Yet with all of this activity, the disappointing fact is that most DAOs are nothing more than a treasury and a chat room. And despite the science-fiction-like possibilites we all fantasized about almost 8 years ago, the decentralized autonomous organization has become a decidedly "soft" area of progress in crypto.

What really is a DAO? A decentralized collective of people doing something together? No. We already had that before the term DAO was coined. A DAO is defined by its code. That's what makes DAOs special and new. They exist as code, and humans (and their bots) interact with them in only the ways the code allows.

As of now, almost all DAO code is entirely centered on the intersection of treasury management and voting. Participants contribute to the treasury in some way, usually as a part of the fee to join the organization. Then members can propose actions for the DAO to take and other members vote. In the most precise cases, these votes are to determine how to spend money. And much of the actual spending of that money ends up being performed by humans in a way that isn't ultimately controlled by the DAO's code.

DAOs are decentrlaized but are pretty far from the definition of "autonomous" I would expect as an outsider. DAOs are, on average, still very soft technology. Until DAOs are truly trustless and autonomous there is nothing you can do with a DAO that wouldn't be possible, and probably better done, with a traditional centralized organization.

So if DAOs are soft, what would a hard DAO be? Here are some attributes I'd look for:

• Programmable - A DAO is ideally defined by its code. The DAOs capabilities are expressed in code. A DAO should therefore be programmable.

• Composable - Programmable organizations can and should also be composable. This means they can be chained together, swapped out, and combined in emergent and unexpected ways. The DAO APIs must be created with composability and interoperability in mind. Frameworks and standards (even defacto) will help define the types of capabilities DAOs might have and allow DAOs of similar capability to be interchanged or to plug in trustlessly to other DAOs.

• Automated - The actions the DAO takes should happen automatically without human intervention. If its code (or code it manages) should be upgraded, these upgrades should happen automatically after a successful vote or other trigger condition. If the DAO is to interact with another contract, protocol or external system, these interactions should happen atomically and automatically when the DAO decides to take the actions. When a system isn't automated, it isn't trustless.

• Tamper-proof - For obvious reasons

• Unstoppable - If humans can intervene and stop a DAO from functioning, it isn't actually autonomous. One of the key powerful concepts in crypto networks is they are censorship resistant and unstoppable. A truly decentralized autonomous organization must be similarly unstoppable and untamperable by individual humans (or other agents).

In the same way we talk of narrow vs broad AI, a hard DAO might be narrow or broad. The "harder" a DAO is, the more constrained it is, which means initial hard DAOs are likely to be restricted to a narrow set of actions. This is OK and possibly even desirable in a complex, network of DAOs/agents as it would encourage modularity and encourage systems to do one thing well vs many things poorly. DAOs are programs, after all, and good programs are small, modular, and decoupled.

The transition towards hard DAOs is more than a technical shift; it's an opportunity to revolutionize digital governance and collaboration. With hard DAOs, we could see the emergence of fully automated marketplaces, decentralized venture capital funds operating with unparalleled transparency, or even autonomous networks managing renewable energy distribution. These DAOs, built to be programmable, composable, automated, and secure, could significantly enhance efficiency and transparency.

I asked GPT4 to read this article and give me some more ideas of the kind of things hard DAOs could do. Here they are, gently coaxed but un-edited.

1. Autonomous Traffic and Transit Systems: Hard DAOs could manage and optimize public transportation and traffic systems in real-time, adjusting routes and schedules based on traffic data and passenger demand, potentially integrating with autonomous vehicle networks.

2. Decentralized Energy Grids: These DAOs could autonomously govern and balance energy production, distribution, and consumption across a decentralized grid, facilitating peer-to-peer energy trading and optimizing for renewable sources based on real-time data.

3. Automated Disaster Response: By monitoring environmental data, Hard DAOs could automatically deploy resources and coordinate response efforts in the event of natural disasters, such as earthquakes or floods, ensuring timely and efficient aid delivery.

4. Smart Public Health Systems: Utilizing health and environmental data, these DAOs could manage public health initiatives, automatically deploying resources for vaccination drives or epidemic containment based on predictive modeling and real-time data.

5. Decentralized Urban Planning: Hard DAOs could facilitate participatory urban planning, using simulation models to predict urban growth and infrastructure needs, and autonomously funding and deploying urban development projects based on community voting and needs assessment.

6. Automated Environmental Protection and Conservation: These DAOs could manage conservation areas autonomously, using sensor networks to monitor wildlife and environmental health, and deploying resources for habitat restoration or anti-poaching measures based on real-time data.

7. Decentralized Autonomous Factories: Factories run by DAOs could manage production lines, supply chains, and maintenance schedules autonomously, adjusting operations in real-time based on demand signals, material availability, and machine status.

8. Water Resource Management: Hard DAOs could optimize water distribution and usage in agriculture and urban settings, allocating water resources efficiently based on real-time data on weather, usage patterns, and reservoir levels.

9. Waste Management and Recycling: These DAOs could optimize waste collection routes and schedules based on data inputs, and manage recycling processes, automatically sorting and routing materials for efficient recycling or disposal.

10. Autonomous Real Estate Development and Management: DAOs could manage real estate assets, executing leases, managing maintenance, and even autonomously buying or developing properties based on market conditions and pre-defined investment criteria.