Reflections from the DWeb Seminar
In August I had the privilege of participating in the DWeb Seminar 2025, an intimate gathering designed to “map the current DWeb technological landscape, learn from each other, and define the challenges ahead”. For those unfamiliar with the event, Wendy Hanamura’s excellent recap captures the spirit and outcomes beautifully. As part of the event we were invited to offer a 15 minute “input talk” to the other participants. I chose to share a fundamental question that has driven Holochain from its inception – and explore how this question shapes not just our technology, but our entire approach to building decentralized systems.
The Core Question: How Does Desirable Social Coherence Evolve?
Everything we do at Holochain (and the projects that I've been nurturing through Lightningrod Labs, like Moss and Acorn) stems from this central inquiry. But what do I mean by “desirable social coherence” and why does it matter?
You can think of social coherence as a group’s long-term stability. Like most things this property exists along a gradient: some social bodies have more coherence than others, which depends on their capacity to respond and adapt to environmental changes as a result of the patterns, practices and organizing principles that they operate by. But therein lies the rub. Some of these patterns provide lots of coherence, but they may not be desirable or pleasant for the individuals taking part in them! It’s no fun for almost everybody involved in an authoritarian regime, but it does have a real degree of stability. My fundamental belief, however, is that not only is it possible to evolve these patterns and processes in directions that participants will find pleasant and desirable, but that doing so actually yields the most long term stability because they will by that fact not contribute to destabilizing it.
The Challenge: Current digital systems scale through centralization and intermediation of critical social functions. Unfortunately, this creates undesirable forms of social coherence – power imbalances that enable both intentional and unintentional abuse. When a few entities control the platforms where billions interact, we may get coherence, but it's often extractive rather than generative. Furthermore our current systems are difficult to evolve because of their very centralization and the interests that want to keep them that way to maintain power.
The Opportunity: Decentralized technology can create substrates for evolvable social coherence – essentially, DNA for social organisms. Instead of rigid, centralized structures, we can build infrastructure that enables new forms of social fabric to emerge and multiple scales, yielding increasing collective intelligence
A key insight here is that there is no single “correct” form of social coherence. What works is contextual, diverse across time, space, and scale. What we need is infrastructure that enables continuous evolution and discovery – balancing stability with emergence.
How This Shapes Our Work at Holochain
This framework isn’t abstract philosophy - it directly informs every architectural decision we make. When building technology to support evolvable social coherence, several principles become essential:
Engagement Spaces as Building Blocks
Human social fabric is built out of layers of interacting and layered “engagement spaces” – essentially social contracts with defined rules. We need infrastructure that makes it easy to create, use, and compose these spaces. The current web may have “solved for” decentralization of publishing - anyone can create a website or blog without permission. But the places where people actually interact and engage with each other (social media platforms, forums, collaborative tools, even finance and accounting tools) remain under intermediary controlled web-servers. Our approach requires protocols where neither the data nor the rules of the group interaction are held by intermediaries.
Agency AND Accountability, Mutually Interwoven
Individuals need genuine agency through their technology - the ability to participate in multiple spaces, move between them, and take their data with them. But this autonomy must be paired with accountability within the contexts where they participate. This tension between empowerment and responsibility is productive, not problematic.
Uncapturable/Unenclosable Carriers: The infrastructure itself must be immune to capture - meaning no single entity can gain enough control to dictate rules, extract value, or shut down the system. We’ve seen far too many examples of infrastructure capture” governments shutting down internet services during protests, platform owners changing terms to benefit their shareholders, or cloud providers being pressured to deplatform users. Even when specific engagement spaces have their own defined rules, the underlying “carrier” of those interactions must remain decentralized. This enables autonomous group formation without intermediation - groups can organize however they choose without worrying that their technological foundation can be pulled out from under them.
Local State, Global Visibility: Rather than forcing artificial global consensus (like blockchains do), we recognize that state is inherently local but can achieve consistent global visibility if nodes share data. Operating this way eliminates unnecessary coordination bottlenecks while maintaining system coherence.
Architectural and Design Consequences
The principles stated above have very concrete design and implementation consequences. For those technically familiar with Holochain you already know how they show up in the design, but here I list some of the key aspects along with pointers to documentation that describe each consequence in more detail.
- Start with a capacity to define & create a known “engagement space”. The “rules of a game”. This consists of the hash of a set of data-types & relations and deterministic validation rules for creation of that data. In Holochain we call this the DNA
- Allow agents to be the authoritative source of all data, i.e. agents “make plays” according to the rules of the DNA.
- Ensure that when this data is shared, it has intrinsic data integrity, i.e. it’s a cryptographically signed append-only ledger for that source (in Holochain we call this the Source Chain), and ensure that it is identifiable as being part of an engagement space by having the first entry in the chain being an agent’s signing of the space’s hash. This is also “I consent to play this game”.
- Share data to an eventually consistent Graphing Distributed Hash Table (DHT), in which other agents validate that all shared data follows the rules of the game.
- Ensure that agents who don’t follow the rules can be blocked/ignored. This prevents capture.
- Allow for “bridge” calls between engagement spaces at the agentic locus (i.e. not at the group level) for composability of spaces. This ensures composibility, autonomy, and accountability
There are of course more details in the design, but these are some of the key ones that fall out of the principles.
Resonance at the DWeb Seminar
What struck me most about the seminar was how much of our framework resonated with challenges other participants were grappling with, even when they approached them from different angles. I would even say that the Seminar itself was fundamentally an example of this thinking. It was a carefully designed set of patterns and processes for a literal engagement space (this time physical instead of digital) whose purpose was to increase the social coherence of players in the p2p domain. These patterns not only included the processes of the input-talks, the unconference sessions, and commitment to production of a collaborative write-up, but also the relational parts of cooking together and sharing non-work time together. All of this together created desirable social coherence. And it’s this pattern that we are all trying to create powerful affordances for in the digital world.
Some further examples: During the unconference sessions, conversations kept circling back to fundamental questions about coordination, autonomy, and accountability.
When we discussed "UI Patterns for Peer-to-peer," I saw it as asking: how do we make decentralized engagement spaces feel natural and empowering to users? When we debated collaborative data model requirements, I saw it as exploring: how do we maintain coherence across distributed participants without sacrificing agency?
When Rae McKelvey shared her focus on "purpose-built apps" that solve real social problems to me that aligned perfectly with the engagement space concept—recognizing that different contexts require different rules and structures.
At the technical level David Thompson's work on object capabilities and Duke Dorje’s work on recryption and identity both live into the same autonomy-with-accountability tension we see as central to social coherence. The ever-present discussions about how best to implement CRDTs (Conflict-free Replicated Data Types, of which Holochain’s DHT is an example) revealed the shared underlying assumption: that meaningful coordination really is possible without central control, that local autonomy and global coherence can coexist, and most profoundly that the infrastructure we build shapes the social possibilities it enables.
But if everything resonated so well, what’s the big deal?
Why This Matters for the Decentralized Ecosystem
Probably the most common complaint I’ve heard over the years from folks who see the astounding potential of decentralized infrastructure goes something like this: “There are so many different p2p solutions, and teams that seem to be working in isolation, why can’t you just agree on a single solution and work together?” On the surface, this sounds like a reasonable complaint, but the lens of coherence helps understand why “working together” is actually such a hard problem to solve.
Recalling from the start of this article: what creates coherence are the patterns, practices and organizing principles of a group. Just because groups have the same goals and want the same outcomes, does not mean that they start their patterns, processes and organizing principles are similar and compatible. In fact, almost always, they aren’t. But this relates to why the DWeb Seminar was so important. It successfully operated according to a higher order organization principle that created an engagement space precisely for the purpose of getting at what patterns, practices and organizing principles folks in the broad DWeb community were operating by, and making them visible and .
So to me this was an example of exactly the underlying principles that we’ve been embedding in Holochain’s architecture from the start.
So, while the decentralized web movement often focuses on technical capabilities – faster consensus, better cryptography, more efficient protocols, we are now seeing the community beginning to seriously see these as means, not ends. The higher level question remains: what kinds of social possibilities can these technologies enable?
This approach enables us to build towards greater “commons enabling infrastructure” - technology that strengthens shared resources and collective capacity rather than extracting value. The creation of digital, unenclosable fabric of engagement spaces is central to this goal. Instead of platforms that capture value from user interactions, we can build infrastructure that enables communities to create and govern their own spaces, according to their own values.
When the decentralized ecosystem embraces this approach, many new possibilities emerge:
- Interoperability with Purpose: We can more easily build bridges between systems that share compatible social intentions. A climate action network could seamlessly share data and coordinate with a local food co-op using a different protocol, supporting community resilience initiatives that address both environmental and food security challenges, while using mutual-credit currencies backed by the productive capacity of the local farms supplying the co-op.
- Governance that Evolves: We can build infrastructure that enables continuous governance innovation rather than trying to solve governance once and for all. A neighborhood mutual aid group could start with simple coordination tools, then gradually evolve more sophisticated decision-making processes as their needs change, without having to migrate to entirely new platforms.
- Network Effects that Serve Users: We can create composable ecosystems where network effects benefit participants rather than extracting from them. As more people join a decentralized social network, the benefits – better content discovery, richer discussions, stronger community bonds - flow to the participants themselves rather than to a platform owner’s advertising revenue.
The Path Forward
The grand challenge of decentralized software is ensuring it actually delivers on evolvable social coherence. This means building infrastructure that serves the flourishing of people and planet rather than extracting from it.
At Holochain, we’re committed to this path, not just in our technology choices, but in how we organize ourselves, engage with our community, and collaborate with other projects. The conversations at the DWeb Seminar reinforced that we’re not alone in this commitment.
The adjacent possibility that Wendy described in her recap isn’t just about new technical capabilities – it’s about new forms of social organization that those capabilities make possible. That’s both a tremendous responsibility and an extraordinary opportunity for all who choose to walk to this path.
