As I understand it, the concept of the “adjacent possible” describes changes just within reach given the current state of a system’s knowledge, resources, and components. It’s a “shadow future” of the possibilities on the edge of the present. The adjacent possible asks: what can you learn from existing building blocks to create new ones?
That is, at heart, what the DWeb Seminar set out to do.
THE EXPERIMENT: Over the course of three days and nights, in a house in the Presidio of San Francisco, can you take 10 core peer-to-peer (P2P) developers, 2 research directors, 1 editor, and 3 stewards, mix them up, expose them to provocative prompts, and see if some breakthroughs in understanding and technical consensus are possible?
THE PRECEDENT: The inspiration and format of the DWeb Seminar comes directly from Professor Christian Tschudin’s P2P Basel, the annual workshop he runs at the University of Basel for offline-first, P2P researchers and protocol builders in Europe. Over the course of five P2P Basel annual gatherings, Christian and his associate research director, Erick Lavoie, have honed the recipe: 10 participants (not more), cooking and doing dishes together (a key lubricant), in a smallish space (no escape!), for three days (it’s exhausting!).
At DWeb Camp 2024, I heard from DWeb Fellow Andreas Dzialocha (P2Panda) that this three-day workshop in Basel had been a turbo-charging event for his work. To be in a place where everyone understands intimately what a CRDT (Conflict-free Replicated Data Type) is, has a basic grasp of cryptography — that allows the participants to move fast and deep very quickly. I wondered: could we replicate what the folks in Europe were doing in the Americas?
THE RAMP UP: In August 2024, on the balcony of the Hackers Hall at Camp Navarro, I approached Christian Tschudin with an idea: would he work with me to create a DWeb version of his workshop, across many DWeb projects and specialties? I started to introduce him to DWeb protocol builders and he attended their technical talks to begin understanding developers’ focus areas in the DWeb ecosystem. We sought to have representatives from every part of the tech stack, and luckily, at DWeb Camp 2024 you can find many accomplished protocol builders just chatting around the campfire.
The Japanese have a core concept called nemawashi – preparing the roots – which is critical to any venture’s success. It’s a time of informal listening, building relationships, coming to a shared vision before any formal announcement is made. Late 2024, Christian and I started our nemawashi. We needed to create a budget, raise the money, so eventually we could hire an Associate Research Director, an editor for an eventual publication, and stewards for the process. I wanted to add on a DWeb Weekend, where the public could contribute ideas and learn from our participants.
Fortunately, the Internet Archive’s founder, Brewster Kahle, was enthusiastic about sponsoring the event from the get-go. He has always been about building working code, testing it in the wild, making tools that are useful for people. So a week-long event anchored at the Internet Archive, dedicated to deepening our understanding of what remains to be built – that resonated with his call long ago to create a “third path” for developers who want to use technology to help humanity. Brewster and his wife Mary Austin even allowed 16 of us to use every nook and cranny of their house, to camp in their backyard, cook in their kitchen, and debate around their dining table long into the night.
THE EXECUTION: We decided to title the DWeb Seminar “Current Science & Grand Challenges” to map the current and future DWeb landscapes, and from August 13-15, 2025, we gathered some of the top researchers and builders in the Americas. There were experts in object capabilities (David Thompson/Spritely Institute) and applied cryptography (Ying Tong Lai.) Our editor, Dmitri Zagidulin brought with him deep knowledge of data types and decentralized identity (DID) standards. Some were founders of venture-backed startups (Brendan O’Brien/iroh) or worked for well-funded P2P companies (Rae McKelvey/Ditto). Matthew Weidner studied CRDTs for his PhD (now on hold.) Others worked on grants and contracts, in small teams or in universities. Many were experts in the data layer of the stack, able to debate the ideal decentralized stack like few others in the world could. We sought a group with a shared vocabulary, diverse specialties, collaborative temperaments, and a willingness to listen to others.
THE PHILOSOPHY: Whenever you gather a small group of highly knowledgeable people, some intimidation and fear can creep in. Many confessed later that they harbored feelings of “impostor syndrome,” and feared that they wouldn’t be able to keep up or be understood. But anticipating this, our research directors, Christian Tschudin and Andreas “adz” Dzialocha created a set of “mantras.”
When I asked our organizing team what they hoped to create, they replied, “A safe space,” “a house with good food, and interesting conversation,” “a caring environment” for growth.
I believe if you create a container where people feel safe, they will feel free to argue, and from that friction often comes accelerated progress. What I wanted was for this deeply talented group of people, who often feel alone and unsupported, to know that others share their burdens, encourage their successes, and are there to help burnish their ideas. In this comfortable house, filled with delicious home-cooked meals, what kind of heat could we generate from the clash of ideas about how to build the DWeb?
THE PROCESS: We started the proceedings with a “keynote” talk by Danny O’Brien and Kurt Opsahl, both formerly with the Electronic Frontier Foundation and now with the Filecoin Foundation, who tried to set the regulatory scene for our builders. When you are heads down in your code, it’s hard to keep your eyes trained on the tech policy regimes in each country. Kurt and Danny were both encouraging. “You have the power to create the rules,” Danny emphasized. “Dreams are what our stuff is made of,” declared Kurt. “Think a lot about how to build good values into your code.”
On the first day, each participant gave a 15-minute “input” talk about a topic of their choosing. They ranged from “Invariant-centric threat modeling” by Ying Tong Lai to “Homomorphic Quantum-resistant Proxy Reencryption (aka Recrypt) by Duke Dorje (Identikey). Rae McKelvey traced her decade-long career from PhD student, DAT developer, Awana Digital developer, Ink and Switch researcher, to venture-backed team manager. She recognized that her throughline is “purpose built apps” that solve societies’ problems with code. Brendan O’Brien (iroh) shared tips for picking partners to build the “Cozy Web, “ while Eric Harris-Braun (Holochain Foundation) explained why social coherence is his “Big Why” for building decentralized tech. Around the dining room table, our group came to understand a little better where each person was coming from. Were we brave enough to share not only our conceptual frameworks, but also the problems we were having trouble solving?
Days 2 and 3 were pure unconference, as the group curated rounds of discussions on topics they wanted to dive into deeply. Redesigning an “Internet 2” stack drew from every area of the group’s expertise; UI Patterns for Peer-to-peer, and collaborative data model requirements were other hot topics.
Over homemade falafel and fudge brownies, the group powered through topic after topic. Byzantine Fault Tolerance (BFT) blended nicely with barbequing vege-kebabs and much CRDT work (aka –Cooking Real Delicious Tri-tip) was accomplished by all. Matthew Weidner proved he had mastered decentralized collaboration, at least with his partner Tiffany as they whipped out fruit galettes. We discovered the best conversations occur over a basin of soapy dishes.
THE PUBLIC: Next this group brought their core questions to the public in the DWeb Weekend (August 16-17) at the Internet Archive. Day 1 focused on unconference-style discussions, while Day 2 was devoted to talks and hands-on workshops. Across five spaces and three sessions, the community brainstormed on topics such as Christian Tschudin’s table of contents for a DWeb Textbook, and Commons Infrastructure with Dmitri Zagidulin. Duke Dorje led us through a simulation of a future-state in which censorship, statelessness, and dissolving trust are answered in part by new forms of cryptographic key management.
On Sunday, people got the chance to play with apps built on Holochain’s Moss or send small packets from device to device using Tiny SSB. In one room was a debate with HyperHyper Space’s founder, Santiago Bazerque, and in another were coders testing out Greg Slepak’s (OK Turtles Foundation) Shelter Protocol and its first implementation, Chelonia. David Thompson (Spritely Institute) explained the nuts and bolts of object capabilities, helping a roomful of CRDT experts imagine how they might build on top of Spritely. Later, Matthew Weidner drew from his academic work to show how collaborative editing might be possible in the DWeb. Akhilesh Thite demoed his Peersky Browser, a browser that natively runs decentralized protocols like IPFS, IPNS, and Hypercore. And rounding out the day, Ying Tong Lai took us through a threat modelling exercise with an eye to the fact that the EU is rolling out digital IDs in just a few years.
THE OUTCOMES: The DWeb Seminar was structured with a tangible output in mind: that the participants would produce a paper synthesizing what happened during our time together. Christian mused it could capture a “timestamped zeitgeist” so that in a decade we can look back at what this group thought was important.
The idea of a DWeb Textbook is being nurtured as an open source, community collaboration, with the assets to be stored in the DWeb’s Gitlab. A group of a half dozen people stepped forward to get the ball rolling.
Among the Seminar participants, there seems to be emerging a first rough consensus on how a “Base CRDT” could be configured. From concept drawings to presentation slides, our group is sharing their architectural ideas with the wider public, refining the principles, soliciting feedback.
As I helped do the final cleanup of our homey venue, the post-its left behind formed a patchwork quilt of key challenges to solve: Pruning, Composability, DAGs, BFTs, Authorship Provenance. These are the “shadow future” on the edge of our DWeb present. But the real impact of our week is undoubtedly human. Can Andreas run P2Panda over David’s Spritely? Should this group be building atop Brendan’s iroh stack? Now that Eric and Santiago share a common vocabulary, is collaboration more likely? Only time will tell. The mission of the DWeb is to “connect people, projects and protocols to build a decentralized web,” but there is a reason we put “people” first.
THE WHY: When Brewster asked me to produce the first Decentralized Web Summit in 2016, we featured the “Father of the Internet” Vint Cerf, and the “Father of the Web” Sir Tim Berners-Lee, pioneers in the brave, new online world we now inhabit. The original 80 people we invited to Builders Day included the next generation of builders of a radical redesign of Web 2.0. Today, nine years later, yet another generation of P2P architects is sketching out the stack they hope others may one day come to use. I believe this week of deeply esoteric conversations has brought us a few steps closer. (Brilliant minds + human connections = potential breakthroughs.)
But why do I commit myself to this vision? In 2023, I retired from the Internet Archive, and yet here I am camped out in a house with 15 developers. It doesn’t take a seer to imagine our near future: massive amounts of data are erased from the Web; truth fractures; climate change and war create people who are stateless and without official identities; the chasm between the powerful and powerless grows ever wider. These problems are creating demand for new and better tools. Users are looking for alternatives in the marketplace. Our friends at Bluesky could not have predicted Elon Musk’s Twitter take-over, but fortunately, their protocol and app were relatively ready for the moment when millions flocked to them away from X. We need to be ready.
The moment for local-first, P2P working code is here. As Larry Lessig said, “Code is Law” and if you write good values into the code, perhaps that can combat the seemingly intractable power imbalances before us. Maybe not. But I want to do everything in my power to support those who are speaking righteously through their code. I believe the adjacent possible is attainable.
NOTE: to watch the talks from DWeb Weekend please visit https://archive.org/details/DWeb-Seminar-2025
