I am posting this on the first full day of Decentralized Web Camp on July 9th, 2026.
TLDR: Infographic!
Doctorow named what we’re all feeling Cory Doctorow gave us the word: enshittification and a description of a pattern that keeps happening.
His original framing, from his January 2023 Pluralistic post:
“First, they are good to their users; then they abuse their users to make things better for their business customers; finally, they abuse those business customers to claw back all the value for themselves.”
Three stages. (A fourth, collapse, is implicit, what happens when the extraction breaks the system.) Amazon, Facebook, Twitter, Google, Uber, Airbnb, every platform that started as a useful service and ended as an extraction machine ran the same playbook.
Doctorow’s point is that this is not a story about bad people. It is a story about structure. Platforms mediate two-sided markets, can reallocate value between sides at zero cost, and once both sides are locked in, the platform’s incentives shift from serving either of them to extracting from both. “Every successful platform becomes a betrayal vehicle.” The diverging of platform interests from user interests is not a failure of the platform. It is the business model.
“We just need a better platform” is not a strategy. A better platform built on the same architecture will follow the same lifecycle. The clock just resets.
But enshittification is the symptom. To run the playbook, a platform first has to have something to extract from. It has to have captured something. That capture has an older name.
Enshittification is what enclosure feels like once you can’t leave
The word for taking something that was held in common and making it private property is enclosure.
Between roughly 1500 and 1850, the English commons — land used collectively by villagers for grazing, firewood, foraging, water — was enclosed. Fenced off. Granted by Acts of Parliament to landowners who could “improve” it for their own profit. The peasants who had used the commons for generations were dispossessed into wage labor. The new private property generated returns for the few who held it. The commons that had fed villages for centuries was gone.
Karl Polanyi argued in The Great Transformation that enclosure was not a natural evolution. It was a political project, sustained by law and violence, that converted a shared substrate of life into private capital. The commons did not vanish because it was inefficient. It was taken. The taking was the act.
The same logic ran through the digital substrate. The things that should have been held as commons — identity, the social graph, group memberships, collective memory, trust relationships — had no protocol layer to hold them. The platforms of the early 2000s emerged into that gap and enclosed them. Each one drew a fence around the part of our social life it mediated and claimed property rights to what was inside. The fences went up because there was nothing to stop them.
Enclosure came first. Enshittification followed. It is only possible to extract from people who have already been captured. The platforms that have enshittified most aggressively are the ones that hold the most of the captured commons. That is not a coincidence. The enclosure is the source of the lock-in. The lock-in is the source of the extraction.
Enshittification is what enclosure feels like once you can’t leave.
What got enclosed
Walk through it concretely:
Identity. You used to be a person. Now you are a Google account, an Apple ID, a Facebook login. The platforms that issued the credentials you log in with own the canonical record of who you are online. If they cancel you, your relationships disappear with you.
The social graph. Microsoft owns LinkedIn, so Microsoft owns the LinkedIn social graph. Every professional relationship you have asserted there — every colleague, every collaborator, every introduction, every weak tie that turned into a job or a partnership or a friendship — is Microsoft’s asset. Not in some loose sense. In the literal, accounting sense: it shows up as goodwill on Microsoft’s balance sheet and as defensive moat in their competitive strategy. Meta owns Facebook, so Meta owns the Facebook social graph. X owns the Twitter follow graph. Discord owns its server membership graphs. Each enclosed by a different platform; none of them held in common; none of them yours.
Groups themselves. This is the one that gets least talked about, and it may be the deepest enclosure of all. A group — a congregation, a soccer league, a watershed council, a neighborhood association, a community garden, a union local, an affinity group, a movement chapter — is a thing in the world. It has its own identity. It has its own roster. It has its own history. It has its own collective will. None of these are facts that belong to a vendor.
But in the platform era, the group does not own itself. The Slack workspace belongs to Slack. The Discord server belongs to Discord. The Facebook Group belongs to Meta. The community has no protocol-level standing. It exists inside a product. If the product changes its terms, the group is at the product’s mercy. If the product shuts down, the group disappears with it. If the host of the workspace leaves the community, the workspace can go with them. Groups, in the current architecture, are not first-class entities. They are features inside someone else’s tool.
This is not how groups have ever worked before the internet. A congregation that has been meeting for two centuries owned itself. A union local owned its membership list. A guild owned its traditions, its standards, its right to admit and expel. The community held itself. No outside vendor held a claim to its identity or its roster or its collective memory.
Digital enclosure changed that. And the result is that every community that has moved its life online has been quietly turned into a tenant in someone else’s building.
Collective memory. Every document, every decision, every conversation, every photo, every record of what your group has been and done — all of it lives on someone else’s servers, under someone else’s terms.
Trust relationships. Reputation, references, vouching, recommendations — all platform-mediated. LinkedIn endorsements. Yelp reviews. Uber ratings. Whatever trust you have built with people you have met online is held in the proprietary databases of whoever mediated the connection.
Each of these was once properly nobody’s property. Each of them is now somebody’s asset. This is enclosure.
Doctorow’s prescription stops short
Doctorow names a structural answer almost as explicitly as he names the structural problem. His prescription has three parts: the end-to-end principle, freedom of exit, and interoperability. He particularly champions adversarial interoperability (or competitive compatibility, “comcom”): the right of third parties to build interoperable tools — scrapers, clients, alternative front-ends — that pry captured platforms open from outside, without asking permission.
This prescription is a retrofit. Doctorow’s argument mostly assumes the captured platforms are the starting condition, and asks: how do we force them to interoperate? How do we mandate that they let users leave with what they came with? His regulatory recommendations follow the same shape — lower exit costs, force open APIs, legalize reverse engineering.
But notice what this prescription accepts. It accepts the enclosure. It accepts that LinkedIn owns the professional social graph, and asks for the right to scrape it. It accepts that Facebook owns your friend list, and asks for the right to take a copy. It accepts that Slack owns your community’s workspace, and asks for an export button. It accepts the fence and asks for a gate.
The deeper move is to refuse the property right at the root. The social graph was never LinkedIn’s to own. Identity was never Google’s to issue. A community’s roster was never Slack’s to keep. These are not facts that belong to vendors. They are facts that belong to the people, the relationships, and the groups they describe.
The architectural form of that refusal is a protocol substrate. Hold the commons as commons, in open protocols, so the enclosure is impossible in the first place. Build the substrate so the fence cannot be built.
This is what Doctorow’s diagnosis demands but his prescription does not quite reach. He is fighting the rear-guard action well. The forward move is to deny the enclosure altogether.
Open protocols already resist enclosure
The open protocols of the early internet already deliver this kind of architecture, in the layers where they exist.
SMTP — the email protocol — was finalized in 1982. It has not been enshittified, because no one owns it. The protocol holds the substrate as commons. Email addresses are not anyone’s property. Mail delivery is not anyone’s monopoly. You can use Gmail or Protonmail or Fastmail or run your own mail server. You can switch between them without losing the ability to reach anyone.
The same is true for the web. HTTP and HTML are open. You can read this article on any browser, any device, any operating system, without asking permission. The web is good not because of any one company. It is good because the protocol underneath it is a commons that has not been enclosed.
These protocols cannot be enshittified because they cannot be enclosed. There is no platform owner whose incentives can decay. SMTP has no CEO, no board, no quarterly earnings, no venture capitalist who can pivot it when growth slows. It is a shared agreement, maintained as a commons.
But SMTP was for messages. HTTP was for documents. Neither was for identity, groups, or trust relationships. When those things became valuable, there was no commons-holding architecture to receive them. So the platforms enclosed them, and we have been living with the consequences ever since.
Twenty years of quiet work
The missing commons-holding architecture has been under construction for over a decade within a community of passionate practical idealists. Beginning in 2005, user-centric identity projects started gathering at the Internet Identity Workshop (IIW). Twenty years later, IIW is still meeting every six months. The work has been slow, technical, often invisible — and steadily productive. The pieces of the missing commons-holding layer have been assembled, one at a time:
- Decentralized Identifiers (DIDs) — a W3C standard since 2021. A way to have a verifiable identifier that no platform owns. Your DID is yours. You hold the keys.
- Verifiable Credentials (VCs) — a digital equivalent of a membership card. Cryptographically signed, privacy-preserving, presentable on your own terms.
- Digital wallets and digital agents — software that lets you hold credentials and identifiers yourself, rather than having them held by a platform.
- The Trust Over IP (ToIP) stack — a four-layer protocol architecture for the trust layer, modeled on TCP/IP. Including the Trust Spanning Protocol (TSP), the protocol-layer equivalent of “any host can talk to any other host” for trust.
- Private channels — pairwise relationships that nobody else can observe or intermediate. Not “encrypted on a corporate server.” Actually peer-to-peer.
None of these are speculative. All of them are shipping. Hundreds of decentralized identity projects use DIDs in production. The ToIP stack is implemented in open source. Major implementations exist in healthcare, education, financial services, and government — quietly, without fanfare, because the work is infrastructural.
Assembling the commons-holding protocol stack
The First Person Project (FPP) is the live, working effort to put the pieces together. Born in 2024 as a collaboration between Linux Foundation Decentralized Trust (LFDT), the Trust Over IP Foundation (ToIP), the Ayra Association, the Decentralized Identity Foundation (DIF), and the OpenWallet Foundation (OWF) — it is over 200 contributors and growing.
Its central insight: digital trust is grounded in first-person trust relationships, and those relationships can be expressed in two new types of verifiable credential:
- Personhood Credentials (PHCs) — issued by an organization, community, or institution that can attest that the holder is a real, unique human within that context. Not a global biometric database. No surveillance. A privacy-preserving way to say “this is a person.”
- Verifiable Relationship Credentials (VRCs) — issued peer-to-peer between people, attesting that two specific humans have a real first-person relationship.
Together, these form a Decentralized Trust Graph. The social graph held in protocols, not in a platform. The cryptography is open. The credentials live in the holder’s own wallet. The governance is open. Microsoft does not own it. Meta does not own it. No one owns it, because it is a commons.
The Linux Foundation is the most concrete current deployment. In March 2025, LF Executive Director Jim Zemlin gave the opening keynote of the LF Member Summit explaining why the Linux Foundation needed First Person credentials. The motivating story: the XZ attack of 2024, in which unknown attackers spent two years building up a fake open-source contributor identity (“Jia Tan”) until they obtained maintainer rights to a utility used in nearly every Linux distribution. A “Jia Tan” with no actual personhood attestation and no verifiable relationship credentials from real contributors could not have gotten near maintainer rights. The work on this is underway in the Linux Kernel project now.
Governance is being built through the First Person Cooperative — a network cooperative that holds the trust network as a commons. Not a startup. Not a foundation grant. A cooperative governance structure that ensures the trust layer cannot be sold, pivoted, or shut down by any single party. It is the IETF model, applied to trust. The commons-holding governance to match the commons-holding architecture.
But individuals owning themselves is only part of it. One layer up sits what FPP calls Verifiable Trust Communities (VTCs) — groups holding themselves at the protocol layer, rather than existing as features inside a vendor’s product.
A community’s identity, roster, and history become things the community itself owns. The deeper treatment of this layer is the subject of the companion piece Free Our Groups: From Platforms to Protocols; for the purposes of this article, the point is that the architecture is real. Project Weave is a live effort to build the protocol layer for communities of all kinds.
Refusing to be enclosed
Doctorow named the extraction. The deeper move is enclosure. Enshittification is what platforms do to the captured commons; enclosure is what made the commons capturable in the first place.
The structural cure is not to pry the fences open. It is to refuse the property right that put the fences up. Adversarial interoperability accepts the enclosure and asks for a gate. Protocols deny the enclosure ever had a right to exist.
When your identity is a Personhood Credential in your own wallet, no platform owns it. When your relationships are Verifiable Relationship Credentials, no platform owns them. When groups hold themselves as Verifiable Trust Communities, the platform becomes a tool the community uses, not a captor that holds it.
You can’t enshittify a protocol because you can’t enclose an open protocol. There is no property right to assert, no fence to draw, no rent to extract. The substrate is held in common. The incentive structure that drives enshittification has no point of purchase.
This is not theoretical. It has been working for SMTP and HTTP for four decades. The reason it has not been working for identity, relationships, groups, and trust is that the protocol layer for those things did not exist. After twenty years of work by a quiet, persistent community, it now does.
The English commons was enclosed because the people who used it could not defend the substrate they depended on. The digital commons was enclosed for the same reason. We have a second chance now. The substrate is built. People can own themselves. Groups can own themselves. The commons can be reclaimed.
The question is whether we recognize the moment in time to act, before the next generation of platforms encloses the next twenty years of digital life as thoroughly as the last twenty.
Kaliya Young is a digital identity expert, co-founder of the Internet Identity Workshop (IIW), and co-founder of Project Weave. She has been working on the trust layer for the internet since 1999.
Companion pieces:
Leave a Reply