Information disappears in many more ways than people realize, and the disappearance has accelerated. A newsroom shutters and its archive goes 404. A journalist dies and their notes vanish from devices nobody else can unlock. A platform's policies shift and a decade of writing is deleted overnight. A government pressures a publisher into removing a story, or quietly de-lists it from search. A scraping bot ingests a piece of writing into an AI training set without consent, and the original creator loses the ability to control what is said about their work, by whom, or for what purpose. A piece of evidence in a court case is taken down for legal reasons and never re-emerges. A community wiki loses its maintainer. A link rots.

These look like separate problems. They share an architecture. Every one of them is a case where the systems we use to publish, host, and connect information assume that someone, somewhere, has both the will and the ability to preserve it — and that assumption fails silently and often. The decentralized web community has been responding to many of these failure modes for years. The workshop assumes that work as a starting point — and asks, with the people doing it in the room, what the next layer of coverage might look like.

The session is a working hour, not a presentation. I'll spend the first few minutes framing the problem space (including a brief reference to web architecture's long history of leaving preservation as somebody else's problem — from the missing HTTP 402 to the assumption that hosting is permanent), and then the rest of the time is the room's. The goal is not consensus or a finished design. The goal is a shared map.