I used to like walking through old libraries, thinking about how each book I touched hadn’t been opened in decades. The internet is much the same. Come have a look what’s here.
The pool is drawn from the web’s least-visited shelves: mirrors of sites dead since 2009, machine transcripts of 1911 newspapers, the second photograph of a museum bird skin, sequencing reads filed once in 1995. Each page arrives with an estimate — error bars included — of how long it has gone unseen by human eyes.
Each page is offered once, to one visitor. The moment you look at it, it stops being a page no one has seen — so after you, it is retired. If someone beats you to it by seconds, you’ll be told you were late.
The ledger records what has been witnessed, and when: pages that waited years for one visit, and got it.
Fair warning: these pages are brittle. Their only regular visitors have been machines, and many now sit behind walls that turn even the machines away — so some can’t be checked before you’re sent, and servers this old fail in creative ways. If you arrive to an error, that is part of the terrain. Sometimes when you open one of those old books in the library, all the pages fall out.