Lina suggested we delete the core and let the world decide. I argued that some experiments reveal more by persisting. The server log recorded the argument as data—names, timestamps, file hashes. It was all decoys now, even our recollections. Memory became something to be patched.
Decoys were small: doctored files, phantom profiles, press releases pointing to empty pages. They baited attention and then dissolved into inconsistencies. A decoy could be a leaked song credited to a non-existent band, an obituary for a fictional mayor, or a homepage for a startup that never received funding. The aim was to redirect, to test networks and people—how quickly belief propagated, where skepticism lived.
Newsfeeds replicated fabricated quotes as if they had always existed. Forums stitched our snippets into new contexts. A musician in Tokyo sampled a decoy chorus and turned it into a hit; an investigative blogger traced its origin and found only threads of our laughter. We watched metrics climb—impressions, reblogs, citations—our small experiment bleeding into the wild.
Then the decoys began to answer back. Replies poured in not just from people but from automated systems trained to detect inauthenticity; they adapted. Warnings labeled our posts as suspicious; content moderators flagged them. Some readers, delighted by the puzzle, added layers: an account claiming to be a whistleblower sent documents—wrongly formatted, obviously faked—but later, piecemeal, genuine evidence surfaced in the spaces we had hollowed out.