LIVE READING / ROLLING 24 HOURS
STATUS: INCONCLUSIVE
Collecting enough signal to make an observation.
FIELD INSTRUMENT / PUBLIC SURFACE 01
The public web looks busy. That does not tell us how many people are still speaking.
What is measured here: requests reaching one hostname, not the authorship of the internet.
LIVE READING / ROLLING 24 HOURS
STATUS: INCONCLUSIVE
Collecting enough signal to make an observation.
01THE CLAIM
The web can look alive while fewer people are speaking.
Bots can fetch, index, imitate and amplify. Humans can browse behind automation. A request proves that software reached this page; it does not prove who intended it.
The dead internet hypothesis cannot be settled by atmosphere or traffic volume. It needs scoped observations, durable records and explicit uncertainty.
Requests, timestamps and the Cloudflare edge that served them.
A likely request class from edge signals and crawler identifiers.
The person, organization or system that ultimately caused the request.
02THE OBSERVATION
One hostname. Approximately 24 hours. Requests, not people. Every classification is an estimate.
LIVE FROM CLOUDFLARE / THIS DOMAIN / WEBSOCKET EVENT STREAM
| UTC | CLASS | EDGE | BASIS |
|---|
Cloudflare edge signals and known crawler identifiers label requests as human-likely, automated-likely, AI-associated, verified crawler or unclassified. A label is not proof of identity.
Counts include GET requests to this homepage. Reloads and automated probes may appear; assets, API calls and missing paths do not. This is not a representative sample of the internet and cannot answer the question above.
03THE STANDARD
This project treats the dead internet theory as a falsifiable claim, not a mood.
Observations must declare their scope. Classifications must expose uncertainty. Authored testimony must retain a reproducible identity even if the interface around it changes.