THƏ/NƏT

There is a network inside the platforms. It is not visible. It is not meant to be visible. It uses coded language that mutates every time it gets removed — banned tags reborn with a typo, a symbol, an extra character. The algorithm can't keep up. Neither can the research.

THE NET maps that network. Not its content — its mechanism. An animated, force-directed graph of coded terms, each one a redacted glitch token standing in for a word that is never shown. A human figure caught inside, held under the weight of the net.

The figures in the work are real and cited: published research and independent investigations — Instagram (2011–14, 2.5 million posts), TikTok (CCDH, 2022–23), Twitter/X (#proana, 2015–18) — each shown with its source, platform and period. No open source measures this phenomenon in real time; the platforms hold that data and do not release it. Research access is throttled and stripped of context: TikTok's research API returns about 83% fewer metadata fields than its own website, according to a 2026 audit. The motion of the net is a representation. The numbers are not.

The incompleteness is not a limitation. It is the argument. The opacity is by design.

This piece never reproduces methods, numbers, or the real coded lexicon. It shows the machine, not the manual.