set the internet to: 1992 1997 1998 2001 2002 2003 2006 2007 2012 2017 2020 2022 2026

home

The Clairvoyance of TikTok

An interactive 2022 For You feed about how recommendation signals, psychic clickbait, and constant connection can make a statistical loop feel like private revelation.

read the original

The Feed Opens

The first stack is broad: a sleepy cat, a thirty-second workout, a weeknight recipe. TikTok’s own explanation says a generalized feed becomes more personal through interactions, with a completed longer watch carrying more weight than some weaker signals. This reconstruction turns that mechanism into a visible toy model. Nothing has been read from your device.

Finish the cat video
Completion counts as interest in this simulation. The next stack leans toward animals and soft comedy. That shift describes a viewing signal, not your personality.
Swipe past the workout
A quick swipe lowers fitness in the toy ranking, so the feed tests an unrelated clip. One skipped video carries no hidden verdict.
Replay the recipe
The replay gives cooking extra weight. More recipes move upward, illustrating how ordinary behavior can rapidly make a feed feel uncannily specific.

This Video Found You

A tarot reader appears without hashtags and says the encounter was destined. I identify this as an engagement tactic that borrows the language of signs and omens. The clip is broad enough to fit many lives, but its arrival inside an already personalized feed can make coincidence feel addressed to one person.

Watch to the last card
The feed queues astrology, reconciliation, and another reading. Completion amplified a content category; it did not validate the prediction.
Replay the final sentence
Replaying adds more weight in the simulation. Similar readings arrive sooner, making repetition look like confirmation even though the loop began with your watch behavior.
Swipe immediately
The next cards turn toward recipes and comedy. The omen disappears because the ranking changed, revealing how contingent its apparent arrival was.

Feedback Becomes Fate

After two long watches, the stack fills with twin-flame promises and videos debunking them. I note that skeptical psychics formed a countergenre to the clickbait readings. Both attraction and rebuttal can keep the topic present. The feed seems to be speaking, yet the visible mechanism is a loop between recommendation and response.

Comment: “This found me”
The comment is treated as engagement, so the cluster tightens. A public declaration can help produce the pattern it appears merely to recognize.
Watch the debunking clip
The feed supplies more skeptical readings. The stance changes, but the subject remains because attention can reinforce a topic without revealing agreement.
Mark the topic uninteresting
The toy ranking reduces the cluster and restores variety. An explicit rejection is modeled as feed curation, not spiritual disproof.

Check the Material World

Now several health clips repeat the same concern. I use this kind of sequence to show how people may read recommendation as warning. Repetition inside a ranked feed is evidence that a topic is circulating and retaining attention; it is not medical evidence about the viewer. A real concern belongs with an independent, appropriate source.

Search inside the feed
More related clips appear, increasing familiarity without independently checking the claim. The recommender is still responding to the same topic signal.
Use an independent health source
The feed receives no new signal in this simulation. The concern moves into a setting designed for evidence rather than engagement.
Ask what the algorithm “knows”
The toy cannot know a diagnosis or private destiny. It only ranks content from the actions exposed in this exercise.

Close the Curtain

I contrast the persistent phone with an earlier ritual: entering a computer room, waiting for dial-up, and eventually leaving. I relay Clinton Ignatov’s effort to make computing deliberate through an offline desktop, books, and Encarta. The point is a boundary the user can feel—a curtain that opens and closes before the feed becomes an ambient oracle.

Close the app now
The stack freezes. No next recommendation is generated, and the abrupt boundary makes the accumulated loop easier to see.
Open the article externally
The simulation leaves the feed and opens a deliberate reading context. The source replaces the personalized sequence as the object of attention.
Scroll one more time
Another tailored card arrives, produced from the prior trail. Persistence is the final mechanism in my essay: the user never quite logs off.