The Algorithm as Magic System

Divination, Diffusion Models, and the For You Page

An essay arguing that divination systems (I Ching, Tarot, Fal-e Hafez), AI image generation (diffusion models), and the TikTok For You Page all run on the same structural engine: focused intention applied to a noise field, with the pattern-seeking mind doing the actual work. The oracle does not predict the future. It reflects the present back at you in a form you can work with. So does TikTok.

What the Essay Covers

The essay opens with a personal anecdote: casting the I Ching and receiving hexagram 36 (Darkening of the Light), then having TikTok serve a thematically identical video the same evening. It then makes a structural argument, not a mystical one, showing that every divination system pairs the same three elements: a noise field, an intention, and a filtering mechanism. The same elements power diffusion models and recommendation algorithms. The technical term for what the mind is doing — finding meaningful patterns in random data — is apophenia. It is not a bug in these systems. It is the runtime environment.

The essay also names the differences honestly: the I Ching has a lineage of three thousand years; a diffusion model has the biases of its creators baked in; the TikTok algorithm has an owner optimising for attention, not clarity. The oracle has no shareholder value to maximise.

Key Argument

Focused randomness is the mechanism of magic. This was true when the first shaman threw bones on the ground. It is true when a prompt engineer types an intention into a latent space. It is true every time you refresh a feed with something unresolved in your chest. You are already doing magic. The only difference is whether you call it a spell or a recommendation.

Explore

By Alex Vizcroft. Also published on Medium.