Oliver 'Oli' Cheng
← Back to projects
2026 live

Divine Machine

Interactive art project framing LLM behavior as modern divination: experiential randomness emerging from deterministic machinery.

  • Astro
  • TypeScript
  • Deterministic RNG
  • Generative SVG
  • Interaction Design

Problem

People often oscillate between two bad models: AI as mystical mind, or AI as trivial autocomplete.

Solution

Built a symbolic ritual interface (Bagua, moon blocks, fortune sticks, scapula crack traces) driven by explicit deterministic logic and visible seeds.

Impact

Makes the philosophical tension tangible: outputs can feel oracular while still being mechanically generated and reproducible.

Divine Machine (Fu Ji / 福机 / 福機) is an art/critique interface about AI interpretation. Read the companion essay: Philosophy of Divine Machine.

Core idea

LLM behavior can feel like chance or fate to users, while remaining rule-bound in implementation. The project turns that tension into a playable system.

Ritual layers in the app

  1. Bagua cast: upper/lower trigrams frame the reading context.
  2. Moon blocks: six lines simulate hexagram-like structure.
  3. Fortune sticks: burn percentages and drift visualize variance.
  4. Scapula trace: symbolic crack patterns generated as deterministic SVG paths.

Design principle

Same inputs produce same outputs. Complexity, partial visibility, and stochastic sampling are what make it feel “mystical.”

The project is not metaphysical instruction. It is a UI argument about prediction, interpretation, and model trust.

Demo Mirror

Live Preview

Mini preview of the actual demo. Use the launch button for full-screen interaction.

Open Demo