Can we convert static inspirations into living systems?

Hypothesis

We collect visual inspiration for years and never touch it again. Can an AI take a static image from a mood board and decompose it into an interactive, living system — turning archived taste into something you can play?

What we built

A generative art piece that extracts a random image from a Pinterest board and decomposes it into interactive digital states. Physics simulations (Hooke's-law spring networks), Lévy-flight random walkers, dynamic color-displacement glitches, and vector flow fields reconstruct and transform the artwork in real time. An interactive Web Audio synthesizer maps the pixel values and coordinates under your cursor to frequency sweeps — you hear the image as you move across it.

Learnings

• Simple physics on top of real pixel data beats elaborate shaders: the spring network alone makes a static image feel alive.
• Sonification is the sleeper feature — mapping cursor position to synthesis parameters makes people linger far longer than the visuals alone.
• "Take something you already own (a mood board) and animate it" is a repeatable experiment recipe, not a one-off.

Log

• 2026-07-10 — Rebuilt as an embedded static site at /experiments/pinterest-art.
• 2026-07-08 — First version: image extraction + spring-network decomposition.