A world model learns to predict (state, action) → next state. Google's Genie 3 uses this at massive scale to generate interactive worlds from images. This demo shows the core challenge: prediction errors compound. Two balls start identical — one follow...

Generative asemic calligraphy. Writing that looks like writing but carries no semantic content. Four procedural scripts — Cursive, Angular, Gestural, Micro — each with distinct stroke dynamics, rendered with simulated brush pressure on ruled pages. An ...