Starting from static text, I used AI to help write generative rules that turn letterforms into dynamic systems of dots. Letter outlines decide where dots gather; noise and motion rules push them off course. By adjusting dot size, density, range of motion and spatial depth, I watched for when the text stays readable and when it starts to become an abstract image. The experiments also extend from two-dimensional dot matrices into three-dimensional point clouds, letting me look further at how viewing angle changes legibility.
The process lets me treat accidents in generation as design material. When a shift produces an effect I hadn’t anticipated, I go back to the code to find the conditions behind it, then decide how to keep or strengthen it, developing a chance result into a visual language that can keep being used.