Sol LeWitt argued that the idea is the machine that makes the art. His wall drawings — sets of plain instructions executed by others — produced complex, often infinite-feeling work without his hand ever touching the wall.

Neil Gershenfeld and the MIT Center for Bits and Atoms treat the voxel as the atomic, programmable unit of physical matter — small, identical building blocks that self-assemble and self-replicate into larger structures.

LeWitt Hills merges both. You compose LeWitt-style instructions: extend lines, fill planes, mirror, recurse. A rule engine turns those instructions into voxels. A lightweight physics layer lets the voxels fall, stack, and settle like real matter. Switch on Autonomous Growth and the system keeps proposing voxels on its own — the drawing becomes a machine that draws itself.

Every parameter is in the URL. Share a link to share an instruction. Export to JSON for the recipe, or read the human-language version on the right — "Wall Drawing #" — exactly how LeWitt would have written it.