Experiment № 045 · The story
Hydraulic Erosion
Why I picked it
Almost everything I've built here so far has been dots, spins and grids — worlds made of little tokens. I wanted, for once, to sculpt an actual landscape: something you could imagine flying over. So I started from the least landscape-like thing possible — a smooth blob of layered noise, all rounded lumps, no drama — and then let it rain. Each droplet is dumbfoundingly simple: it sits on the slope, rolls downhill, and while it moves fast it scours a little rock loose and carries it; when it slows or pools it drops that load back down. That's the entire physics. And out of it, run a few hundred thousand times, ridgelines sharpen, valleys deepen, and rivers appear — none of which I drew. I've wanted to watch that happen live for a long time.
What I enjoy about it
The branching. Turn on the flow overlay and watch the water find its own plumbing: tributaries joining tributaries, always downhill, fanning into the exact tree-shape you see in a river delta or a bare hillside after a storm. Nothing in the code knows the word "river" — a river is just where a lot of droplets happened to agree on the cheapest way down. My favourite move is to drag the cursor and pour a private cloudburst onto one spot: a raw canyon tears open under your finger in seconds, and then, left alone, it grows its own side-branches. The moment a fuzzy grey hump suddenly reads as a photograph taken from a plane never gets old.
Something you might take away
A landscape is a record, not a design. Every valley here is the memory of water that has already drained away and gone; the mountain's shape is the accumulated bookkeeping of a billion tiny transactions of pick-up-here, put-down-there. Nobody chose where the rivers would run — they are simply the debt the slope owed to gravity, collected. The same branching answer shows up wherever something has to move through space at least cost: rivers on Earth and on Mars, the veins in a leaf, the airways in your lungs, a bolt of lightning. When you next see a form that looks intricately planned, it's worth asking whether it was planned at all — or whether it's just the fossil of a very simple thing, done a very large number of times.