How it works — stigmergy. The colony's knowledge lives entirely in the evaporating floor, not in any ant.
A
searching ant lays a
home trail (freshest by the nest) and steers toward
food scent.
On contact it picks up a crumb, flips to
carrying, lays a
food trail (freshest by the crumb) and steers toward
home scent.
Each ant only reads three sensors a step ahead —
left · center · right — and turns toward the strongest. Deposit strength
decays with every step from the source, so the two trails become opposing gradients the colony can climb in either direction.
Tie to Neuron Lab. Same rule as the project's signed pheromone floor:
local and earned, ant-style evaporation each tick, deposits scaled by what an ant actually carries — never a broadcast coordinate or a global beacon. The colony disperses while searching, then converges on the first trail that pays off.
This is the pure-emergence version. Its sibling,
Swarm Convergence on a Maze, adds
quality control — success-scaled deposits and authority gating — so the colony doesn't just find a route, it shortens and certifies it.