MECHANISMS IN MOTION

Watch It Compute

Six compute functions, shown not told

The Catalogue names the kind of computation at work in each system. Some of those functions are easier to see than to define. Below are six of the most common — each an abstract sketch of the mechanism itself, not the thing it runs in. Click any panel to run it.

Feedback control set-point click to run

Feedback Control

e.g. a thermostat, sweating, a beating heart

A system measures the gap between where it is and where it should be, then corrects — overshooting, sensing the new error, correcting again, until it settles on the set-point. The wobble is the computation: each correction is a reading of its own last result.

Path optimization start goal click to run

Path Optimization

e.g. a river, an ant colony, lightning

Many routes are tried, faintly; the cheaper ones are reinforced and the wasteful ones fade, until a single efficient path remains. No router plans it — the path is computed by the trials themselves, the good ones surviving.

Equilibrium-seeking unbalanced click to run

Equilibrium-Seeking

e.g. a star, acid meeting base, a see-saw

Opposing pressures push until they cancel. The system swings, overshoots, swings back with less force each time, and converges on the point where the forces balance. That resting point is the answer it was computing.

Threshold / trigger threshold charging… click to run

Threshold / Trigger

e.g. a neuron, a volcano, a dividing cell

Something accumulates quietly — charge, pressure, signal — doing nothing visible until it crosses a line. Then, all at once, it fires. The computation is a comparison run continuously: is the total over the limit yet?

Emergent aggregation click to run

Emergent Aggregation

e.g. a starling murmuration, a school of fish, a city

Each agent follows only its neighbours — match their heading, keep close, don't collide. No one directs the whole, yet a single coherent shape forms and moves. The pattern is computed by everyone at once, and by no one in particular.

Energy minimization released high click to run

Energy Minimization

e.g. a folding protein, a water droplet, a snowflake

Left alone, a system rolls toward its lowest-energy state, the way a ball settles into the deepest reachable valley. The final shape isn't designed — it's whatever costs the least to hold. Finding that low point is the computation.

Six sketches, one lesson: in every case the answer is not looked up or handed down — it is arrived at, by the process running itself forward. That arriving is what the Autoverse means by computation.

END OF MECHANISMS
Return to the Book