The bullet train in Japan used to make a noise problem. Every time it shot out of a tunnel, the air it had compressed ahead of it released with a crack loud enough to disturb people living hundreds of metres from the track. The engineer who fixed it, Eiji Nakatsu, was a birdwatcher. He had noticed that a kingfisher dives from air into water — two media of very different density — almost without a splash. The bird's beak is shaped to part the water cleanly. So the train got a longer, beak-like nose. The noise dropped, and the train ran faster on less power.

That is biomimetics in one sentence: looking at what living things do and borrowing the principle, not the appearance. It sits naturally inside cognitive science because so much of what we copy from nature is not structure but behaviour — how an organism senses its world, makes decisions, and acts with very little information.

The classic example is Velcro, invented after a Swiss engineer looked under a microscope at the burrs stuck to his dog's fur and saw tiny hooks. But the more interesting cases are the ones that copy a process rather than a shape. Ant colonies find efficient paths between food and nest with no central planner; each ant follows a simple chemical rule, and the efficient route emerges from the group. That insight now sits inside routing algorithms for telephone networks and delivery fleets. Termite mounds in Africa stay at a steady temperature through passive airflow, and architects have used the same trick to build offices that need far less air conditioning.

What I find compelling about biomimetics is the humility built into it. It assumes the answer might already exist, refined over a few hundred million years of testing, and that the designer's job is to notice it. That is a different posture from the usual engineering instinct to build from scratch. It rewards observation over invention.

There is a catch worth naming. Nature optimises for survival and reproduction, not for our goals, and copying a biological solution without understanding why it works can mislead. A bird's wing and an aeroplane's wing both generate lift, but planes do not flap, because rigid wings plus engines suit our materials and speeds better. The skill is in separating the principle from the implementation — taking the kingfisher's gradient-crossing geometry without trying to build a train that hunts fish.

The field is quietly everywhere now: self-cleaning paint based on the lotus leaf, adhesives based on gecko feet, wind-turbine blades shaped like humpback-whale fins. The common thread is attention. Nature has been running the longest design experiment there is, and most of the results are sitting in plain sight.