Moravec's Paradox

Moravec's Paradox is a surprising observation about robots and computers: the things that feel hard to us — chess, arithmetic, logic puzzles — turn out to be easy for machines, while the things that feel effortless to us — walking, grabbing a cup, recognizing a face, keeping your balance — turn out to be brutally hard for machines. In short, what is easy for a human is often hard for a robot, and what is hard for a human is often easy for a robot. The name comes from the roboticist Hans Moravec, who pointed this out in the 1980s.

Why does this flip happen? The skills that feel automatic to you — seeing, moving, touching — are the ones evolution spent hundreds of millions of years polishing, so they run beneath your awareness and seem free. Deliberate reasoning, like doing long division, is a newer trick that we have to think about consciously, which is exactly why it feels difficult — and also why it is easier to spell out as step-by-step rules a computer can follow. A machine can be handed those rules directly, but it cannot inherit the vast, hidden know-how packed into a body that has learned to move. This is a big reason computers could beat people at chess long before any robot could reliably tidy a kitchen, and it still shapes which jobs are hard to automate today.