in this sense, the errors scale linearly with the scope of the task. Imitation learning is far, far worse, they discovered. Because a single mistake could cause the system to see things it had never prepared for before, once it makes a first mistake, all bets are off.

If intelligence is, as computer scientist John McCarthy famously said, “the computational part of the ability to achieve goals in the world,”88 then reinforcement learning offers a strikingly general toolbox for doing so.

At its most basic, reinforcement learning is a study of learning by trial and error, and the simplest algorithmic form this trial (or groping, if you prefer) takes is what’s called “epsilon-greedy”