»Imagine that you have risen through the ranks to the top tennis tournament in the world and you are now poised on a green court facing the planet’s greatest tennis robot. This robot has incredibly miniaturized components and self-repairing parts, and it runs on such optimized energy principles that it can consume three hundred grams of hydrocarbons and then leap all over the court like a mountain goat. Sounds like a formidable opponent, right? Welcome to Wimbledon—you’re playing against a human being.

The competitors at Wimbledon are rapid, efficient machines that play tennis shockingly well. They can track a ball travelling ninety miles per hour, move toward it rapidly, and orient a small surface to intersect its trajectory. And these professional tennis players do almost none of this consciously. In exactly the same way that you read letters on a page or change lanes, they rely entirely on their unconscious machinery. They are, for all practical purposes, robots. Indeed, when Ilie Nastase lost the Wimbledon final in 1976, he sullenly said of his winning opponent, Björn Borg, “He’s a robot from outer space.” But these robots are trained by conscious minds. An aspiring tennis player does not have to know anything about building robotics (that was taken care of by evolution). Rather, the challenge is to program the robotics. In this case, the challenge is to program the machinery to devote its flexible computational resources to rapidly and accurately volleying a fuzzy yellow ball over a short net. And this is where consciousness plays a role. Conscious parts of the brain train other parts of the neural machinery, establishing the goals and allocating the resources. “Grip the racket lower when you swing,” the coach says, and the young player mumbles that to herself. She practices her swing over and over, thousands of times, each time setting as her endpoint the goal of smashing the ball directly into the other quadrant. As she serves, again and again, the robotic system makes tiny adjustments across a network of innumerable synaptic connections. Her coach gives feedback which she needs to hear and understand consciously. And she continually incorporates the instructions (“Straighten your wrist. Step into the swing.”) into the training of the robot until the movements become so ingrained as to no longer be accessible. Consciousness is the long-term planner, the CEO of the company, while most of the day-to-day operations are run by all those parts of her brain to which she has no access. Imagine a CEO who has inherited a giant blue-chip company: he has some influence, but he is also coming into a situation that has already been evolving for a long time before he got there. His job is to define a vision and make long-term plans for the company, insofar as the technology of the company is able to support his policies. This is what consciousness does: it sets the goals, and the rest of the system learns how to meet them. You may not be a professional tennis player, but you’ve been through this process if you ever learned to ride a bicycle. The first time you got on, you wobbled and crashed and tried desperately to figure it out. Your conscious mind was heavily involved. Eventually, after an adult guided the bicycle along, you became able to ride on your own. After some time, the skill became like a reflex. It became automatized. It became just like reading and speaking your language, or tying your shoes, or recognizing your father’s walk. The details became no longer conscious and no longer accessible.«