Hod Lipson, a mechanical engineer who directs the Creative Lab Machines at Columbia University, has shaped most of his career around what some people in his industry have called the word C.
On a sunny morning last October, the robotics born Israeli sat behind a table in his laboratory and explained himself. “This subject was taboo,” he said, a smile exposing a slight gap between his front teeth. “We were almost prohibited from talking about it -” don’t talk about the word C; You will not get the mandate – so at the beginning, I had to disguise it, as if it was something else.
It was back in the early 2000s, when Dr. Lipson was a deputy professor at Cornell University. He worked to create machines that could note when something was wrong with their own equipment – a broken part or a defective wiring – then modify their behavior to compensate for this deficiency without the guidance hand of a programmer. Just like when a dog loses one leg in an accident, he can learn to walk again in a different way.
This type of integrated adaptability, supported Dr. Lipson, would become more important as we become more dependent on the machines. Robots were used for surgical procedures, food manufacturing and transport; The applications for machines seemed almost endless, and any error in their operation, because they integrated into our lives, could spell a disaster. “We will literally abandon our life to a robot,” he said. “You want these machines to be resilient.”
One way to do it was to be inspired by nature. Animals, and in particular humans, are good for adapting to changes. This capacity could be the result of millions of years of evolution, because resilience in response to injuries and changing environments generally increases the chances that an animal will survive and reproduce. Dr. Lipson wondered if he could reproduce this type of natural selection in his code, creating a generalizable form of intelligence that could learn more about his body and his function, no matter what this body looked like, and what whatever this function.