What if the surest way to fail at something ambitious is to have a clear plan to achieve it? What if a robot that doesn’t know it’s trying to walk learns faster than one explicitly...