In Gleick's book he mentions that, for instance, Feynman talked about inventing his own notation for various mathematical operations when he was young, and that the habit of thinking things through on his own, without reference to other people's approaches, persisted throughout his career. You see that in his famous Lectures: he expressed disappointment that his approach in the third book is the standard one, that he was not able to come up with his own, because he must not understand it well enough.