I read Hadamard's book 8 years ago and liked it a lot.

What I missed is that I mistakenly thought that Poincare's style of thinking was reserved for supergeniuses, and that all that someone like me could do was to clumsily use explicit reasoning.

I found out otherwise when I worked on my speed dating project. Something very primal in me came out, and I worked on it almost involuntarily for ~90 hours a week for 12 weeks. I finally had the experience of becoming sufficiently deeply involved so that the problems that I was trying to solve percolated into my subconscious and my intuition took over. I rediscovered a large fraction of standard machine learning algorithms (it was faster than learning from books for me personally because of my learning disability). Before this, I had no idea how capable I was. It made me realize that being a great scientist might be within the reach of a much larger fraction of the population than I had thought.