"[On writing the proof] It was like, connecting dots. Some nights, I, I could connect, three or four of them, and some nights they'd be really far apart, I'd have no idea how to get to the next one, if there was the next one."