“Our knowledge advances not when we have the best theory, or the best data, or the best lab, but when we have really good problems”