Alán Aspuru-Guzik, a theoretical chemist at Harvard, is developing methods that could one day do away with the need for approximations altogether--and lead to better drugs or solar cells.