Aligning One Matrix with Another / hacker news