That was always a big argument when I was in and around a university computer science department: how much time you spend on "general enlightenment" topics like linear algebra and how much to spend on specific, valuable skills, like x86 assembly.