Depends on the application. A lot of stuff, especially in CRUD world, doesn't need any. A lot of stuff really leans hard on it. The least terrible one sentence

Depends on the application. A lot of stuff, especially in CRUD world, doesn't need any. A lot of stuff really leans hard on it. The least terrible one sentence rule of thumb is that the more SIMD could help make your code faster, the better at math you should be. It's a really weird rule, but for some reason it works. Specifically regarding the generalized good at math, therefore good at problem solving, don't think too much into it. There's a lot of correlation/causation problems in there. A BS in mathematics will not make you a better programmer than a BS in computer science will. If you're going to do four years in college anyway, sure, pick up a few extra math classes. (especially linear algebra) If you're going the self taught route, (I couldn't make the transition without the piece of paper, ymmv) watch the mit ocw (or from whatever other college) series for the first semester of linear algebra, discrete mathematics, and calculus. You don't need the second semester. I do GIS, which is a lot of spatial modeling. Linear algebra is an everyday thing for a lot of us. Many others don't need any, because they deal with networking, databases, static UI, etc. Don't think you'll be locked out of a certain field just because you don't know the required math, but also know that knowing the right math will grease the wheels.