This is a good point. When people talk about the "usefulness" of math/science, they tend to assume that learning math/science is analogous to being trained in h

This is a good point. When people talk about the "usefulness" of math/science, they tend to assume that learning math/science is analogous to being trained in how to operate a forklift: it's thought of as a "skill" that's useless if one never needs to directly apply the content. But learning math/science encourages certain cognitive habits which are useful in practice. For most people, I expect that these cognitive habits are where the value of a math education lies. As an example, in math, you often need to define a new object out of thin air, and then think about whether the definition is useful (e.g., let u=log(x), then rewrite the integral in terms of u and see if it gets you anywhere). I hypothesize that even after you've forgotten the math content, this cognitive habit (constructing a new mental object ex nihilo, thinking about whether it's useful, and perhaps giving it a name if it is useful) sticks with you, and contributes to general intelligence in other contexts.