This ignores that imaginary numbers (useful for both electronics and Pythagorean triples) were originally thought to be that kind of “abstract nonsense”. You h

This ignores that imaginary numbers (useful for both electronics and Pythagorean triples) were originally thought to be that kind of “abstract nonsense”. You have to think up a new ontological idea (eg, imaginary numbers to “complete” the solution space of polynomials) before you can apply that to modeling things, eg signal analysis in electronics. Category theory to some degree has this problem: the ideas are useful, but only once you’re at the level of “I know many areas of math and have noticed proofs have similarities… how do I discuss that?” …but it turns out that abstract problem is useful for computer science, eg design patterns.