If you're going to say that you need to study math exclusively for many years to understand your formulas then you are not using abstraction well.
Once again, my point is that people are trying to take shortcuts with abstractions that are not grounded in reality. That is a matter of self-discipline, of priorities, of putting the cart before the horse. Consider string theories: we have worked out so many ways in which strings can behave, etc. with so many possibilities and permutations. However, we never proved the ground reality for strings, we just ran with a bunch of assumptions and then parameterized them, went meta a bunch of times, and called that a research program.
All of that mathematical sophistication and model-building could have went to, e.g. perfecting QCD, or even in other directions.