I think academia is right to distrust "teaching to the fad" because concepts don't change, and people who know the concepts will be good at picking up the fads. Math -- to a lesser or greater extent, it's arguable -- is one of the concepts.

Analogy: if you'll be programming in Java and Python, why learn C or assembly? Don't "tough" concepts like pointers and managing your own memory turn off a lot of potentially great programmers? Well ... apparently they wouldn't have been so great. Without these concepts, you think you're fine bubbling along in Java till one day you try reading in a 10-MB file by using String str = str + nextLine() and don't understand why it's taking forever.

So -- this is also about standards. What standards a student has for herself, and what standards a school wants to set for its students. You think most of the programmers hired at Google don't know calculus and linear algebra? It's not about whether you integrate functions on a daily basis, it's that (1) you've proven you're smart and those smarts often correlate with good coding, and (2) you have additional tools and a better understanding in many cases. In industry, that pays.

And to be blunt, if someone told me they hadn't had calc or probability, I would distrust their overall understanding of the field of "computer science". Nothing against their programming skills, it's just that a lot of what I call "CS" includes mathematical concepts, and I think that a CS degree should include them.