I think Scott Alexander is a terrible example.

He wrote a rebuttal.

There are several messages that would be good to send to lots of calculus students. One is that math is diverse. Another is that classes are twisted by the focus on evaluation and grading. Calculus is particularly bad: no one computes closed form integrals -- not pure mathematicians, not engineers.

But Scott is not a calculus student. He has encountered math in many contexts outside of math class and outside of school. He needs statistics. Telling him that the statistics that he grapples with is "not real math" is not helpful. And he really does grapple with it. There are people who refuse to deal with statistics or just go through the motions without understanding. Perhaps they could unlearn their helplessness by reevaluating their belief that they are bad at math. Or perhaps that would be bad for their careers. In any event, Scott is not one of them.

Perhaps there are people who struggle in calculus class and then turn into research mathematicians. I'm skeptical that the skill needed to make it through badly taught calculus class is not also very important to doing "real math." Maybe more people should be exposed to "real math." But Scott has been exposed to "real math," as detailed in that essay.