> All that said, I agree with Mark that math.e is at best an attractive nuisance.

Why don't we fix the nuisance part without making it less attractive:

class _E(float):
    def __pow__(self, other):
        if self is e:
            return exp(other)
        else:
            return pow(self, other)

e = _E(2.718281828...)

> It's okay if Python occasionally shows its lighter side in unexpected places.

I would rather see

from math import π

work in the future Pythons.

I like Vi Hart, but I am firmly on the π side in the π vs. τ debate.  The problem with τ is that it is visually smaller than π, actually twice smaller: π ≈ ττ, but the actual definition is the opposite.