With the rename to
argmax, and dims now a kwarg, it seems very tempting to me to define a 2-arg flavor, argmax(f, domain).
This would be equivalent to x[argmax(f(x) for x=x)] but argmax(f, x) is just so nice. Here, x could be any iterable and argmax(f, x) would quite literally implement the mathematical definition of argmax. Has this been considered?