Agreed.

Your comment is mostly true, when you do binary search in something like an array, yes.

But you can also do binary search in any monotonically increasing function.

For arbitrary functions you usually want floating point, which solves the overflow problem in a different way.