Or when optimization extrapolation fails.
Or the cost of non-final decisions.
Or avoiding dynamism for optimal performance.
Or predicting future in C++.
Or broken zero abstraction promise.
Or don't let work go to waste.
Or one more reason to move away from raw loops.
Or be careful with floats.
Or true cost of optional abstraction.
Or doing more at compile time.
Or how not to double the size of the pointer.
Or caching hashing.