Way back when, I did a small study (4 cyclists, 5 reps). The question: Could the rider tell which was heavier between a water bottle filled with lead shot, and one with glass beads. (2.5kg vs ~ 1kg).
1 rider managed to get more than half right. And he was a frame builder riding his own bike.
And there's nothing wrong with having optimized code. This seems to be a discussion about having the right priorities, not on the virtues of the optimizations themselves.
1 rider managed to get more than half right. And he was a frame builder riding his own bike.