Big-O is a promise, not a stopwatch

$O(n)$ does not mean fast. It means the running time grows at most linearly as $n$ grows.

A $O(n)$ algorithm can lose to a $O(n^2)$ one on small inputs. Constants matter.