... which is __the__ problem I was referring to, because these two implementations have essentially become black boxes with different performance on differnt ru

... which is __the__ problem I was referring to, because these two implementations have essentially become black boxes with different performance on differnt runtimes. Should all library writers detect runtime and choose the optimal one? I don't think so. They're far more likely to simply write `a+=b`, which is NOT quadratic without any knowledge of the underlying impl.