Actually, it does negate the existence of two discrete categories of people that you can ascribe intrinsic biological differences to. You'll always have to delineate the categories at some arbitrary point, and since the delineation is arbitrary, you'll therefore never be sure if your categories are correct.