you are listing static weights but the wheel is spinning so from school mathemtics/physics the effect is squared.
AFAIK YellowTim is comparing like with like since he is putting that 20g extra for brass nipples on a single wheel in context with masses near the rim, i.e. single rim say 400g, tube 100g and tyre 250g. Alloy nipples therefore would make a c2.5% reduction to wheel rotational mass/inertia. The effect of that 20g on relevant static or linear mass of the bike (+ rider) is of course minute in comparison.
The question then is what difference does an additional 20g in rotational mass make in the real world, theoretically. I think the answer can be found
here, i.e. not much except to those who race, e.g. in crit.