I think the trick is to define the kg in terms of something that is not only a universal constant but can also be practically measured. If I redefine the gram

I think the trick is to define the kg in terms of something that is not only a universal constant but can also be practically measured. If I redefine the gram as the mass of 6e23 atoms of carbon-12, and you give me a big pile of carbon-12 and say "what's the mass?" and I can't count exactly how many atoms are in the pile, then my definition isn't very useful.