No I agree, it seems stupid. What we did is take a block of metal and call it a kilogram, figure out how many atoms are in 1/1000 of it (a gram) and call that a mole. What we should do is redefine the gram to be whatever 1e10 atoms of Hydrogen weighs, or something similar. When I read the headline I thought thats what they did, and I admit first reaction was "oh god I'm going to be dealing with conversions and associated errors for the rest of my life".
> What we should do is redefine the gram to be whatever 1e10 atoms of Hydrogen weighs, or something similar.
That was actually an alternative proposal for redefining the kg. The kg would have been 1000/28 the weight of a mole of silicon-28; you could build a sample by counting 6.023x10^26/28 atoms of silicon-28, and making a sphere out of them.
Initially the watt balance seemed to be less precise than atom counting, but then it was improved to a point where defining the kg on top of the mole became less convenient than the definition they are adopting now.
Disclaimer: I took high school chem, that's it.