1. They are not both fundamental, a mole is defined as the amount of substance that contains an equal number of elementary entities as there are atoms in 0.012kg of the isotope carbon-12. So the mole is defined in terms of the kg.
2. This is essentially Arnold Nicolaus' project (described in the article). He wants to define the kg as the mass of a certain number of atoms. But we don't yet have an accurate way of counting atoms, so if we pick a number (i.e. define Avogadro's number by fiat) today, then later when we count more carefully the mass of everything (denoted in kg) will fluctuate.
2) We don't exactly know Avogadro's constant. We've measured it fairly accurately, but, for your proposal to work, we'd be forced to define all 23 digits. We can then declare that 12g of carbon 12 is exactly 12g, but how do you know that you have exactly 12g of carbon twelve? Just from basic counting statistics, you'll probably be off by a few hundred billion atoms, and that's assuming that you have someone is and count all 6.02 * 10^23 atoms individually. The result is that the mass of the kilogram is now a derived and empirical value, adding uncertainty to every measurement of mass in every experiment. It's far preferable to leave the kilogram as an exactly defined unit and put the uncertainty into Avogadro's number, where it will only crop up when someone needs to make the conversion.
1. Why do we need both kilogram and mole as fundamental units when they measure essentially the same thing (at least in case when the thing being measured is pure)?
2. Why not reverse the definition: postulate that 1 mole is Avogadro's number of elementary particles; 1 mole of C-12 has 12 grams and derive (kilo)gram from that.