Recently, Kai-Liang brought to my attention an extension form of the classic inequality of the arithmetic and geometric means.

It is easy to see C2=1. It is not easy to calculate C3. I spent a couple of days on this problem and found the solution.

My method is very standard and hence very DULL, unfortunately.

Anyway, I copied here in case anyone can propose a better method and provide solutions to other cases.