> Shouldn't entirely new life develop all the time on earth?

The fact that it doesn't is a pretty simple solution to the Fermi Paradox.