In the excellent article on Mersenne primes
http://www.utm.edu/research/primes/mersenne/
I found the following, to me, cryptic passage
"Are there infinitely many Mersenne primes? The answer is
probably[ital] yes (because the harmonic series diverges). Could
someone please explain the "because"? Does the same argument
work for Fermat primes? for primes of the form 10^n + 3? for (a^ n)+b,
gcd[a,b]=1?