[ozyschmozy]

Well in this case you weren't wrong about the first 10^10 numbers:

> the smallest counterexample is n = 906,150,257, found by Minoru Tanaka in 1980.[5]