(In reply to some observations...
by Erik O.)
Your first observation is correct, all the numbers in the sequence are odd.
Your fourth (multiples of 3) is a consequence of the algorithm I used to generate the sequence.
Your predictions of 1000-3000, 100-500 for the next two numbers are correct.