"of which we know at least one is a boy" limits the possibilities to the sample of families with at least one boy.
I don't think we can tell from that statement whether the family was chosen from a sample of families containing 2 children of which one was a boy (1/3) or if the family was chosen from a sample of families containing 2 children (and the statement about the sex was made afterwards based on that choice) (1/2).
I don't think we can tell from that statement whether the family was chosen from a sample of families containing 2 children of which one was a boy (1/3) or if the family was chosen from a sample of families containing 2 children (and the statement about the sex was made afterwards based on that choice) (1/2).