From 2018 - no use correcting it now - but an instance of drawing conclusions about probability after ignoring the predicates. So, if one child is a son, what is the chance both children are sons: (1) after a daughter (1/2), the next is a son (=>1/2); (2) after a son (1/2), the next can be a daughter (=>1/4) or another son (=>1/4). The three allowed outcomes does not imply they are equally probable.
That fallacy continues in that article, or am I wrong, just like the whole world about that well-known vos Savant's answer to the Monty Hall conundrum. I had to write a little program before I believed that one.