Point taken.
It uses the US Social Security Administration's list of baby names from 2012 and does a weighted random selection, so that more popular baby names chosen in 2012 are more likely to be chosen (but less popular ones still come through pretty often.)
The gender-neutral names work similarly, by finding the intersection of boy/girl names in that dataset and weights the names by the popularity of the name and the evenness of the boy/girl distribution.