Building data models is often hard, but some are harder than others. Gender is a particulary tricky thing to model for a variety of reasons: it's complicated, our understanding is changing, and, yet, our knowledge is uncertain. Working out how to model gender touches on deep philosophical issues about how we determine what is correct, what is good and what is not.
It's also difficult for another reason. It's uncommon when discussing data models to consider the possibility that they might be personally offensive and deeply political. But this is entirely the case when discussing gender; consider the enormous amount of political debate in recent years over where people should urinate. This form of political sensitivity raises some interesting modelling issues, including the choice between two semantically identical models, one of which may be offensive, the other not.
I will probably talk about pizza as well. It is also a contentious topic and not about gonads.