I said nothing about matching engineering's gender distribution to the user base's. Please don't put words in my mouth.
The thread-parent's comment made it clear that, in his situation, it made things better. Is it going to in every case? No. But using edge cases to argue against the median is even more specious than the argument you assert I was trying to sneak in.
Because if you're ignoring (by simply not hearing) the perspective of a meaningful representation of that ~half of your user-base, you do not, and probably can not understand them.
You don't need to have parity between those ratios, but you might want to do better than the "token diversity hire" — assuming you do at all.
The thread-parent didn't make that clear. Rather, he defined success as replacing men with women and then said he'd done that, therefore, it was a success.
His post tells us nothing at all about what impact that had on the quality of the resulting software.
The thread-parent's comment made it clear that, in his situation, it made things better. Is it going to in every case? No. But using edge cases to argue against the median is even more specious than the argument you assert I was trying to sneak in.