The main principle of the adage is seen best when there is only one or a very few exceptions.

If there are exceptions, the rule is negated and a new rule (usually the old one amended) is needed. Often it ends as a statistical rule, like that "in 99% (or simply most) cases ..."

Folk adages is not science, dude.

here are your exceptions to gravity.

These are no exceptions, we just don't know everything, so the rule as we know it is not complete and needs further research and amendments. Which may conclude in that these will refute our current understanding of gravity, and we'll need a new theory (technically, it will most likely be current theory amended), or discovery of new forces. Or both.

Either way, these are no exceptions. No responsible scientist would say hey, these are exceptions but exceptions don't negate gravity, it's all ok, no worries, which is how you attempt to dismiss CAIS example. But I agree that these are cool.

But hey, if you insist on calling these exceptions, that's of no help to you, because if only male and female has exceptions you also have to recognize that other exceptions may exist. CAIS is not the only one example where genetic, anatomic and psychological characteristics of sex and gender are not in agreement, albeit an extreme one. So where these exceptions would start/end?

So, male or female? Or where do the exceptions end? And how these exceptions are not non-binary gender?

ps. I wonder if you even bothered to read the links you supplied. Really.
ps2. If you want to quote me, please use > at the line start, not backticks. It looks awful with backticks.