> I am referring to the way many on the far left speak of intersectionality as a number of diversity or victimhood points to be tallied.

IME, to the extent that occurs, that's not at all typical of the far left, but (to the extent it happens at all), the weakly-ideology tribal-identity left.

> Here is an example. Being male is not a privilege for a young black man being stopped by a police officer in the United States.

It actually is in some ways; the ways gender interacts with race in that scenario is complex.

> On the contrary, it greatly increases the odds he will soon be unjustly killed.

It greatly increases the chance of being unjustly subject to non-sexual violence by the police; OTOH, it very much seems to (though numbers are much harder to come by) greatly decrease the chance of being subject to sexual violence by the police. Now, police sexual assault gets less attention (for many of the same reasons sexual assault more generally doesn't, regarding publicity being retraumatizing for victims, etc.), But it's a serious issue, with significant racial aspects, too.

Most intersectionality can be seen as simplistic statistical modeling that has flaws (such as male not being a benefit whenever the legal system is involved). I don't actually think the theory is wrong, but I do think the current models are way too simple and more often than not picked not for their correctness but their usefulness in achieving political goals.