(Disclaimer: overly shallow/un-nuanced post about to follow because deep and nuanced would be longer than an HN discussion thread or two)
There are a number of reasons those topics are taboo and, to some degree, maybe should be until further notice. For example, you say "reality is what it is and all the words in the world won't change it" but when the reality in question is the words, actions, and beliefs of people we all know that that statement is categorically false. The wrong words, the wrong discussions, the legitimisation of wrong arguments can all have a disproportionate effect and impede or even reverse some of the progress we've made. For example, it may or may not be the case that the average XY male has an inherently slightly higher propensity for physical strength as compared to the average XX female but any time a study suggests "yes" or "no" the discussions surrounding it greatly exaggerate the potential difference discussed. Especially with the way science is reported today, proof positive of a real difference that is still dwarfed by training would enter the public consciousness as an insurmountable gap that some would take to justify various forms of oppressing women.
Of course, we are talking about science and some people hold the ridiculous belief that the human element doesn't matter and we should just discuss the truth and thus everything above is irrelevant. Except, even if we legitimize that point of view, the truth is: we don't know. The state of the research on real, meaningful, inherent differences is largely inconclusive except for some highly specialized and specific results. For example, it's common knowledge that the best distance runners are disproportionately often Kenyan but this doesn't take into account the relative effect from genetics, environment, and training and it also doesn't imply that being Kenyan makes you better at running. When we look at it we can't say there's a difference between races or nationalities because we don't know that; we just know that some combination of factors has balanced the scales in an interesting fashion. With the exception of certain disease/condition frequencies, this is some of the best data we have for real differences between races or sexes. Generally speaking, even if we could do a study to compare everybody alive now from group A with everybody alive now from group B (say all XY-males with all XX-females) and properly accounted for other variables and measured a real difference we still have no way of knowing where the differences come from as far as "nature" or "nurture" goes.
Furthermore, there are adequate studies suggesting that, regardless of natural tendencies, the environmental/training effect can be greatly affected just by the simple act of telling someone they can or cannot do something or fit some role. That is, it might be the case that American XY-males today are better (by whatever quantitative measure) at math and science than American XX-females today strictly by environmental factors and that by repeating "men are better at math than women" we are exacerbating and continuing this problem.
Now, all this said, it shouldn't be the case that all differences are strictly off-limits since, for example, there is a high incidence of lactose intolerance among south Asians and an incredibly low incidence among northern Europeans so menus for both groups absent other information would likely have necessary differences. And, given an equal society with a different mind towards things data on other differences might be interesting. But, widely speaking, discussions on potential differences is necessarily an area of little to no information and has high potential to be actively harmful to society.