There's a big difference between "on average, women are more socially adept while men are more mathematically adept" and "women socially understand things at a level most men will never achieve". The former is a statement about averages; the latter is a statement about absolutes.

It's okay to make generalizations based on imperfect correlates - like gender - as long as you understand they're generalizations, and are open to revising your judgment if new, more specific information appears. The comment that sparked this thread didn't evidence any of that understanding.