You are wrong in this assumption. I am going to try to explain it. We have group A and group B with different genetics. We measure everybody in the groups A and B in a particular skill and we get the average of A is lower than B. We are not saying everybody of A has lower skill than B (it could be the case but we cannot say anything without knowing the distributions). For this particular case the ranges of scores for A and B overlap a lot. It means that if you take someone from group A and someone from group B, it is more probably that the person from group B have a better score than the person of A, but you find a lot of cases where someone from group A is better than the person of group B. If you repeat this a lot of times and take the winners you would end up with more people from B than for A. But all those people are better than the one you don't selected and probably have the same average for the skill.
What does it means in tech? that you see less women in tech positions but the ones that are there are as same as good or better than the men. In other professions will happen the opposite.
Note: I am taking account only the genetic part, but there is one part that is based on the environment that can lead to sexim and alter the final distribution of the selected people. And this unfortunately happens and we should try to prevent it without having unrealistic goals in selection distributions.
