I have two questions. First which claim? I think I see at least two separate claims in memkpos post (that cognitive differences exist and that they don't matter). Second, what experiments are you referring to that disprove memkpos claim? I ask only to clarify.
What percentage of programming performance is visualizing rotations of 3d objects in your head?
The parent's post is based on a reasonable heuristic: Programming is an activity with challenges that require a broad set of skills both technical and social. If such work requires a sufficient diversity of skills to be done effectively, differences in individual skills are less important than the proficiency of the group as a whole.
Heuristics are the only thing that can guide here. How many studies are there that bridge the gap between basic cognitive tasks and something as complex as creating software at a company?
https://en.m.wikipedia.org/wiki/Sex_differences_in_cognition