Interesting paper! Thanks for sharing it!

You make some good points, but the opposite points are worth thinking about too, to get the whole picture.

The brain-sex-classifier that they trained--using older, less effective methods than the current state of the art neural networks--was able to distinguish male and female brains correctly 77.8% of the time.

There's overlap between the distributions, but also, quite often, a big enough difference to know who is whom.

Then the question is why there is the difference--which the article doesn't address. Is it because of cultural expectations? Or because of hormone-induced developmental differences? Or non-hormonal sex differentiation pathway-induced developmental differences?

And the next question is whether there's a functional consequence and if so what--which the article also doesn't address.

And this only covers what they could infer from the activity of brains when the participants were at rest. If they had a more extensive dataset, presumably they could do better.

Anyway, it's a very interesting set of data, but even more than usual it provokes more questions than it gives answers.

(Also, the "androgynous brains are healthier" thing is very shaky. It's barely at p < 0.05, and the fits are dominated by a very small number of outliers. Since it's mostly a male effect, even if the conclusion holds up, which it may well not, it can be explained by "highly-feminized-brain men are anxious and withdrawn". See left side of graphs in Figure 4.)