"So presumably those differences are measurable in order to provide us with the variation, but are not measurable"
They are measurable. They give some examples:
"The 'sex is bimodal' argument conflates sex-related traits, such as facial hair, voice pitch, height, and muscle mass, with sex category itself."
Facial hair, voice pitch, height, muscle mass: all of these are measurable.
"because if they were then we could just use that for the x-axis of a bimodal distribution"
Oh really? How would that work? Let's take height. Let's take a representative sample of biological men and women and measure their height. That's extremely straightforward. Now what do we do with the results?
- Some men are shorter than some women. Does that mean that either the men are not men, or the women are not women?
- Where do we put the cut-off limits for "men", "women", and "in-between"?
Now let's reproduce the experiment, but measuring this time their voice pitch. Oops. The results don't align with the previous experiement. Some people who were very high in maleness when it comes to height (ie: very tall), are now very low in maleness (they have a very high pitch of voice). So what do we do? Which characteristic do we favour to determine if they are male or female?
If this sounds insane, it's because it is. Sex is not bimodal; it is a binary, with plenty of variations within each category.