https://www.scientificamerican.com/article/broken-mirrors-a-...
STEM and Autism: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3620841/
Empathy: https://pubmed.ncbi.nlm.nih.gov/27246093/
Third one is probably extracted from the last two and colored by personal experience. I would tend to agree with the author though "appalling" is a little too strong. When I had a professor who was very obviously on the spectrum the class was often difficult due to either communication issues or assumptions about student skill level. It's not a universal truth, of course, but common enough in my Math and CS classes many of them were often frustrating to take. Stereotypes are often built in reality and while the author's language was strong I don't think the greater point they are making is necessarily wrong, nor unsupported by science as you seem to suggest.
I casually googled for sources for some of these claims: These studies do exist. But even if not, I would still think it is valid to talk about this phenomena (taking into account that measuring empathy, teaching ability or mass diagnosing large student bodies are difficult research tasks).
His use of "on the spectrum" could be considered offensive (similar to "retarded"), depending on whether you read it as him calling them that or read it as him describing what they call themselves. But the problem is not "good science."
Says the author without providing any piece of data or source of information. In particular
- where is the data that shows there's a significant portion of peoples on the spectrum in these fields?
- where's the data or source of information that proves that people on the spectrum are less empathetic?
- where's the data or source of information that proves that people on the spectrum are appaling teachers?
You can't launch such a statement without providing any source of information or data. Please, do good science, specially when talking about a vulnerable group.