I don't know enough above the subject (and what I could make of the article isn't helpful) to know if I agree or not. We should split Autism if we can conclusively separate people into the different diagnoses and then give them the correct treatment (which would be wrong for the other). However if we still give the same treatments in the end there isn't any point even if we can find different symptoms to result in a different diagnosis.

As science learns more (or I learn more) I reserve the right to change my position.