If I could summarize, Uta Frith's view is that there is "autism classic" (those "significantly" impacted individuals quickly diagnosed in early childhood) and all the rest is "something else". I appreciate where her mind is trying to go here, but likely not accurate nor as simple as she would like unfortunately.
I am thinking the ASD-1, 2, and 3 is sufficient in categorizing the condition, but on the other hand, I can understand the concern with the ASD-1 category and all it's apparent inclusivity, as well as, this apparent societal trend towards needing an "identity" and a "label". I can appreciate and value neurodiversity within the human population. I think our contributions strengthens the species and is key to our long-term survival, not to mention pushing us forward in technology, culturally, and lived experience. Perhaps there have always been a sub-population of what we call "ASD-1s" within society all along, but now we have the tools to identify these individuals. Perhaps this is what we are "struggling" to understand and accept... this perception that there are "too many". Humans, and frankly, all social creatures on Earth tend to have "issues" with individuals who are "different" to the point of exile and even killing those individuals. There's a strong "primal" drive for this sort of behavior. We may be living in 2026 and we would like to think we are "civilized" and "above" other creatures, but at our core, there still appears to be a primitive ape.
Personally, "thinking" you are autistic and actually "being" autistic may be fraught with error. I also think the DSM is a good general guideline... but there seems to be some inconsistency in HOW and WHO is doing the diagnosing, especially for those potentially within the ASD-1 category. Knowing you are "different" and being able to somehow "check all the boxes" within your own mind can sometimes be a result of cognitive biases, and I would say, even for those who are doing the diagnosing. So, I am more concerned with the HOW individuals are being diagnosed and am suggesting a more comprehensive, structured, and agreed upon testing method.