A lot of the problems of diagnosis happened because schools only referred the children who were really disruptive and really unco-operative, nearly always boys (since girls tend to rebel in quieter ways, or tend to be better behaved though there are exceptions to every rule).
So...they developed a checklist that matched the behaviour, not the cause.
We can be friendly. Many of the people with ASD I know are very friendly indeed and would like to have lots of good friends (perhaps one at a time, though). But we're not able to use body language, tone of voice or facial expressions very well. We're not able to see others do those things, so we miss 85-90% of the communication and can make terrible mistakes about what's meant.
Consider the phrase "Thanks very much!" If there's a with it, we can guess that this means thankyou.
Consider someone saying "Thanks VERY much"
We can't hear the difference, we can't see the sceptical face. We'll probably still think it's a nice thing for them to say
We're not good at remembering who's who, or getting quick information about that person out of our brains. It means that if we're somewhere familiar with one familiar person to chat with, we can cope in a friendly way. But put us in a mixed social group with a lot of noise and unexpectedness and watch us panic after a while. We can't handle it. As adults we get better at handling it for longer, but it is SUCH a strain.
Friends get very fed up with us getting it all wrong. They don't bother after a while. Well, some do (bless them)
What's needed is very sophisticated body language monitoring to see how different that is and whether there's natural copying, natural understanding of tones of voice, but none of the tests even look for that stuff yet. They've hardly begun to look at the sensory issues, and those are such a huge thing for us.
Like I say, the old ways of assessing an ASD were based around "is this a badly behaved unco-operative boy who won't look at me and is defiant about change". Not a good way to spot the rest of us.