The trouble with disallowing self-diagnosis of high-functioning autism - especially in adults - is that it's possible to conceal outward signs of the atypical thought processes or avoid triggers entirely. I have a (not self-) diagnosis of Asperger's syndrome, but since I've moved away from home, I'm in almost complete control of my surroundings, associating with people of my choosing, and generally not subject to anything that would be likely to make me behave in any unusual way. I doubt I could get a diagnosis if I walked into a psychologist's office today, but this isn't because I no longer freak out in contact with certain offensive textures or because I have learned to read minds. It's because I have created an environment for myself that is free of those textures and have made friends who are able to explain themselves to me, whereas as a child, my parents were not-particularly-accommodating obstacles between me and environmental/social self-determination.

Agreed, in an ideal world self-diagnosing would be simply foolish, for all the reasons you listed. Even in this one, it's not always a good idea.

In my personal case, I already have a diagnosis of being neurologically atypical, which was made when I was very young and specifically re-tested when I was in my late teens, so I'd be able to claim most if not all of the benefits of belonging to the neurodiverse community anyway. Questioning that the correct diagnosis was made is not the same as questioning that there was something diagnosis-worthy going on in the first place. This is a relatively common story, given that the 'high-functioning' variants of autism only became well known about 10 to 15 years ago, and before that time people who were only mild--to-moderately autistic got other labels instead.

In most 'true' self-diagnosis cases that I'm aware of, the person was already very aware that there was something unusual about themselves, and had looked at and rejected one or more other explanations before realizing that autism was the one that fit. Those people generally did seem to also be autistic (the similarities in processing style tend to be pretty obvious when you're interacting with someone in realtime, if you know what to look for) and the few of them that pursued official diagnoses got them easily.

I did run into a few people who did appear to be using the diagnosis for reasons like the ones you suggested; they stuck out like sore thumbs. My general observation about those people is that their general sanity waterline was very low, and that if they hadn't been able to use autism as their attention-getter, they would just have chosen another one. The self-diagnosis appeared to be a symptom, not a cause, of poor rationality in those cases.

By the way, if you're wondering why I haven't gone for a diagnosis myself, it's because a series of very bad experiences with psychiatrists and other mental health workers as a teenager left me with a bit of what appears to be PTSD. I could probably force myself to go into a situation like that, but the chance of my having a panic attack and getting myself into serious trouble would be significant, and the chance of me actually being able to interact with them well enough to get any benefit out of it would be negligible. (I don't consider the diagnosis itself to be a significant benefit, just a prerequisite if I wanted other assistance, which... not goona happen.) (This is, by the way, a good example of that 'quiet emotional system' thing... I'm having a very quiet anxiety attack in the back of my head, which isn't strong enough to interfere with my thinking much. If I were in the real situation mentioned, there would be enough of a feedback loop that the anxiety would go from being quiet to being overwhelming in about 5 to 10 minutes without extensive conscious management.)