Non-falsifiability on account of self-report is a pervasive problem in mental health. But just because something isn't falsifiable doesn't mean it isn't true.

If I told you I had to consciously monitor my tone of voice, and mannerisms, and level of eye contact to pass for normal, and had to silently drill these things in my mind before starting or joining a conversation, that I basically had to be like a method actor and get into character before speaking, what would you propose as a way to falsify or verify my claim?

You're right: it's non-falsifiable. And you'd be right to raise an eyebrow about the idea that keeping it up would lead to clinical depression. Doesn't mean it isn't true.

I don’t think anybody has said AUTISTIC burnout is a thing. Certainly, I never did.

And having raised a “high functioning” autistic son, I definitely saw when interventions helped. And I also saw the struggles my son had that his coworkers would have just written off as “quirky”. Sadly, he stopped treatment, was overcome by those struggles, and took his own life, so at least for this sample of one, yes, the interventions help. Right up until they don’t any more.

There is survivorship bias at play as well.

Theres people who go in and DONT get diagnosed with ASD, but get found out to be having other issues, ranging from endocrine to trauma.

The evidence for treatment is the improvement in quality of life for people who do manage to get help.

The control would be people whose challenges make their lives, and the lives of those around them, harder.

The literal survivorship bias is from those who've managed to not commit suicide.