I'd argue high intelligence alone is a neurodiversity, we just don't pathologize it. It has plenty of undesirable side effects too.
We also have situations where 3+ sigma outliers in brain function don't see any benefits at all and need constant support to survive in our environment (natural and societal).
So, no autism isn't the next stage of evolution, but having somebody with 2 sigma deviation from normal cognition existing and contributing to society is probably more of a good thing than a bad one.
[1] I'm thinking of evolutionary-level time frames here. There are certainly problems on a shorter-term, individual level, which most of the comments address.
I am curious to see how the population changes over time as a result of increased freedoms for women. It's only really been a couple generations so far (and only in some countries), but now that women and men are on more equal footing when choosing a partner (and both sexes are more likely to settle outside their hometown) you end up with a lot more pairings that match career status and to some extent intelligence.
There's definitely going to be (on average) a difference in offspring between an MIT dude that married his secretary and one that married someone also from MIT, and for all we know the former may be healthier, especially as you extend this N generations. Of course that's a drop in the bucket relative to overall population size, but I think the effect could exist across various societal groupings.