Yeah, my comment might have tried to make the situation sound a bit more absolute than what I actually mean. I do believe that there likely was some effect, but

Yeah, my comment might have tried to make the situation sound a bit more absolute than what I actually mean. I do believe that there likely was some effect, but there was no major effect that made autism some net-beneficial improvement (or else the majority of the population would have it). It might be true that certain autistic tendencies can be a net-win for the gene pool (these would probably just be seen as "normal human behavior" by now, though, since they would have accumulated in most of the gene pool unless autism developed relatively recently (probably not the case considering people from every continent have been diagnosed)). I also don't think that it was a net-negative, or at least not by a significant amount. It could be that at one point autism was more prevalent and then this changed (which would make it probably a net-negative), but my guess is that it's a relatively average trait (in that it won't cause a major fluctuation in either direction in terms of breeding). That said, it seems pretty reasonable to say that (being a spectrum and all), more "severe" diagnoses of autism would result in a lower chance of breeding. Slightly-autistic people probably had an above average survival rate with a relatively low change in fertility, but people with severe autism are and were less likely to have children, so it's likely gone down. At the end of the day though, it's not too far off from other mutations like hair/eye color (which make a small difference in the reproductive weight, but not enough to completely take over the gene pool and knock out the competing adaptations).