Evolutionary pressures aren't as simple as "if it lowers your chance of reproducing it's bad." You can read "The Selfish Gene" for more info.

Consider a gene that promotes compassion, selflessness and sharing. While it may lower your individual fitness, groups with some individuals with this gene may fare better as a whole.

Same as autism. For example, say autistic people see connections that other people don't. Having a low rate of "different thinkers" is positive for the group as a whole, even if the individuals may be disadvantaged in other ways.

Why would an individuals fertility be increased because the fertility of his group is increased? The gene would not propagate unless he has more children than his peers. There is no evidence of evolution by natural selection to begin with, though.

If the group with the trait has on average a slightly higher reproduction rate than a group without, that is sufficient.

This doesn't mean that any particular individual will have a higher rate (although most will).

One hypothesis for the spread of Christianity in the Roman Empire is that Christianity provided social services that improved reproductive rates by a couple of percentage points. Maintain that over a particularly stressful upheaval in the social order and you don't need many generations to significantly outbreed your rivals.

> Christianity provided social services that improved reproductive rates

Social services, and perhaps this might have some impact:

"In 1968, Pope Paul VI issued his landmark encyclical letter Humanae Vitae (Latin, “Human Life”), which reemphasized the Church’s constant teaching that it is always intrinsically wrong to use contraception to prevent new human beings from coming into existence."

https://www.catholic.com/tract/birth-control

The gene can be propagated by individuals who did not express it - it could be sex-linked, it could be recessive, it could require two or three conditions to be met, etc. (Also, propagation doesn't require 'more children than his peers', it just requires children that go on to reproduce).

> As far as I know there's no way for evolution to "sprinkled a bit of autism" across a population for an increased chance of population survival - I don't see how such a mechanism could exist.

Not strictly autism, but sickle cell anemia is a disease caused by carrying two copies of a recessive gene. If you have two copies of the recessive gene, you suffer from a debilitating illness. If you carry one copy of the normal gene and one copy of the disease causing recessive version, you have increased resistance to malaria, which has killed more humans than everything else combined. So evolution definitely does contribute to the prevalence of sickle cell anemia. Almost everyone in the US we suffers from sickle cell anemia is of sun Saharan African descent, where malaria is especially relevant.

While autism certainly isn't as simple as sickle cell anemia, there does exist the possibility that genetic components which contribute to autism can also contribute to other, more positive traits with fewer risk factors.

Well, I think one issue is that evolution acts on base-pair mutations and not traits. So (to wildly make a possibility up) suppose that mutations X, Y, and Z are each individually beneficial, but having all of them at once gives you autism. Then the population probably converges on a distribution that has some nonzero probability of X, Y, and Z.

That’s another 100% plausible evolutionary explanation. Autism is caused by genes that correlate with intelligence, which is positive for evolutionary fitness. If you get a certain mixture of those genes, you get autism. That won’t stop the genes from propagating though.

You're right and evolution is easily misunderstood as deterministic when it is only the result of random mutations in a population over time. Every variation you see does not mean it has a positive evolutionary advantage, only that those it wasn't prohibitive to reproduction