Iirc, this idea comes from a study that found people who used Tylenol while pregnant actually did have a slightly heightened chance of birthing an autistic child. However, that study did not in any way address why this was the case, so incurious fools take it at face value and never examine things any further than that.

This means that there is some very small degree of corelation, even if that corelation has failed to hold water when put to later tests. One of the many, many things that these idiots refuse to look in the eye is that this data point is a singular, contextless point that has way more evidence against than there is for.

I hear where you're coming from, but this one wasn't ripped out of thin air, making it that much more annoying to talk these people out of. So it's not just made up bullshit, but that is still about 99.99% of it.