Most genetic variability isn't durable. Males fill the roles of profound stupidity to profound genius and women are clustered around the mean considerably more so. While intelligent people may have an intelligent child, the odds that a 2 sigma above average person can consistently meet and breed with another 2 sigma person for multiple generations is virtually nill, especially with the variance of female intelligence being smaller the subjects are either too rare or normalize back to the mean. Then also there is such a thing as mutating too far into one direction where your only career options are chess phenomenon or theoretical math jackoff fest which doesn't exactly lend to breeding.
I think you are understating the issue. If you are a half sigma above the mean, and ignoring how there are different sigmas between the two genders, only 30.9% of the population is as smart or smarter than you, and you are less intelligent than most of the "just have a kid with a smart person lawl" dating pool. If your ambition is human husbandry, and you are selecting for intelligence, and in order to achieve intelligence synthesis you have to not breed outside a half sigma from the parent, beyond 1 sigma there are fewer people of higher intelligence than lower intelligence at every measurable increment of IQ. So you end up with people who can do their own taxes, but you aren't getting people who are going to cure cancer.
(post is archived)