This would be controversial if it were advanced in a more scientifically rigorous way:
If it turns out that the heritability of intelligence is relatively high in the developed world, then it may be that the Left-progressive project of ameliorating class based differences in access to cognitively enhancing environments has succeeded to a large extent. Barring genetic engineering this is the “end of history” for this project. It is a matter of when, not if (i.e., if you reject that the project has hit sharply diminishing marginal returns, logically it should at some point if the Left-progressive project succeeds). Assortative mating and more transparent meritocracy should allow for cleaner sorting within the population, and inter-generational class churn should decrease and stabilize at a basal level dictated by the random environmental variables which no amount of social engineering can squeeze out of the system. A perfect meritocracy would replace cultural class with biological caste.
I’d like to come back to this later, but what Khan is arguing is pretty straightforward: that relative affluence having been achieved, those environmental factors that prohibit the full expression of genes for intelligence–i.e., how they would express themselves given an optimal environment–are removed, and those with purely better intelligence genes will rise to the top. The analogy is to height, where given appropriate nutrition and air quality, which the West largely enjoys, the nearly 1.0 factor that determines your height is your genes.
I think what Khan misses is that gene expression is not quite so clean. There are different optimal environments for different genes. Genes are not single units but usually bundles of different individual genetic units, and they operate in not only the physical environment, but also their own genetic environment. For each organism, genes will thrive in different types of environments depending on the various other genes they have.
In other words, the “nature versus nurture” debate is probably incoherent. Gene expression is too complex, and where “gene” begins and “the environment” ends.