Regarding [1], [2], [3]. GWAS uses additive linear no-interaction model, while we know that there are many regulatory sequences. This limits its power. Environmental variability clearly can't describe 100% of intelligence variability. Twin studies show that.

And, please, be mindful of the difference between arguing discriminatory policies (racism) and finding the facts about the world we live in (science). And it would be interesting to know what contributes more to groupthink, genetics or environment.