> the vast majority of genetic variation exists within populations

This particular argument (I am not talking about anything else) always looked to me as "inkblot defense" (Cephalopods muddy water to defend themselves).

Genome is discrete. A single nucleotide polymorphism can have far-reaching consequences. So it's a bit like arguing that this collection of pentagons is not statistically different from this collection of hexagons because radius variation within collections is greater than between collections.

One day I've got into trouble by pointing to another genetic adaptation (EPAS1 SNPs) rather than the poster child of genetic differences: an SNP in the 6th codon of the β-globin gene. But that's another story.