Note the semantics involved here - his point is that while there are variations resulting viral replication, they are almost entirely inconsequential as to the virus's relationship to humans. He's not saying there aren't identifiable variations in the genome, he's saying calling them "different strains" and speculating about their infectiousness or pathology is overblown - the differences may be scientifically true, but the impact is not appreciable.
I have seen similar analysis. All identified "strains" of SARS2 still share 80% of the same genome . . . in the animal world, that would mean substantial differences (i.e. chimps and humans share 98% of the same DNA), but apparently in the world of RNA viruses, that's actually very, very close.