I think I would have to argue against the conclusion here about there being no modern Capoids in East Africa. Ive seen pictures of Sandawe and Dahalo people and they have the South African facial features of bushmen, or al least Nguni like Nelson Mandela. For instance this Dahalo man could be a Nguni speaker
http://www.ipacc.org.za/default.asp?sPage=/gallery/ea_sept02_SD/ea_sept02_sd01.asp and the Sandawe man on the right looks even more like a bushman
http://dsc.discovery.com/news/briefs/20030421/gallery/eve_zoom.jpg .
(Although I don't think anyone has ever suggested that most East Africans are descended from Capoids anyway. Where he says there is "no relationship between living East Africans and their South African counterparts", is he referring to East Africans in general? If he is, hes making an unfair comparison.)
I don't understand what he means by "Like the mitochondrial and Y-chromosome DNA haplotypes shared by modern Ethiopians and KhoiSan, these should be taken as signs of survivorship of ancient genetic polymorphisms and not survivorship of ancient people."
These haplogroups hes talking about, are a monophyletic side branch to other human haplogroups, or at least the paternal haplogroup I is, I'm not sure about the mtDNA lineages, but I was under the impression that they were also a monophyletic haplogroup.
And although hes correct in saying that language isn't genetic, and I know that the inclusion of East African click languages in Khoisan isn't very well supported, its controversial to say that the click is primitive. (Why would the click be lost?) So I still think that the correlation between genetic evidence, physical anthropology, click sounds and rock art strongly supports the old idea of East African bushmen, in my own opinion.