I just finished the book, and I have to say that in addition to liking the book on the whole and finding the evidence and argument compelling, I was very satisfied with Diamond's use of linguistic evidence.
It was very clear that he had really done his homework, unlike some
other works that I have criticized. He used linguistic evidence to support other evidence and to help establish timelines and migration routes. For instance, I didn't know that the Austronesian family originated in southern China; I had assumed it came from Indonesia because that's the more central location.
Only once or twice did he mention something that sounded a little questionable. He mentioned that the Afroasiatic family probably originated in North Africa, but he didn't provide anything to back it up other than the fact that most Afroasiatic languages are spoken there. To support this claim, he should have mentioned that the Afroasiatic languages of Africa are more diverse, while the Afroasiatic languages of the Middle East are more closely related, which would show that they split off from the rest more recently.
He also mentioned glottochronology* in connection with the expansion of Bantu languages across sub-Saharan Africa. As
this article mentions at the end, glottochronology is only reliable when it is used on language families whose historical phonology is well known. However, he seemed to use the glottochronological evidence accurately, so this wasn't a big deal.
I was still nonetheless very impressed with his consistent and accurate application of linguistic evidence to show the expansions and movements of certain groups of people. I learned a few things I didn't know before, particularly concerning African, Southeast Asian, and Austronesian languages. Good stuff. I think I'll see if I can get my hands on a few of the books he used as references.
*Porter, was this the word that you were thinking of but couldn't remember?