How is it that Google hires PHDs in HCI to design this stuff and they're still getting it wrong?
I don't think they're getting it wrong; I think they're just judging success by a different metric than you are. As other comments upthread have suggested, their success metric appears to be how many users they can keep within their walled garden. So the HCI experts aren't figuring out how to make Google's apps easier to use; they're figuring out how to make them harder to escape.
Case in point: a bunch of people in this thread have pointed out various things wrong with GMail's web UI, but nobody, as far as I can tell, has said "Screw this, I'm using a different email client." [Edit: I have now spotted one such person, but that's still effectively none in a thread with this many posters.] (For the record, I use KMail on Linux, the KDE 3 version, and I don't use GMail anyway, for reasons which go way beyond its UI.) To Google's HCI experts, that's success.
You're right but this is an anomaly to me. How is it that Google hires PHDs in HCI to design this stuff and they're still getting it wrong?
Perhaps I have misconceptions about the organizational structure of large companies but they do hire these people for a reason no?