He makes a number of interesting points, most of which I agree with. In particular, I think that OpenID while interesting in concept, is attempting to solve a problem that most users and businesses don't think they have (the notion of a common identity that is free from "lock-in", or something like that).
This is a common problem in technology: we identify a problem -- a real problem, mind you -- and we set out to provide a solution. The problem is, we fail to take into account whether the people with the problem actually care about having that problem solved. This is particularly important when our solution requires them to do something different from their normal routine. OpenID is, I think, an example of this sort of thinking.
As an aside, earlier today, I was looking at some new site (dopplr.com, to be exact) and noticed this weird little eye staring at me:
Turns out, naturally, that this is the OpenID logo. Okay, now that's just a bad logo. It looks like a freaky little eye with an exclamation point in it. Not exactly a warm-and-fuzzy feeling which i'd say is important when dealing with the idea of credentials. Call me crazy.