Given the number of subtleties involved with contributor agreements I don't see how refusing to accept some means they're being inconsistent. He admits he doesn't yet know what a good one would look like and so maybe he looked at those and decided that they weren't good enough.
More useful would be comparing those they have accepted with those they haven't. The ones they reject, would you have accepted them? Do you think the ones they accepted are good ones? That would make it clearer whether you agree with his goal or not.
I'm still more stuck on the practical sides. I really hope we don't get contributor agreement proliferation, because that would be even worse than licence proliferation.