Dave McClure was right.
But also, the economic factors that would lead to more interop simply aren't there. The [perceived] costs of building another silo'd social network are always lower than building an interoperable one, where you have to wait for everyone else to move in lockstep. Unlike hardware which has a much longer refresh cycle, software is revved daily — and waiting for standards to evolve/keep up with market/product demands is a losing proposition — especially considering the real value that accrues from "owning" a social/interest graph (partial or otherwise).
Really, the economic upside for interoperating — considering that advertising is still the primary way to make money on the web — is just too low.