It will be interesting to see how this experiment plays out. I wish Kurt success, but personally, am extremely skeptical. By removing Museums from the mainstream of social media to its own island you risk making them a backwater. Refining this metaphor, you reduce the likelihood that casual browsers and the merely curious will stumble across Museum content (and then become future seekers) in their everyday ramblings. I think people are more surfers than seekers, initially going to one of a few trusted sources (their e-mail, their banking info, the front page or funny pages of their hometown newspaper) and then letting their curiosity or social proof ("Hey, what's interesting all *those* people?") lead them onward. We appreciate well-crafted linearity, but learn and explore associatively, from one tangentially-related distraction to the next.
The dead link safe haven established and promoted by the government and briefly championed in the early 2000s by Smithsonian's Center for Education and Museum Studies (Smithsonian Education) haven't been resoundingly successful, which is perhaps not a surprise. Even little kids realize that you sell more lemonade from a street corner than from a cul de sac -- the advantages from increased foot traffic overwhelms the appearance of safety. What's more, it's a false dichotomy, since Museums like other users aren't limited to pursuing a single outreach methodology, except to the extent of their staff's limited resources. Why should museum staff invest duplicate effort? I've come to think that Facebook's power lies in its unlimited and free access to all, its enforced simplicity (despite its less customizable new design), and chiefly, its reinventability (the thing that makes Museum apps like Artshare work). Plus it has already grown its own audience.
Seems to me that offering a new business model is by itself insufficient to achieve the goals of promoting museum content in new ways on the web. Anyone wanting to establish a Facebook competitor will have to offer materially distinct functionality (say, shared web tools a la GMU's Omeka, or the kind of annotated blending of media attempted once upon a time by Smithsonian Folkways' Synchrotext, about which more here, here, and in parallel invention, as used by the New York Times and better still, in Washington Post's Debate Decoder). Even at that, substantial investment would be required to seed content and establish value. Users cannot (or perhaps better, should not) be obtained through fiat, but by the consistent presentation of valued, superior content with the least barriers to entry (even where those barriers are merely those of direct navigation).
Content providers like Museums succeed where they coexist with, invite the participation of, and facilitate feedback from an unrestricted audience and wither when they establish ivory towers. Like I said, should be an interesting experiment. As they always do, truth and talent will [win] out.