There have always been proprietary corners of the Web. For example, unless you're a subscriber to the online Wall Street Journal, you're locked out of most of the Journal's content.
The service makes hitherto fee-based content available for free to Google users--at least, the first page is free, as anything beyond the first page (or "click") requires payment--and therein is both the hook and the danger, as Philipp Lenssen describes:
There once was a time when Google search tried to be a neutral bystander, watching the Web without getting too actively involved. There once was a time when Google instructed Webmasters to serve their Googlebot the same thing served to a site's human users.
Now Google is officially telling Webmasters they can serve one thing to people coming from Google Web search, and another thing to people coming from elsewhere.
Google's organic results thus become not any view onto the Web, but (rather) a special one. You may prefer this view--when using Google you're being treated as a VIP, after all!--or dislike it. And it might force you to rely on Google even more than before, if some publishers start creating one free Web site for Google users, and another free one for second-class Web citizens.
Carr describes this as a "helluva good business idea," and he's right. He's also right to point out that it's a "subtle form of lock-in," as it makes Google's Web better to browse than the vanilla Web. This may not be a bad thing for users in the short term, but is it good for the Web in the long term?