Dave Winer mentioned a good article, from Microdoc News, called What Google Leaves Out.

With the iCite net, every iCite will be indexed. The prototype central index I am building, for the sake of simplicity, doesn't dig down deep into the iCites' content, but it is nevertheless updated immediately when a new iCite is created or updated.

My hope is, assuming iCites catch on, that someone who is really into indexing content on the Internet (like a Google) will plug into the iCite net and use its essentially realtime reporting on content creation / updating to produce a truly inclusive and current searchable index.

I don't think that Google's 98% or so coverage of the web is lacking for most searches for information. But, for example, with iCites and indexing, it would be conceivable to search on "friends of mine eating lunch in Noe Valley right now"--assuming people registering their lunching whereabouts might be as common, at least among some people, as posting to their blogs.

Of course, with the iCite net, that particular search probably wouldn't need a centralized search index--it could be done more peer-to-peer (iCite to iCite). So, maybe centralized search engines won't be interested in these kinds of "ends" / "edges" of super-specific querying. But, the iCite net will be open to integrate with centralized search engines, so it will at least be a possibility.