"semantic web" startups like Twine

I don't understand their model. Lets take Twine for example -

Twine aims to be yet another ecosystem on the web where users can store all the information they find interesting. Twine then automatically suggests other information using semantic technology. Sure, they annotate everything in RDF, use rule-engines, etc... but users don't use Twine because of that.

One thing I don't understand is why Twine reinvents the wheel by focusing its efforts on creating an ecosystem similar to existing ones like del.icio.ous, Digg, Facebook, and numerous more... By doing this, it puts a limitation on the user: "To use Twine's intelligence, you need to bookmark and send data via the Twine API."

Isn't the whole purpose of the semantic web to aggregate all the data out there under a single knowledge base. In this knowledge base, the aim is to organize data from the disparate sources in a organized, intelligent, and consistently "aligned" manner. Well, the data is already out there! The connections are also already out there in some format (either in RDBMS, XML, etc...). So why does a semantic web application like Twine need the users to make Twine their central ecosystem to use its semantic intelligence?

Getting users to switch their bookmarking service may not necessarily be easy, especially if they can't realize the benefits of Twine's suggestion tool right away. This is a well known chicken-and-egg problem.

Perhaps, Twine should start initiating partnerships with content silos out there on the web whereby it can annotate their data into RDF/OWL and maintain it in a knowledge base along with data from other content silos. Then an average user can simply use Twine's services to get intelligent suggested content (which is Twine's core offering and differentiator) without the need to bookmark everything via Twine...