The following is some work from Tim Berners-Lee’s group (I think) which I think gives some insights into how the semantic Web (on SAFE) can be used to deliver highly personalised services and applications without giving up control of your data to third parties.
In fact, they extract your data from v Facebook and Google+ and use this to pull together and mashup with data from your other browsing activity.
I hope some will find this gives ideas for how to build really interesting Web applications on SAFE, as well as illustrating the potential of semantic Web standards (LinkedData / RDF / Turtle, SPARQL etc).
@joshuef Maybe some ideas for Peruse 
PAPER: The Web Browser Personalization with the
Client Side Triplestore
My notes…
Browser with personal semantic data collection
- chrome browser plugin
- collects semantic data during browsing
- saves in browser storage with triplestore.js searchable API
- harvests personal contacts, messaging from Facebook /Google+
Features based on user’s personal data mashup:
- suggest text input (eg for search)
- recognises similar documents
- can group related resources (photos, travel schedules, people)
- query via facial recognition in photos
Two of three authors from MIT CSAIL (TimBL’s group?), this from Canon Japan.
Uses various JS libraries and parsers.