Imagine this scenario:
You are using your media player/browser, and you stumble across a track on a webpage. You preview it. You like it. There's no information on the webpage about where it's from, but luckily, the mp3 is tagged correctly.
You hit a button on your media browser, and bam! You can find popularity information on Last.fm, prices on Amazon, full album information on Musicbrainz. Last.fm suggests a similar artist. The process rinses and repeats. You get to stuff yourself full of delicious music.
To see (some of) this in action; fire up Songbird 0.3 or greater, play some tunes, and go explore.
Last.fm's API is being used to render information that the page asked songbird about.
Other notes and odds and ends:
Songbird now takes regular firefox extensions, (say hello to greasemonkey) and Freebase is awesome.
Basically: Freebase has an API, it's got data from wikipedia, and you can do things like search for episodes of Battlestar Galactica. Or shows produced by JJ Abrahams.
So, mix the power of Songbird's webpage api with the vast data of wikipedia/freebase, sprinkle in a few applications, and you've got something far more fun looking than a simple google maps mashup.
Throw in greasemonkey and operator, and you can scrape information from pages into microformats. Then export them.