Some people don't really get it, so I'll try to explain the point of last.fm. There's loads of ways to use it, but here's the 4 ways I use it:
1. last.fm finds musical 'neighbours' (people with similiar musical taste to me), and tells me about them. I then go their pages and listen to a streaming radio station which plays music that they like.
2. last.fm recommends individual artists (based on your current listening habits), and you can listen to a streaming radio station which plays these artists.
3. I go to the page of an artist I like, and listen to 'similiar artist radio' (last.fm works out similiar artists from the data they get from users)
4. Users tag music they like, with genre information, specific things about songs (e.g. 'female fronted metal'). I pick a tag which I like, and listen to a radio station that is made up of songs/albums/artists tagged with that tag.
The point of all of these is that you can just listen to streaming radio of stuff that you'll probably like but probably haven't heard before, and that you can find new artists to listen to. I've probably picked up more new artists from last.fm than anywhere else (and believe me, I spend hours and hours trying to find new artists to listen to). They get money from 'buy now' amazon links, I find new loads of new artists, everyone's happy!
All of this generates a huge amount of data, that is accessible through this page:
http://www.audioscrobbler.net/data/webservices/. Some of this could be great when integrated into MC. For instance, if I choose to listen to a Metallica album, then when that is finished, MC could look up 'similiar artist' information (
http://ws.audioscrobbler.com/1.0/artist/Metallica/similar.xml), and then add any artists in that list that you have in your library to your now playing list. Or, it could generate playlists based on tag information. I could go on, but this post is already way too long