I have been in a web 2.0 conference all day and been pondering the Last.fm/CBS deal. One of they key tenents of the new business paradigm encompassed in "web 2.0" is that interactions with your customers need to move from a "push" model (you will take what I give you) to a "conversation" (tell me what you want and I will try to provide it to you). [Those of you that have read The Cluetrain Manifesto, this will sound very familar to you.]
I have to think this is one of the key hurdles that terrestrial radio needs to overcome - and Last.fm gives CBS (and their network of radio stations) the ability to overcome. Instead of radio station music directors programming what you should hear, they will move towards curating what you hear. The difference? In the former, they push.... in the later, you make suggestions and then select and refine. In the case of Last.fm, CBS can leverage all of the data collected via the millions of installed audioscrobbler plug-ins on users machines. That data is extremely powerful. Not only does it tell you what the "head" of the long tail is, it gives them visibility into what in the tail is starting to make it's way up the body. By aggregating the listening behaviors and trends of the audience for all of their stations, the users passively program the stations. Do terrestrial radio station music directors need to fear for their jobs? I don't think so, but I do think they need to learn the same lessons as the rest of us. Listen to you customer and give them what they want - don't give them something and tell them they should want it.
Last.fm also represents quiet a nice hedge for CBS in that they still have a (very compelling) play when broadband IP makes it way to the car. In addition to tuning into one of their standard "group" stations, users can also select their very own "personal" station (delivered via wi-fi or 3G networks).
There are lots of other interesting ramifications/opportunities represented in this deal... but those are the ones I'm currently chewing on. Thoughts?