home | hauntedcastle.org | sao bento music | mobile | about
www.flickr.com
|
Recent Posts -> so much goin... -> mobisodes in... -> mobile album... -> NIME 2005... -> game music... ->
« digital watermarking | Main | dynamics for designers »
April 21, 2003
calculating my interests @06:40 PM
http://www.synapseai.com/
http://www.tivo.com/
http://www.amazon.com/
The idea of tracking user interests has been around forever - it is the fundamental principle behind the idea of the demographic. I was reading the first link above, synapse, a plugin that tracks your mp3 playlists and then generates music that you want to hear. There are very interesting technologies at work behind this that deal with concepts of machine learning. But what it really did was get me thinking about what these machine selection / recommendation processes do to the control structures of information. My friend has TiVo. He set up his system, went through the "setup wizard" that tracks what kind of movies you like, a few sports you enjoy, etc. He tuned his tiVo to be smart, left for work, came back and was left with a hard disk filled with Martin Lawrence movies and high school field hockey or something. Obviously you can erase these things - but with TiVo, you can't go back and reset the machine's memory, as far as I know. This means that the machine keeps imposing this crap on him - he is thusly never in complete control of the machine. Amazon recommendations work a similar way, but it allows you to go back and edit your purchasing history so that the recs can be adjusted. For example, I can go back and erase the Puff Daddy cd I bought for my sister's birthday so amazon doesn't flood me with recommendations that come straight from the Bad Boy catalog. Thank god for that. So I have a little control over that system - and as a result, it works better for me- I like the idea of recalibration.
I think that such mechanisms are important, and can be linked back to the idea of who is in control of these systems, and what their motives are. For instance, amazon can make recommendations without thinking about which recommendations will pad their wallets, because it doesn't really matter if you buy an indie cd or britney - amazon is going to end up with the same profit. This type of system is therefore a type of cornucopia of the commons - amazon benefits, and the user benefits, and there is no strain on either side. However, when huge corporations get involved in this equation, then the commons is abused - the corporation pushes their own interests at the expensive of the user. There is a clear control / power mechanism at play in this equation. This is payola - if Amazon was bought out by, let's say, Time Warner - then I'm sure the "coincidences" of Warner Bros. artists bubbling to the top of my recommendations list would be funny indeed. And of course, by funny, I mean really sad.
How can these systems be used best? How can we alert people to things they might like, yet exist outside mainstream culture or Total Request Live?
I'm beginning to realize that everything is related.
posted by will | comments (0) | 
permalink
| all rights william carter |
| view cc license |