We chatted in the Com Theory course about the intersection between Horkheimer and Adorno’s “Culture Industry” and the ideas surrounding diffusion. The theme I’ve been pushing all semester has been where human agency exists and where it is constrained. I think that this is the most interesting and important question that communication has to answer, and since I’m the prof, I get to “profess” such a claim.
In our discussions, we touched a bit on pop music and other fashions. If individual will and agency were not at play there, couldn’t we predict hits. Strangely, I have a little background there, having had a chance to chat with a few people about the question of pop music and models of diffusion (particularly in the J-Pop context) while at the Santa Fe Institute.
Now comes this: an algorithm called Hitpredictor that seems to be pretty good at predicting Billboard hits. If it is true that it is possible to detect factors in the content of a song that will make it a hit, it raises questions about where taste is made. It also immediately forces some second order effects. As if the tin ear of an A&R person wasn’t already a poor excuse for a low pass filter, we now (potentially) have computers taking a first cut.
Diffusion, individuality, and Hitpredictor
We chatted in the Com Theory course about the intersection between Horkheimer and Adorno’s “Culture Industry” and the ideas surrounding diffusion. The theme I’ve been pushing all semester has been where human agency exists and where it is constrained. I think that this is the most interesting and important question that communication has to answer, and since I’m the prof, I get to “profess” such a claim.
In our discussions, we touched a bit on pop music and other fashions. If individual will and agency were not at play there, couldn’t we predict hits. Strangely, I have a little background there, having had a chance to chat with a few people about the question of pop music and models of diffusion (particularly in the J-Pop context) while at the Santa Fe Institute.
Now comes this: an algorithm called Hitpredictor that seems to be pretty good at predicting Billboard hits. If it is true that it is possible to detect factors in the content of a song that will make it a hit, it raises questions about where taste is made. It also immediately forces some second order effects. As if the tin ear of an A&R person wasn’t already a poor excuse for a low pass filter, we now (potentially) have computers taking a first cut.
Share this: