I have been reading Liz Pelly’s Mood Machine: The Rise of Spotify and the Costs of the Perfect Playlist (Atria / One Signal Publishers, 2025), which is an interesting trip into the work of Spotify playlists. I got it on a hunch and it plays out very well against the previous Spotify book that I was reading.
It was not quite what I thought it might be, but there is a really powerful run through of the way way that playlists work and how they have affected artists. Pelly follows this through pointing out how both listening and creation bots / farms are creating an own ecosystem that ultimately puts independent creators out of work. I do need to follow up on some other project links, but part of me is wondering about a digital methods approach to this that works with and develops Spotify Teardown. It’s interesting, but not overly surprising, that music becomes a rapidly forgotten part of the study. Rather, it focusses on a ‘hollowed out’ set of ambient and muzak (that I doubt the originators of either might recognise) that are created on spec and the ways that re-categorisation and classification are carried out. It does remind me of the subject of a PhD thesis on the algotorial side of music that I came into glancing contact with a few years ago.
What I like about the book is that that it does set out some avenues forward and ways that these work. There are things that can be tried and it becomes somewhat hopeful. For me, the book was of interest, but somewhat at a tangent to current research. It does provide a counter to the potential work on algorithmic accountability and places it in social context that might, with a re-reading of the more data focussed tome, provide a useful piece of work or experimentation.
No Comments