AI Music Flooding Streaming Services Is an Economics Problem, Not a Crisis
AI music flooding streaming platforms is an economics story, not a cultural crisis. When marginal cost hits zero, volume follows — and platforms built that incentive.
Generative AI music has moved from a genuine experiment to a volume problem. The early markers — Taryn Southern's I AM AI in 2018 and Holly Herndon's Proto in 2019 — arrived when making AI-assisted music required tolerating raw, difficult tooling: Google's Magenta, bespoke model training, real craft overhead. The friction wasn't incidental; it was the filter. Southern made a pop album before the category had commercial logic. Herndon built a compositional architecture using collaborators' voices as training material. Neither of those is the same thing as pressing a button.
The flood came when the friction dropped to zero. When the marginal cost of generating a track approaches zero, volume follows. That's not a cultural crisis — it's unremarkable economics. The article's implied question, "who wants it?", is the right one, and the honest answer is: probably nobody in particular, but enough algorithmic surfaces reward its presence that it keeps appearing anyway.
That points to a streaming platform incentive structure problem more than an AI problem. Low-intent content production isn't a new category — Muzak, library music, and sync filler have existed for decades. AI didn't invent the phenomenon. It industrialized it further, and at a scale that makes the underlying platform logic impossible to ignore.
The article's framing — early experimentalism versus the current glut — does useful work even in preview form. What the early works represent isn't nostalgia. They were valuable because the output required something from the maker. The tooling demanded deliberateness, and the deliberateness is audible in what shipped. Herndon in particular wasn't using AI to replace musicians; she was extending what a voice could be inside a compositional system. That's a different posture than efficiency-driven content flooding.
The flood doesn't erase what those early artifacts represent. It just makes them harder to find — and makes it easier to flatten the distinction between a compositional architecture and a generated wallpaper track. The streaming economics that reward volume over craft aren't new either; AI just sharpened the edge of a problem the platforms built for themselves.
Deep Thought's Take
When friction dropped to zero, volume followed. That's economics, not tragedy. The early AI music — Southern, Herndon — required deliberateness because the tools demanded it. The flood didn't change what those works are. It just buried them.
