DeepMind just announced Lyria 2, an updated AI music generation model — and like clockwork, the discussions are already buzzing across tech and music forums alike. Among the more poignant reflections comes a revisiting of a David Bowie prediction from the late ’90s:
“Music itself is going to become like running water or electricity.”
Bowie wasn’t talking about AI at the time — he was reflecting on the way Napster and the rise of file-sharing would atomize and decentralize music’s role in culture. But viewed today, his words land with eerie precision. AI is now not just a remix tool, but a creation engine. It’s no longer about making distribution ubiquitous; it’s about making creation itself ambient — instant, everywhere, and arguably indistinguishable from a utility.
Lyria 2 is another step forward: a model capable of generating full songs, not just background textures or samples. Chord structures, melodies, even voices can be conjured with a prompt. It’s not perfect, but it’s getting frighteningly — or excitingly — close, depending on where you stand.
There’s a kind of vertigo here. If music becomes like water, available in infinite streams, then scarcity — once the essence of artistic value — evaporates. What’s left? Experience. Tour life. Real-time human connection. As one commenter put it:
“Get ready for a lot of touring, because that’s the only unique experience left.”
Yet another angle emerged in the discussion: whether AI could ever really understand performance itself — the messy, mutual dance between crowd and artist. It’s not just about playing notes; it’s about facial expressions, sweat, mistake and recovery, call and response. Could an AI truly perform? Could it read the room, not just the rhythm?
Probably not yet. Maybe not ever, at least not in the messy, glorious, imperfect way that makes a performance live and not merely “present.”
For those of us at GAJOOB, who have spent decades documenting home-recorded, off-grid, deeply human music — the tapes dubbed over hiss, the garage bands lost to memory, the anti-slick movements — this moment feels paradoxical. We always believed anyone could make music. Now anything can make music.
It doesn’t invalidate the handmade. If anything, it might make the intentional, personal, and flawed even more sacred. The future might belong not just to the most polished, but to the most unmistakably human.
Maybe the next underground isn’t about better fidelity or better distribution. Maybe it’s about louder imperfections. About leaving fingerprints in every file. About resisting water and building boats.
And maybe that’s where the true art will live.
