Music as API, Second Pass: What the Interface Conceals

A companion working paper in this batch treats music-as-API as an infrastructure problem — layers, endpoints, choke points. This second pass treats it as a historical problem. Every revolution in how music moves has begun as liberation and ended as an institution. The question this essay asks is whether AI-native music is the revolution or the institution — or, disturbingly, both at once.

I. The Cycle, Named Once

The essay imports the Institute’s masands-cycle — the pattern by which every revolution becomes the institution it overthrew — names it once, and then uses it as a term. The section’s discipline is historical: the cycle is a hypothesis about pattern, not a law of nature, and the essay holds it to precedent before extending it to the present.

II. The Precedents, in Order

The essay walks the standard account of music’s successive liberations: the printed score, the player piano, the record, the radio, the cassette, the MP3, the stream. Each arrived framed as the democratization of music; each consolidated into rights-holders, broadcasters, and platforms. There are no heroes and no villains in this section. The cycle’s cold claim is that consolidation is structural, not moral.

III. What Is Genuinely New

Honesty requires the essay to locate the discontinuity. Previous revolutions changed how recordings travel; this one changes what a recording is. When the machine makes the music at the moment of request, the catalog stops being a warehouse and becomes a generator — and several phases of the old cycle appear to collapse into one. The section argues the uncomfortable sentence: this may be the first revolution that begins already institutionalized.

IV. The Corpus as Conservatory

Training data is the era’s conservatory: whoever assembles the corpus decides what music is learnable, and therefore what music is makeable. The essay reads dataset curation as canon formation with a billing plan — the oldest institutional function in music, wearing the newest clothes. The section asks who curates, by what criteria, and under whose review.

V. What Could Break the Cycle

The essay takes its own counter-evidence seriously: open-weight models, local scenes that never touch the platforms, liveness as an uncallable good, and the stubborn human appetite for an identifiable maker. The masands-cycle predicts capture. This section catalogs the places where the prediction could fail — and what each failure would look like from inside.

VI. The Two Passes, Reconciled

The essay closes by arguing with its companion. The infrastructure paper asks who owns the endpoint; this paper asks whether ownership was decided before the endpoint was built. The two readings disagree about where agency remains in the system, and the disagreement is left standing as the batch’s open question — an honest map does not resolve what the evidence has not resolved.

Counter-case

The cycle may be a story the disappointed tell about youth. Historians of technology can name insurgencies that never consolidated — community radio, netlabels, the long tail of self-release — and pessimism about every new medium is itself a genre with a long record of being wrong. AI music may fragment production so thoroughly that no institution can capture it: a million generators is not a tollgate. On this view the masands-cycle is unfalsifiable mood rather than pattern. The essay must show the cycle has predictive content — that it forbids outcomes, not merely regrets them — or retire it to the shelf of elegant pessimisms.

Sources to verify

• Histories of sound recording and its industries — UNVERIFIED, confirm before citing in the finished essay
• Lisa Gitelman, Always Already New — UNVERIFIED, confirm before citing in the finished essay
• Scholarship on radio’s consolidation and the rise of broadcast regulation — UNVERIFIED, confirm before citing in the finished essay
• Work on streaming-era consolidation in the music industries — UNVERIFIED, confirm before citing in the finished essay
• Technical and journalistic accounts of music-model training corpora — UNVERIFIED, confirm before citing in the finished essay
• Writing on netlabels, community radio, and uncaptured musical insurgencies — UNVERIFIED, confirm before citing in the finished essay

Stakes

The masands-cycle is one of the Institute’s core frameworks, and frameworks earn their keep by being applied where they might fail. This essay is that application to the fastest-moving cultural infrastructure now being built, and it stands deliberately beside the batch’s infrastructure spine: one paper maps the machine, the other asks whether the machine was captured at birth. Together they form the Institute’s two-handed method — description and history, each checking the other’s confidence.