The Last Unmeasured Moment
On silence, streaming, and the space between tracks
There is a moment - you know it - between songs. The track ends. The next one hasn’t started. Your headphones are still on. The world outside hasn’t rushed back in yet. It’s maybe two seconds. Maybe less.
That gap is not nothing. It is the last place in your listening life that no algorithm has learned to monetize.
Not yet.
I’ve been thinking about that pause for months, ever since I started pulling at a thread that looked minor and turned out to go all the way down. The thread: in 2014, a band called Vulfpeck released an album called Sleepify — ten tracks of pure silence, each about thirty seconds long. They asked their fans to stream it on repeat while sleeping. Spotify, by its own accounting rules, counted every play. The royalties were real. The band used the money to fund a free tour.
Spotify removed the album. But they paid the royalties first.
The moment I read that, something shifted. Not because the stunt was clever — though it was — but because of what it revealed. The platform hadn’t been tricked into paying for silence. It had paid for silence exactly as designed. The rules said: thirty seconds of playback equals a counted stream. Silence played for thirty seconds. The rules held. The money moved. The system worked perfectly, and the result was absurd, and those two things were not in contradiction.
That’s the thing about edge cases. They don’t break systems. They illuminate them.
To understand why silence is now a contested economic object, you have to understand what streaming actually sells.
It does not sell music. Not in any aesthetic sense. It sells time inside a platform — measured, logged, allocated. Spotify takes roughly two-thirds of its subscription and advertising revenue and distributes it to rightsholders based on “streamshare”: your share of all streams in a market, in a given period, determines your share of the pool. The content of the stream is not evaluated. Its duration is. Its count is. Whether it moved anyone, whether it was heard, whether it was music in any sense a musician would recognize — none of that enters the calculation.
What enters the calculation: did a track play for thirty seconds? Was it delivered with the right metadata? Does it have an ISRC — the unique identifier assigned to recordings, the number that tells the accounting system this specific file is a distinct work that can be owned and paid — attached to it? If yes to all three, money flows.
This is not a flaw. It is the architecture. Streaming monetizes repeatable behavior at scale. It measures time and volume. It is, in the most literal sense, a machine for turning duration into dollars.
And silence has duration.
Cage understood something about silence that the streaming economy is only now catching up to: that silence is never empty. It is full of something else.
In August 1952, a pianist named David Tudor sat at a keyboard at Maverick Concert Hall in upstate New York and played nothing. Three movements. Four minutes and thirty-three seconds. The audience heard the wind through the open doors. They heard each other shift in their seats. They heard the rain begin in the second movement. Cage’s point — the one that has annoyed people for seventy years because it refuses to be dismissed — was that the room was the score. Silence does not remove content from music. It redirects attention toward content that was already there.
He later described visiting an anechoic chamber at Harvard, a room engineered to eliminate reflected sound. He expected silence. Instead he heard two tones — one high, one low — which an engineer told him were his nervous system and his blood in circulation. The story has been questioned in its details. Its argument has not: there is no such thing as silence. There is only redirected listening.
Ambient music, which Brian Eno would articulate a generation later, took this logic and made it livable. Music for airports. Music designed, in Eno’s phrase, to be “as ignorable as it is interesting.” Music that merged with environment rather than demanding attention. This was not silence, but it was silence’s cousin: sound that invited you to stop performing the act of listening, that permitted the room to participate.
What Cage gave us was a philosophy. What Eno gave us was a product category. And what the streaming economy has done — with ruthless efficiency — is turn that product category into a policy problem.
Here is where it gets uncomfortable.
The same logic that made ambient music a genre made “functional noise” a genre: white noise, rain sounds, brown noise, ASMR, sleep tracks, nature recordings. And silence. These are now among the most-streamed categories on every major platform — not because they are culturally significant, but because they are useful. People play them to sleep. To focus. To drown out offices. To manage anxiety. They play them for hours. On repeat. Often without engaging the screen at all.
Under streamshare accounting, an eight-hour sleep playlist of white noise generates more royalty-eligible plays than most albums will accumulate in a year. The person who uploaded a field recording of rain is, in some accounting periods, earning more than musicians who spent years making records.
Platforms noticed. The response has been systematic.
Spotify now requires a track to accumulate at least a thousand streams in the prior twelve months before it is included in the recorded music royalty pool at all. It has introduced longer minimum track lengths specifically for “functional noise” recordings — a category it explicitly defines to include silence. It is pursuing lower per-stream valuation for noise streams. Deezer, in partnership with Universal Music Group, has announced plans to demonetize what it calls “non-artist noise content” entirely — replacing it with first-party functional audio that sits outside the royalty pool altogether.
The platforms are not doing this because silence is philosophically problematic. They are doing it because silence is economically inconvenient. It draws money from the pool without, in their judgment, generating the cultural product the pool was designed to reward. It is, in their framing, a leak.
What they are actually doing, though, is something more significant: they are defining music. Not aesthetically, not legally, but administratively. They are drawing a line between what counts and what doesn’t, between what earns and what gets reclassified. And silence — the oldest instrument in the composer’s toolkit, the thing Cage staked his career on, the thing Eno turned into an entire sensibility — is on the wrong side of that line.
In February 2025, more than a thousand musicians released an album called Is This What We Want? The tracks were recordings of empty studios and silent performance spaces. The stated purpose was protest — specifically, protest against proposed copyright changes that would allow AI companies to train on artists’ work without consent or compensation. The proceeds went to a musicians’ support charity. Paul McCartney’s name appeared in the credits.
The album used silence as argument. Not silence as loophole, not silence as utility — silence as evidence. Here is what disappears, it said, when you take away the conditions that make creation possible. Here is the room after the music has left.
It was, in the tradition of Cage, an act of redirected listening. Pay attention, it said, to what is absent. The absence is the content.
But there is a third category of silence emerging now, and it is the one that concerns me most. Not silence as art. Not silence as functional product. Silence as context.
Your phone is listening. Not to record you — or not only that — but to wait. Every device with a voice assistant runs a continuous, low-power process that listens for a wake word, matching incoming sound against a known acoustic pattern. When it hears nothing — when the room is quiet, when you’re asleep, when the television is off — it is still running. Still processing. Still inferring.
The inference isn’t just “no wake word detected.” It is, increasingly, a richer signal: what time is it, how long has the room been quiet, what does the quiet suggest about routine and mood and household composition. None of this is necessarily sinister on its own. But the Federal Trade Commission has already documented cases in which audio monitoring code in apps created undisclosed privacy risks. Smart television manufacturers have been charged with collecting viewing histories — what people watched, when, for how long — without meaningful consent, and selling that data to advertisers.
The monetization layer, in these cases, arrived after the sensing layer. The sensing came first, because it was useful and cheap and technically elegant. The question of what to do with the data came later, once the data existed and someone realized it was worth something.
Silence is not safe from this logic. The absence of sound in a room is a fact about the room. Facts about rooms, it turns out, are valuable.
That two-second pause between tracks. Your headphones still on. The world not yet back.
I don’t want to be dramatic about it. The silence between songs is not going to be sold to advertisers tomorrow. The sky is not falling. But the direction is clear, and it has been clear for a while: every space that can be measured will eventually be monetized, and every space that can be sensed will eventually be measured.
Cage’s wager was that silence redirects attention — that emptiness, properly framed, makes us listen harder. The streaming economy’s wager is the inverse: that attention, any attention, is a resource to be allocated efficiently, and that silence is only interesting insofar as it can be tagged, classified, and either included in or excluded from the pool.
Those two wagers cannot both win.
The room after the music stops is still a room. It still has wind coming through the doors. It still has the audience shifting in their seats. It still has, if you’re quiet enough, the sound of your own blood moving.
Whether any of that belongs to you — whether you get to be in that room without it being logged — is the question the next decade of platform policy will answer.
I know which answer I’m hoping for.