The Unfunded Wave

TAM-EAR.3-03 · The Standard with Ears · The Approximate Mind

Marcus has signed bands for thirty-one years, long enough to have done the job in three unrecognizable versions of the same industry. He keeps a cassette on his desk, a demo from 1996, four songs by a band nobody remembers, and he keeps it because it is the first thing he ever signed against the room. Everyone at the label heard the tape and heard a mess. He heard a mess too, and signed them anyway, because he had been to the shows, and the mess in the room was doing something to two hundred people that the tape could not carry. The band made three records. The second one mattered, in a small permanent way, to a scene that produced other bands that mattered more. The label lost money on all three.

Last month Marcus sat in a meeting where an act he wanted was scored. The system his company licenses had read the act’s streams, the engagement curves, the playlist velocities, the audio features against the features of everything that has ever recouped, and it returned a number, and the number was low. He argued for the signing. He is senior enough to argue. The room heard him out, looked at the number, and passed, and the thing he could not say, because there is no way to say it in a meeting, is that the number was not wrong. On everything the number measures, the act will underperform. That was true of the cassette too.

The hunch that funded the future

#

Production culture, film and music both, has always turned on a decision made in ignorance, and the ignorance was structural, and the structure is what this essay is watching close.

William Goldman put the industry’s epistemology in three words that everyone in it has quoted since: nobody knows anything. He meant it precisely. Not that the executives were fools, but that no one, however experienced, can know in advance whether a picture will work, because the audience’s response does not exist yet and cannot be sampled. Every greenlight was a bet placed before the information existed. The music business ran the same way and said so less elegantly: sign ten, pray for one, let the one carry the nine.

The nine are the point. An industry that cannot tell hits from misses in advance is forced to fund misses in volume, and the misses were not waste. They were the accumulation period, running inside the market’s own budget. The strange records that recouped nothing trained the musicians who made the records that did. Scenes lived for years on the label money that arrived by mistake, on signings like Marcus’s cassette, bets placed by people who had been in the room and could not prove what they had felt there. Motown ran a Friday meeting where every release was voted on, and the question Berry Gordy put to the room was whether, with a dollar, you would buy this record or a sandwich. The meeting was a filter, and it was a human one, which means it leaked: records got through on an argument, on a defender, on someone hearing something the room did not. The filter said no constantly and was constantly wrong in both directions, and the wrongness in the generous direction is where several futures came from.

The next wave was never protected by the old regime. It was funded by mistake, and the mistakes have stopped.

Nothing in the old arrangement loved deviation. The system was trying as hard as it could to fund only sure things. It funded waves anyway because it could not tell a wave from a mess, and that incapacity, embarrassing to everyone inside it, was the subsidy on which unformed things lived.

The number arrives

#

The score in Marcus’s meeting is the hearing, applied at the greenlight.

Read its construction plainly, because the construction is where the argument lives. The model is trained on what worked: on the audio and the metadata and the trajectories of everything that has already succeeded, which is to say on the accumulated output of every previous wave after it crested. Against that training, an unformed thing has a characteristic signature in that training, and the signature is noise. It resembles nothing that recouped, because the things it will eventually resemble do not exist yet; it will make them. The title essay, The Standard with Ears, found that newness fails both filters, unsellable because nobody knows what it is for, uninflammatory because nobody understands it well enough to be angry. Here the same anti-correlation appears in a purer form, as a scoring function: the model prices deviation against the past, and deviation’s whole value is its distance from the past, so the better the model, the more precisely it prices the future at zero.

The filter that runs this domain is the monetizable one, outright; a greenlight is a purchase decision about culture, and the score exists to protect the purchase. What the score removes is not the no. The old rooms said no constantly. It removes the argument, the defended exception, the leak. A number does not hear Marcus out. His thirty-one years enter the meeting as one voice against an instrument the company bought specifically because voices like his are expensive and inconsistent, and the instrument is not wrong on its own terms, which is what makes it unanswerable on its own terms.

And the suppression, if that is the word, happens earlier than any censor ever operated. The old standard interfered with finished culture: it banned the record, cut the film, dropped the band after the flop. The score interferes with formation. The deviation is priced before production, which means the wave is not stopped on the beach. It is unfunded at sea, before it has gathered, and an unfunded wave does not crash and does not register. There is nothing to point to. No banned film, no dropped band, no scandal. A set of things that would have existed, and now will not, leaves no evidence of the kind an argument can hold up.

The standard does not refuse the new wave. It declines to capitalize the accumulation, which no one can protest, because the wave that was not funded never happened.

What the misses were doing

#

The defense of the score writes itself, and honesty requires stating it at strength. The old system’s mistakes were not romantic to the people who bore them. The nine unrecouped signings were nine bands in debt to their label, nine sets of lives spent on advances that were never coming back. The slush was not a grants program; it was waste, borne mostly by artists, and an industry that wastes less is not obviously worse. If the model funds fewer messes, it also ruins fewer people in the making of them.

What the defense cannot account for is where the next form comes from. Every wave anyone can name was, at formation, indistinguishable from the failures around it, and was funded, when it was funded at all, by the same incapacity that funded the failures. The precision does not select better among the unformed. It selects against the unformed as a class, and redirects the capital toward the extrapolated proven: the sequel, the soundalike, the generated continuation of whatever the training data says the audience already wanted. Production has never been more efficient at supplying the present. What it has stopped buying is tickets on the future, because the future, scored honestly, always looks like a bad bet.

The unformed does not disappear when it is unfunded. It moves to whatever can be made for nothing, the bedroom recording, the phone film, and some of it will form there, at the pace poverty sets. Whether formation at that pace, in full hearing, can still compound into a wave, or whether waves needed exactly the combination of money and deafness the old industry accidentally supplied, is a thing the arc can only watch for. Other domains in these essays will show deviation surviving by other means. This one shows what happens when the deviation needs capital, and the capital has learned to listen.

We wonder how a culture would ever find out what it lost this way, since the missing evidence is the loss, and whether the first sign will be nothing more dramatic than a long, well-supplied sameness that no one can date the beginning of.

The cassette on Marcus’s desk still plays. He checked, a few years ago, on a machine borrowed from the archive room. Four songs, badly recorded, the mess intact. Somewhere in the second song is the thing he heard in 1996, and he can still hear it, and he has stopped trying to describe it to the meeting, because the meeting has a number, and the number has heard everything that ever worked, and the one thing it cannot hear is the thing that has not worked yet.

The band, if they came through today, would score low. That is not the system failing. That is the system working, on every measure it has, which is every measure except the one the cassette is on the desk to remember.

This essay is the third of the domain arc of The Standard with Ears. Its domain runs on the monetizable filter of The Standard with Ears applied at the point of capitalization, and its central observation extends the title essay’s anti-correlation: a model trained on what worked prices the unformed at zero by construction. The unfunded wave leaves no evidence, which places this domain’s losses in the class the corpus has examined as load-bearing absence. The bedroom remainder, deviation retreating to what capital is not required for, is observed here and left for the countercurrent arc. The scoring of culture before production stands in structural rhyme with the guideline that answers the nephrologist in The Counterparty: correct on every measure it has, and unable to be argued with in its own terms.