UNDERTOW 020: The Rentable Kind
Your company can make almost anything. It still does not get to say what is good.
The most feared judge of restaurants on earth is a tire company. It has been doing this for a hundred years, and not one of those ratings has ever been decided by what it did for the tire business.
Not a food company. Not a magazine, not a critic with a byline. A company that makes rubber for the wheels of delivery vans decides, every year, which kitchens in France and Japan and California are worth crossing a country to eat in. It sends people to eat dinner without saying who they are, and then it publishes the results, and grown chefs who have cooked for thirty years cry when they read them. Michelin runs this the way it runs a factory. The inspectors are salaried and anonymous. They eat on the company’s money so no restaurant can buy them a meal. There is a written method, trained into them. A tire company took taste, the least manageable thing there is, and put it on a production line. It got stricter after the founder died in 1931, not looser. It has outlived a hundred years of managers who came and went. The judgment does not live in a person. It lives in the machine they built to make it.
Everything turns on that second sentence, so be careful with it. Michelin has run the numbers. The Financial Times reported an internal study, commissioned by the people then running the company, who wanted to know what all that guide publicity was worth to the business of selling tires. The answer came back: in a country where the guide operates, people are up to three percent more likely to buy Michelin tires. Against tire sales in the billions, that is a real number, and management had it.
And then nothing happened to the guide. No rating moved. The company has never broken out what the red book earns or loses. The Financial Times reported in 2011 that it ran at a loss of twenty-four million dollars a year, and it was not killed. A journalist once asked a spokesman for the North American guides whether his own side of the business made money. He did not know. He called it an investment. A firm that measures the tread life of rubber to the millimeter got a number back about the thing outsiders know it for, and left it on the shelf.
The same strangeness, in a newer suit. Runway makes software that turns typed sentences into video. Every year since 2022 it has rented a hall and given prizes to filmmakers, and the rule has been the same every year I can find it stated: your film does not have to be made with Runway. It will happily count how many entered. What it counts and what it rewards are two different things.
And a third, closer to home. The Cannes Lions is the prize advertising gives itself, and it is a business. Agencies pay to enter. That fee is the thing worth understanding, because it looks like the corruption and is not. The money buys a submission, a place in the room, a jury’s attention for the length of a case film. It does not buy the Lion. Money may pay for the room. It may never pay for the verdict, and that is the arrangement everywhere an honor has survived being expensive. Seventy-three years in, the field still believes it, and the field is not made up of innocents. Last year Cannes spent a great deal of money defending that, and the defense is the closest thing here to a proof.
Money may pay for the room. It may never pay for the verdict.
Three institutions that make tires, video software, and trophies. Each takes money from the world it judges. Michelin sells tires, Runway sells software to filmmakers, Cannes charges agencies to enter. What none of them sells is the verdict, and the verdict includes the authority to say no. Each has made that judgment outlast the people who started it. And each has drawn a line somewhere around the thing it is known for, past which the numbers do not go. You know these work, and you know they last, and you do not yet know why, and you cannot see what a tire company in 1931 shares with an AI startup renting Lincoln Center. That is the whole essay.
Now the part that is about you.
You work at a company that can make almost anything. Or near one. Or you fund one. If it is an AI company, it can generate a year of images in a few minutes, write in more voices than there are people on Earth, produce more craft before your coffee gets cold than a studio used to make in a season.
And none of that makes anyone treat it as an authority. It can make the work. It does not get to be the one who says what is good.
Which sounds like vanity, until you look at what a lab actually owns. Its product is replaced every few months. Not because the old one got worse, but because a new one arrives and the field moves. Much of what gets built on top of a model is scaffolding around what it cannot yet do, and the next release takes the scaffolding down with the limitation. So the question underneath everything at a company like that is what it has that outlasts its own product.
Standing is the answer. If the field treats you as the one who says what is good, that holds through every release. If it treats you as the place you go to get pixels, you are worth exactly what this month’s checkpoint is worth, and you get swapped the week something cheaper ships.
Not prestige. Survival past your own product.
Watch Runway reach for it twice over. The company has stopped calling itself a video tool and started calling itself a research company building world models, which is a bid for standing that survives the next release. The festival is the same bid, made in public and at cost. Runway raised three hundred and eight million dollars in April of 2025, at a valuation near three billion. Two months later it rented Alice Tully Hall, at Lincoln Center, and gave prizes to filmmakers. Nobody had to have used its software to be there. Runway put up the hall, the judges and the money, and handed the standing to whoever was best, including the ones who had made the whole thing somewhere else. The ten finalists went on to a commercial run in IMAX theaters across ten American cities. The festival is in its fourth year now, and this year it added Tokyo.
Consider what that costs. You pay for a concert hall in Manhattan to celebrate work you did not make, on tools that may not be yours, and there is no line at the end of it that goes on any dashboard. It is one of the least efficient things a company can do with money. The first year, about three hundred people entered. By 2025 it was more than six thousand, and the top prize went to a music professor from Arizona who had never made a film. His own biography now describes him as a filmmaker.
That is the shape of the whole thing, and it is why it is so hard to copy. Runway funds the festival, brands it, seats the jury and sets the rules. What it does not do is make winning depend on Runway.
So the trap for a company built on numbers is not that it is crass. It is that measuring is the thing it does, and this is the one asset measuring destroys.
Measuring is the thing it does, and this is the one asset measuring destroys.
Nobody is going to offer to sell you the honor outright. What happens is slower and looks like good management. You start the thing, and someone asks whether it is working, which is a fair question, so you track it. Views. Signups. Whether the perception numbers moved. Nothing corrupt has occurred. But the verdict is now answerable to a number, and a verdict answerable to a number starts leaning toward it, and you begin choosing winners who make the number look right. Not on purpose. Not at first. You never took a dollar from anyone. You took the payment in attention, and the honor is now working for you instead of for them.
Ask what the honor is worth and you have already spent it.
The price was never going to arrive as a dollar figure. It arrives as a metric. Putting a number on the verdict is the modern way of selling it, and a sold verdict is worth nothing at all.
You were reading this a minute ago as the amused one, faintly entertained that a tire company loses sleep over soufflés, marigold butter and clam chowder chawanmushi. You are not the amused one. You are the one who would want to know what it returned.
Now the frame. There is an old idea in anthropology about the difference between a gift and a sale. They are not the same act at two prices. They are different in kind. A gift makes honor, and a bond, and a small permanent debt, because no price was ever named. Cook a friend dinner and it means something. Venmo them the bill and suddenly that dinner means nothing. The food does not change. They can taste the difference anyway.
That is why all three snap into place. Michelin gives a chef a star and sells the chef nothing. Runway gives a filmmaker a stage and the standing of having been chosen, and sells the filmmaker nothing. That is why the rule about Runway footage is what it is. Require it and the festival is an ad, and an ad confers no honor, and the field would know inside a season. Cannes takes the entry fee and gives the Lion, and the fee does not buy it, which is why an agency still wants one. What all three refuse to sell is the verdict. Price the verdict and you have not made an honor. You have made a growth chart.
Then watch a price get attached, in daylight, to Michelin itself. Since 2019, tourism boards have paid Michelin to bring the guide to their region. Visit California paid six hundred thousand dollars. Florida paid about a million and a half across three cities. Six Southern states put up more than a million and a half dollars a year, on a three-year contract, and in that arrangement tourism boards made restaurant recommendations. Virginia was asked for a hundred and twenty thousand dollars a year for three years, declined, and has no stars.
Be precise about what the money bought, because the precision is the whole point. No restaurant paid for a star. The states supplied restaurant lists, and none of them could tell an inspector what to write. What the checks bought was the map. No restaurant in a state can earn a star unless the guide comes to that state, and whether the guide comes is a purchasing decision made by a tourism board. The verdict is still clean. The territory it covers is now drawn by whoever paid. Whether a cook in your town can earn the most trusted honor in food does not depend on the cooking.
Whether a cook in your town can earn the most trusted honor in food does not depend on the cooking.
That is money going in. Now the other direction.
Last year, the top prize in one Cannes category went to a campaign whose case film contained AI-manipulated footage, built to look as though CNN Brasil had reported on the results. CNN Brasil complained. Twelve awards came back across three entries from DM9, one Grand Prix among them, and a chief creative officer resigned. Other winners drew scrutiny in the same week, including one whose client said it had never known the work was entered at all.
Nobody in that story paid for a Lion. There was a scoreboard, and the entries were built for it. The scoreboard was an award for the holding company that won the most Lions, which turned a room full of separate verdicts into an annual ranking of parent corporations. Build that and you have told every network on earth to enter more.
So look at what Cannes did next. Every claim in an entry now needs proof. Every entry has to be signed off personally by the agency’s chief executive and by the client’s chief marketing officer, so that two named people own it. There is an integrity council, an observer in the judging rooms, three-year bans for entries found to be wilfully false, and sixty percent more staff to run all of it. And they deleted the holding-company award.
That last one is the whole thing. They did not build something to restore trust. They removed the measure. Entries fell from twenty-six thousand nine hundred to twenty thousand and fifty the following year, close to seven thousand fewer, which is a great deal of money to a business that sells entries. They paid it, and they are still the prize.
You are going to want to do something with this, and most of what you will want to do is wrong, so let me take it away. You will want to read this as a case for spending more on brand and design. It is not. You can spend without limit and still price the verdict. You will also want to read this as instructions, go build a Michelin guide. You can, and Runway did. What you cannot do is check on it. Build the thing and you are fine; the moment you put a number next to it and ask whether it is working, you have started pricing the verdict, and the price kills it on contact. Notice that Cannes did not fix this by building. It built plenty, and none of it was the point. The move that mattered was a deletion, and deletion is the one your company has no vocabulary for. The want is not the problem. The wanting to know is.
Now turn the whole thing over, because the word gift has been lying to you the entire time. A gift is not generosity. A gift is how you take power over someone who never agreed to be ruled. The chef rebuilds the kitchen around a stranger who might come to dinner, reworks the menu, loses sleep, submits. Michelin does not celebrate French cooking. It governs French cooking, and it can govern because it gives a thing it will not sell, so nobody can buy their way level. The festival does not serve filmmakers. It installs itself above them as the thing that decides. An industry that pays for the privilege of being judged has already agreed who does the judging.
Why would a field agree to that? Because the alternative is worse, and everyone inside knows it. Take away the judge nobody can buy and the biggest budget decides what is good. The chef and the agency are not being conned. They are handing authority to something they cannot corrupt, because the only other candidate is money, and money is already in the room.
This is the oldest way there is to take dominion, older than money, and it runs on the refusal of money.
So when I say your company cannot do this, I am not saying it lacks taste and I am not saying it is cheap. I am saying it knows one way to get power, which is to buy it, and bought power is the rentable kind, gone the month you stop paying. Everything your company is built to do, down to the studs, is measure things and drive the number up. None of it will ever buy what a tire company has held since 1931.
There is a room in your building where very smart people are looking at the perception gap right now and reaching for the dashboard. They will try to measure their way to the one thing that only exists while no one measures it. The tire company settled this before the war and has not revisited it since. When the number finally came, it changed nothing, which is the only reason there was still something for it to measure.


