On October 29, two investors, Will Manidis and Joe Roberts, published a 145-page essay called The Commanding Heights. It asks where the money from AI will end up. Their answer is that most of it won’t stay with the companies that build the models, because models get cheaper every few months. The lasting fortunes, they argue, will go to whoever owns the places AI has to pass through to reach people. Their examples include the apps billions of people already use, the payment systems underneath them, and the dollar itself. They call those places crossings, and the ones worth owning are the Commanding Heights. The essay spread fast among investors, and its launch post has been viewed more than two million times.
Last week’s Run Their Test on Media examined how companies can challenge the tolls collected at these crossings. But I think the investors are missing something else about media. The work moving through the pipe comes with people attached. Writers, actors, musicians and photographers all have a stake in what happens to it, and some of them have won real power over what happens next. The people whose work moves through a crossing can change how it may be used and who still owes what, even when they cannot stop the deal.
One rule in this year’s actors’ contract shows how. A studio can sell its rights in a digital copy of an actor, and the actor’s protections travel with the copy. The studio can hand off its own responsibility for those protections only if the buyer takes them on in writing. The union then gets three weeks to object to whether that buyer can pay. If it reasonably objects, the studio stays on the hook. It can’t sell its way out of what it owes.
What the essay argues
The essay is written in a high, almost scriptural register, but its core argument fits in a paragraph. Technologies rarely make the most money where they are invented. Gutenberg went broke. The great railroad fortunes went to people who never built a locomotive. The authors expect AI to follow the same pattern. The labs will keep spending to stay at the frontier while the price of what they sell falls toward the cost of the computing behind it. The durable money will settle at the crossings. They define those as scarce, controllable positions that capture a share of the activity running through them and get harder to get around as that activity grows. The essay’s advice is for investors, and it sums that advice up in one line, that “one either owns the rails or pays those who do.”
Where I part ways
The essay already accepts that other people can lay claim to a toll, the cut a crossing’s owner takes from everything that moves through it. Its own test of last resort for a crossing asks who else can claim its rent, and it names regulators, ratepayer advocates, local interests and popular backlash. What its map cannot tell a reader is which of those claims leaves a seller liable, which forces a meeting, and which only puts an argument before a judge. In media, those differences decide what an objection is worth. Keeping a seller liable means someone still owes the performer after the deal closes. Forcing a meeting gets the union a conversation after the license is signed. A brief only asks a judge, and the outcome is the judge’s.
The last installment of the Legitimacy series ended on the same gap from the other side. In the provenance chain’s own published documents, the photographer is party to none of the governance. What follows asks what the people who are party to something can actually do with it.
