Attention Was Always the Product

Nobody in the room at an upfront is buying shows. They’re buying audiences. Demographics, minutes, verified eyeballs. The show is the lure, not the product.

From inside the ad and content stack this is so obvious it barely gets said out loud, and from outside it’s still weirdly invisible. Media companies never sold content. They sold aggregated attention, and content was the harvesting mechanism.

The history runs in a straight line once you see it. Newspapers sold readers to advertisers; the cover price never came close to covering the newsroom. Broadcast sold ratings points. The web sold impressions. Platforms sold targeting. Each generation, the content got cheaper relative to the attention-measurement apparatus wrapped around it. The trendline was always pointing at where we are now.

What AI completes is the inversion. When content cost approaches zero, the pretense drops entirely. Content becomes a pure attention-harvesting function: generated on demand, optimized per viewer, disposable by design. The feed doesn’t need hits anymore. It needs your next thirty seconds.

Once you see it that way, a lot of otherwise confusing behavior becomes legible. Why platforms optimize for retention over any particular piece of content. Why “quality” arguments rarely move the people running the machine, no matter how eloquently critics make them: the machine is scored on attention, not craft. Why creators often feel like contractors for an attention wholesaler. That feeling isn’t paranoia. It’s a fair read of the incentive. It was true before AI. AI just makes the pretense harder to maintain.

There’s a practical turn here for builders, and it’s the part I’d want a founder to walk away with: know which side of this trade you’re on. If your product monetizes attention, you’re in the harvesting business whether you admit it or not, and the harvest is industrializing. If you charge users money, you’re in a different and currently rarer business, one where the customer and the product are the same person. Neither is wrong. Confusing them is, and the confusion gets more expensive as content costs approach zero.

Now the concession, because “media was always just attention arbitrage” is too clean, and I’d rather break my own frame than have someone else do it. Subscription businesses genuinely sold content. HBO then, paid newsletters now, and they’re growing precisely as the ad models strain. And even the attention business required making things people chose. The lure had to be good, and “the lure had to be good” is where all the culture came from. The cynical frame explains the machine but not why the machine produced The Wire.

The truer story is that the attention business and the content craft coexisted in productive tension for a century. The business side needed the craft to harvest, so the craft got funded, and occasionally something great slipped through a machine that only wanted eyeballs. The real event of the AI era may be that the tension is snapping, because the machine no longer needs craft to harvest. Retention can be optimized directly now, no showrunner required.

Whether anything culturally worth having survives inside a harvesting machine that no longer needs it, or whether craft migrates entirely to the businesses people pay for, is the question the next decade answers. Both futures are already visible from here. They may both happen.