Ezra Marcus of the New York Times wrote an excellent piece about the OnlyFans economy. It describes the people and systems that operate behind the scenes to ensure "creators" make as much money as possible. For those not familiar, OnlyFans enables individual people to broadcast themselves to paying and non-paying viewers. It also enables online "creators" to chat with and sell digital goods to their viewers and fans.
OnlyFans can be used in a variety of ways, but it is most notable for its use by sexual performers who charge their fans for access to racy content. But OnlyFans is not merely a porn site. As I pointed out in Lovers and Leavers, the fans don't pay to watch; instead, they pay for personalized messages from their idols, videos that mention their names, and for having a say on what happens next.
Using OnlyFans is not a passive act of mass consumption; it's a relationship. As one anonymous analyst put it, OnlyFans provides a "middle-ground between porn and a girlfriend." Like Starbucks, OnlyFans is a third place, something between a business and a home. You pay for the coffee, but someone takes the trouble to write your name on it and call it when it's ready.
The highest-paid performers on the site generate upwards of 10$ million per month. This is an early indication of the fact that even the most in-person services can be performed online and can be scaled. A single performer can serve hundreds or even thousands of customers in parallel. As Marcus's piece illustrates, creativity and cognitive power can transform even the most physical and personal tasks. Marcus described the emergence of agencies that help OnlyFan performers optimize their online business. Such agencies offer various solutions like web design, social media management, and analytics. But they also offer something more: They have an "army of "chatters" who speak and interact with customers:
"These chatters work in shifts, responding to incoming messages and reaching out to new subscribers, trying to coax them into buying expensive pay-per-view videos. They tell particular subscribers that a video was recorded just for them; in fact, the same clip might be sold to dozens of people. The chatters earn a small percentage on most sales, and the rest is split between the agency and the model. The subscribers presumably think they’re talking directly to the woman in the videos, and it is the job of the chatter to convincingly manifest that illusion."
The online persona is, in fact, a collection of individual employees across the world: A graphic designer in Croatia, a manager in Miami, a performer in Vladivostok, and a "chatter" in the Philippines. The division of labor — the division of a single individual into multiple people, each performing they're best at, and each compensated based on the specific value-add they contribute and their relative pricing power.
This is an unimaginable increase in the division of labor. The coordination afforded by the internet means that fitting each person to specific tasks is no longer enough. Instead, the person itself is divided into several functions. In a way, every part of the creator's body and mind is handled by someone else — the posing, the thinking, the speaking, the typing.
This may sound depressing. But it might offer some hope as well. It shows that humans might be able to remain competitive and useful in the face of growing automation. Let me explain.
A couple of years ago, I wrote about the emergence of agencies that develop virtual humans: Imma is a Japanese model that works with brands such as IKEA and Salvatore Ferragamo. She has hundreds of thousands of followers on Instagram and TikTok and has even collaborated with several other models and entertainers. It looks like she has everything going for her. The only problem is it doesn't exist.
Imma was created by Tokyo-based Aww Inc., a "virtual human" agency. Aww works with Nike, Dior, KFC, Porsche, and Adidas to create online personas that appeal to different consumers. Some of these personas — Imma and her colleagues — have appeared in Vogue and Harpers Bazaar.
When I originally wrote about Imma, I was wondering how long it would take before virtual humans of her kind will push real humans out of business. What's happening on OnlyFans offers an alternative scenario. Instead of producing new virtual personas, skilled people can identify actual humans that have potential and help them scale their "business" and reach a broader audience.
This process is similar to what's happening in other parts of the media landscape. A growing share of music and video are no longer produced by media giants. Instead, skilled individuals (and algorithms) sift through content that was created by millions of individuals and figure out which pieces (and people) to promote.
The lesson here is that whatever AI creates will not simply compete with what skilled humans create. Instead, it will compete with millions of things created by multiple people and sifted by other people and by AI itself. In many cases, the result of the latter process will result in a more successful product. And at some point, these processes might converge, and AI's own creative approach will rely heavily on mixing and mashing work by multiple humans. The important thing to note is that humans will be necessary — both to produce "the raw "material" and to help provide feedback on what might appeal to other humans. All this, of course, is already happening but to a limited extent.
The type of performers that do well on OnlyFans also offers some hope. Jayson Rosero, the Miami-based "orchestrator" interviewed in the New York Times story, explains that people are attracted to relatable, imperfect performers: "They want something that's, like, attainable, or that they can see themselves with... someone they can imagine talking to in r" al life."
Yuval Noah-Harari predicted that many people are about to become useless. He's wrong. We will remain useful, and our primary use will be entertaining and providing human attention and affection to other people. And we will be paid to do so. Is this better than being I'mless? I'm not sure.