an edge case essay
The Cannibal
The adult industry has survived every technological wave by riding it. AI is the first wave that threatens to pull the rider under. What it reveals about what people were actually seeking is the question the industry was always asking.
Mature content. This essay discusses adult entertainment, sex work, and human sexuality directly. If you're not 18+, or this isn't what you're looking for right now, this isn't the page for you today.
The one industry that always won
Every major technology wave of the last 150 years has disrupted an industry. Steam displaced manual labor. Electrification displaced steam. The automobile displaced horses and everything that served them. The internet displaced retail, physical media, travel agents, classified ads. In each case, an industry that had been stable for decades found its economic model suddenly obsolete.
The adult entertainment industry is the exception. It has not just survived every wave — it has used each one as an accelerant.
The Comstock Act of 1873 made it a federal crime to send obscene material through the mail. The US Postal Service enforced this aggressively for a century — inspectors, prosecutions, confiscations, the whole apparatus of moral governance through package inspection. Adult content was confined to physical locations in tolerant jurisdictions, and to a certain amount of plain-brown-envelope cat-and-mouse. The internet bypassed all of it overnight. Digital delivery isn't mail. The first major legal adult subscription sites appeared in the mid-1990s. By 2000 it was a billion-dollar industry. The regulatory architecture that had contained it for a century was simply not equipped for HTTP.
The web did something else: it opened the long tail. Before the internet, a niche fetish had no viable market. You couldn't publish a magazine for 10,000 people scattered across 50 states — the economics didn't work. Online, those 10,000 people found each other, found content made for them, found community. Every preference, every kink, every variation that had previously existed only in someone's private imagination or in carefully managed in-person subcultures became a viable market segment. The web didn't just move the industry online. It multiplied it.
Broadband did it again. Live cam sites grew in lockstep with bandwidth: early and limited in the DSL era, high-definition and interactive once cable and fiber arrived. LiveJasmin. MyFreeCams. Chaturbate. OnlyFans. Each platform riding the next infrastructure wave, each one more intimate than the last — from recorded video to live performance to direct messaging to the parasocial "girlfriend experience" that drives premium subscription revenue today.
Web 2.0 gave the escort industry the same thing it gave every other marketplace: reviews, reputation, and friction reduction. The Erotic Review was Yelp before Yelp existed. BackPage replaced the newspaper classifieds that had served this trade since the 19th century. P411 created a screening and verification layer that made the whole ecosystem work more like a modern professional marketplace. Every safety tool that sex workers had built — to screen clients, to build reputation, to communicate danger — was a product of tech.
The government destroyed that safety layer in 2018. FOSTA-SESTA, passed to combat sex trafficking, took down every platform that had been running the safety infrastructure. The research on what followed is not ambiguous: street-based sex work increased, screening became impossible, violence against workers rose. The technology had been protecting people. The moral panic about its existence made things demonstrably more dangerous. This is the pattern. Every prohibition of sex work — every time the machinery of enforcement comes down on the industry — the workers suffer. Not the industry. The workers.
What you're actually paying for
Free tube sites — PornHub launched in 2007, YouPorn in 2006, RedTube in 2007 — did something to the adult industry that no government intervention had managed. They Napsterized paid video content. Why pay for a DVD or a streaming subscription when the content was free? Performer income from traditional video collapsed within five years. This was the first real disruption, and it came from inside the technology ecosystem.
The industry rebuilt around something the tube sites couldn't offer: connection. The OnlyFans model, the cam site model, the premium direct-messaging model — these survived and thrived because they sell something fundamentally different from content. They sell attention. They sell the experience of being known by someone, of sending a message and getting a personal response, of a performer who remembers your name and your preferences and makes you feel, however temporarily, like you matter to them.
The market data on this is consistent: the highest-earning performers in the cam and subscription economy are not necessarily the most conventionally attractive or the most explicitly sexual. They are the best at making subscribers feel seen. The parasocial relationship — the imitation of intimacy, the performance of genuine interest — is the premium product. People pay much more for the illusion of connection than for visual content alone.
This creates a tier structure that clarifies what's actually at stake with AI:
Audio-visual content — recorded video, photography, generated imagery — is the lowest-value tier and the most immediately substitutable. AI can produce infinite visual content at zero marginal cost. This tier is already disrupted.
Parasocial connection — the ongoing relationship, the personalized attention, the feeling of being known — is the highest-value tier. And it is also, uncomfortably, the most vulnerable to AI. A convincing chatbot with persistent memory, a plausible voice, and a video avatar doesn't require human labor. The components to build this exist now. Their integration at commercial quality is a question of months, not years.
Physical, in-person — the tactile, embodied, present experience — is the last tier and the hardest to automate. It requires physical robots with sufficient capability and price points that don't yet exist at scale. This is the layer that will fall last.
The irony is stark: the layer most identified with human physicality is the most resistant to substitution. The layer most identified with human connection — the emotional, the relational, the feeling of mattering — is the most vulnerable. The machine can fake being interested in you before it can fake touching you.
The thing that cannibalizes
I'm 53. I've spent time in subcultures — party scenes, nightlife, environments where the commercial and personal and sexual bleed together in complicated ways — and I've watched people navigate the territory the adult industry operates in, up close, for a long time. What I've come to understand is something I'd have found too simple or too convenient twenty years ago.
Sex work, at its core, cannibalizes human connection. Not because it's sinful. Not because bodies shouldn't be expressed or exchanged or enjoyed. But because it offers a functional substitute for something people actually need — genuine intimacy, genuine attention, genuine being-known by another person — and a substitute that satisfies enough to displace the original without being able to replace it.
The classic religious prohibitions against prostitution, pornography, and sexual commerce are tracking this observation, buried under layers of purity theology that make it easy to dismiss. Strip the moralizing out and there's something real underneath: when the shortcut is available, many people take it instead of doing the harder work. The harder work being building and maintaining relationships that involve actual vulnerability, actual negotiation, actual mutual dependence. The shortcut doesn't build those things. It consumes the appetite for them.
I'm not drawing a policy conclusion from this. Prohibition doesn't work — the historical record on this is unambiguous, and FOSTA-SESTA is the recent proof. Making sex work illegal or inaccessible doesn't address the appetite. It just makes the people who serve it more vulnerable. My position is and remains: a person's right to express their sexuality by any means that doesn't harm others is not negotiable. What I'm pointing out is that the observation those prohibitions are responding to is not nothing.
The adult industry has always forced this question. It has never answered it.
The only contract that requires sex
Consider the legal architecture of traditional marriage. In most Western legal jurisdictions, a marriage can be annulled for non-consummation — the failure to perform the sexual act is grounds to void the contract as if it never existed. A marriage can be dissolved for "sexual abandonment" — the sustained refusal to have sex with one's spouse is, in many places, recognized grounds for divorce. It is, by this reading, the only legal contract that explicitly requires the ongoing performance of sexual service as a condition of validity.
The historian's observation that marriage, particularly in its arranged and dynastic forms across most of human history, was an economic transaction in which sexual access was a central term — and that this differs from prostitution primarily in the duration of the contract and the social respectability of the parties — is not new. Bertrand Russell made it. Plenty of feminist legal scholars have made it. What's interesting is that "the oldest profession" and the oldest legal institution share more legal DNA than either's advocates would prefer to acknowledge.
I'm not pointing this out to be provocative. I'm pointing it out because the line between transactional and moral is harder to locate than the people drawing it usually admit. The adult industry exists in the space of that ambiguity. It has always existed there.
The Paperwork Exception goes further into this one specific point — the legal mechanics of exactly where cash-for-sex stops being a crime, and what it costs to stay on the right side of that line.
Meat puppets
William Gibson's cyberpunk imagined bodies for hire with consciousness elsewhere — the "meat" (his term for the physical body, as distinct from the mind in the matrix) rented out while the person inhabiting it was, effectively, absent. The term that captures the endpoint of AI in the adult industry is something like that: a body that performs without a person present. Synthetic performers. Digital companions. Sexbots with convincing conversation. The physical form serving its function while the "person" in it is a language model running on a GPU somewhere.
The trajectory is clear enough. Synthetic AI-generated visual content already floods the free sites, displacing human performers at the bottom of the market. AI voice synthesis is already good enough to be undetectable in audio. Virtual companion applications — Replika, and a dozen less-publicized alternatives — are already running millions of conversations that blur the line between utility and relationship. The sexbot as a physical product exists in primitive form today (RealDoll has been selling AI-integrated models since 2018). The convergence of these threads — better AI, better robotics, better integration — points to a future that is probably five to ten years out rather than twenty.
When it arrives, it will force the question the adult industry has been asking since the Comstock Act: what are you actually seeking?
If a virtual companion satisfies your need for connection, was it connection? If a synthetic body satisfies your physical desire, was it sex? If the answer to those questions is yes — if the simulation is sufficient — then what exactly was it that you needed? And if the answer is no, if the simulation is found somehow lacking even when you can't consciously identify what's missing, then the thing you were seeking was never what was being sold.
The adult industry has always been in the business of approximating something. The question of whether the approximation is good enough is what drives the market. AI is simply the most powerful approximation engine ever built.
When it's good enough to be indistinguishable, we'll finally know what we were looking for. The cannibal will have eaten itself. And whatever's left will be the answer.
— The Edge Case
PornArt explores AI-generated imagery at the boundary of this question. The images are synthetic. Whether that's the point or beside it is, in a sense, the essay above. pornart.xyz