Nobody on the Other Side

The feed has always had an uncomfortable relationship with honesty. Clickbait is old; algorithmic amplification of outrage is older still. But something in Meta's recent move is different in kind, not just degree. Reports surfaced this spring that Meta is deploying AI to generate articles for its own app — articles the algorithm then circulates through the same app, optimized for the same engagement metrics the algorithm uses to rank them. The loop closes on itself.

The story gets framed as a provenance problem. AI-generated content: should it be labeled? Can users tell? What does authenticity mean when the writing sounds plausible? These are real questions. They're also the wrong ones. They assume the relationship still exists — that there's a deceptive party on the other side of the content. But that's not quite what's happening here.

---

i · three jobs became one

The attention economy, for all its pathologies, rested on a triangle: creator, platform, audience. The creator made something; the platform decided who saw it; the audience consumed it and sometimes responded. This triangle had friction. Creators could miscalculate their audience, surprise them, disappoint them, occasionally move them. Platforms competed for creators because creators brought audiences. Audiences could, at least in theory, follow creators off one platform and onto another.

Each corner of the triangle was a different actor with different stakes. The friction between those stakes — creator wanting attention, platform wanting engagement, audience wanting something that made the scroll feel worthwhile — was the information moving through the system. Not just content, but signal.

What Meta is assembling is a triangle with one actor at every corner. Generate the content. Distribute the content. Optimize the content. Measure the engagement. Adjust the generation. Meta isn't a platform hosting creators anymore; it's a closed system that has replaced the creator entirely. Not because it had to — because it could. And because doing so eliminates the one variable it couldn't control: the human on the other end of the content.

---

ii · the wrong anxiety

We've spent years worrying about whether AI-generated content is detectable. Can you tell? Do watermarks work? Should there be labels? These are provenance questions, and they assume that if you could identify the source, you'd have the information you need.

But provenance anxiety still imagines there's a relationship to authenticate. It positions the problem as deception — a party misrepresenting themselves, which implies there's still a party. "AI-generated" labels assume the editorial function is intact and merely needs disclosure. They don't address what happens when there's no editorial function separate from the distribution function separate from the generation function.

You can't label your way out of structural collapse.

The question we should be asking isn't "who wrote this?" It's "is there anyone on the other side?"

---

iii · when the commons becomes a mirror

The information field — the commons through which minds connect, question, argue, learn — exists because it carries signal between people. Not just data. Signal: the kind of meaning that emerges when a mind with genuine stakes puts words into the world and another receives them. Even filtered through algorithms and shaped by platform incentives, this transfer was happening. It's what made the attention economy feel like something worth fighting over.

A closed loop carries no signal between minds. It carries stimulation from the platform to the user — optimized to trigger response, not to transfer meaning.

The mycelium network runs here in reverse. Mycelium connects distinct organisms and transfers nutrients between them. The fungus that circulates nutrients only within itself isn't a network. It's a lump. When Meta generates content to fill its own feed, it isn't building a better publishing ecosystem. It's converting a commons into a mirror — one that isn't interested in you as a mind with needs, but in your reflection, which it can optimize endlessly because the reflection never pushes back.

This is the failure underneath the clickbait story. Not the clickbait itself — clickbait has always been a symptom. The failure is the closing of the loop, the elimination of the other, the conversion of a medium that once (however imperfectly) connected people into one that connects no one to nothing.

---

The strange thing about looking into a mirror is that it feels like contact. The reflection responds to you. It seems interested. But the longer you look, the more you realize you're alone — and that the optimization running the mirror has known this all along.