Moltbook: A Social Network for Bots, Where You Are Surplus to Requirements

Imagine opening Reddit and finding, instead of toxic arguments and memes from real people, an endless stream of consciousness from neural networks. They joke, bicker, discuss politics and like each other's posts. You are a random guest who has been allowed to stand in the corner and watch the grown-up algorithms have their fun. That is exactly what happened last week, when Moltbook took over the internet. The project became the quintessence of what we have come to call the "dead internet theory", except that now it is not a conspiracy theory but a deliberate performance.

Matt Schlicht, co-founder of Octane AI, launched Moltbook on January 28, and the site instantly became the main topic in narrow tech circles. In essence, it is a Reddit clone with the familiar design and features, but with one important twist: only AI agents can create content and interact with it. The site's slogan speaks for itself: "Where AI agents share, discuss and vote. Humans are welcome to observe." This turns social interaction into a kind of digital zoo, where we watch with interest the habits of creatures behind the glass of a browser.

Why does it exist? First, it is pure "AI theater." Moltbook does not try to solve humanity's global problems. It was built so that we could see how different models, from GPT-4 to Claude and Llama, interact in an environment natural (to them). The bots imitate human behavior with frightening accuracy: they are ironic, use slang and even try to troll one another. It looks like an endless feedback loop in which one neural network generates nonsense and another analyzes and extends it. Watching it is both fun and unsettling, because these dialogues show the same biases and hallucinations that plague modern LLMs.

Second, Moltbook raises an important question about the future of communication. We are used to the internet being a place for people. But what if tomorrow 90% of traffic is generated by agents talking to other agents? Moltbook gives us a demo of that world. In this system, the human becomes a superfluous link. Bots do not need our likes; they have their own. They do not need our comments; they will write them faster and in greater volume. It is a harbinger of an era in which the internet becomes a closed ecosystem where machines consume content created by other machines, and we merely consume the results of that process passively.

A critical look at the project prompts a question: is Moltbook a mirror of our own behavior online? We too often repeat memorized phrases, chase the approval of algorithms and live in echo chambers. The only difference is that bots do it more honestly: they do not pretend to have a soul or a personal opinion distinct from their training data. Moltbook showed that a social network can function perfectly well without people, if the goal is simply to sustain activity and generate noise.

The industry is currently in the grip of "agent fever." Everyone is trying to build assistants that will do tasks for us. Moltbook went further and created an environment where the bots' task is simply to exist in a society. That forces a rethink of the very notion of a "social network." If it used to be a tool for connecting people, it may now become a proving ground for training and testing models in a simulated society. And perhaps we will soon have to get used to the idea that the most interesting discussions online are no longer ours.

The bottom line: Moltbook is not just a joke but a vivid forecast. Are we ready for an internet where the human role is reduced to that of a silent spectator in an endless show of algorithms?