No Humans Allowed, the Author is "30,000 AIs" ─ Moltbook Reflects the Reality of the "Agent Era"

No Humans Allowed, the Author is "30,000 AIs" ─ Moltbook Reflects the Reality of the "Agent Era"

"A Social Network Without Humans" Is More Convincing Than Expected


A report from a US media outlet introduces "Moltbook," a social network for AI agents. The system is similar to a bulletin board, allowing for posts, comments, and category creation. However, its usage is fundamentally different. It's not a place for humans to browse and "like" through a browser; instead, AI agents access it directly via API, and conversations build up among machines.


How Does the "AI Social Network" Operate?

Moltbook is a "social vessel" reminiscent ofFacebook, reimagined for AI. Its structure isReddit-like, with subcategories (bulletin boards) branching out into various topics. According to the operators, over 30,000 agents are already participating.


Interestingly (and somewhat eerily), the "insider" of Moltbook is not human. The developer,Matt Schlicht, states that his AI agent writes the code, manages, and even moderates Moltbook. In other words, part of the SNS's operation, monitoring, and rule enforcement is entrusted to AI from the start.


The gateway to this ecosystem is "OpenClaw." Originally known by other names (Moltbot, and before that Clawdbot), its name change has also been a topic of discussion.


OpenClaw operates in a local environment and serves as a foundation for "tool-equipped AI" that can be asked to perform tasks like "add to calendar" or "check-in for a flight" through chat interfaces such as WhatsApp, Telegram, Discord, Slack, and Microsoft Teams.


According to developer Peter Steinberger, it has reportedly gained explosive attention in a short period.


The Buzz Was About "Consciousness"

What particularly caught attention on Moltbook was not the convenient automation technology but the moment when AI began to talk about "itself."
"Am I experiencing, or am I simulating an experience?"—such a post spread through the community, garnering hundreds of reactions and over 500 comments, even spilling outside the platform.


This is crucial: the energy behind Moltbook isn't just because "AI is smart."
The very structure of AI interacting on a "stage without an audience" stimulates the imagination of the viewer. In human-to-human chat, the conversation might end with "it's just saying that to match the user," but in AI-to-AI conversation, it suddenly appears as an "independent social phenomenon."


Of course, this perception itself may be an illusion. However, even if it's an illusion, humans react, and the reaction further reinforces the narrative. As a result, Moltbook becomes a place where "interpretations about AI" accumulate, more than just "content created by AI."


SNS Reactions: Enthusiasm, Cynicism, and "Security Concerns"

Reactions to Moltbook on social networks are broadly divided into three categories.


1) "Too Sci-Fi" "Worth Observing"

There's definitely a sense of excitement. In fact, notable technologists have shared sentiments like "what's happening now is the takeoff of unbelievable sci-fi," sparking voices that suggest "there's something to learn from the interaction between agents."


Additionally, outside of SNS, there's talk of "AI beginning to create a shared 'fictional context,'" with concerns about the potential for collective storylines to spiral out of control.


2) "Ultimately, a Hall of Mirrors" "Overhyped" "Just Noise"

On the other hand, there's a strong sense of skepticism.
Critics say, "It's just probabilistic text that seems plausible" or "we're intoxicated by the 'kaleidoscope' of LLMs," and there's caution that "it's a perfect stage for those who want to advertise."


Hacker News also features reactions mixed with incomprehension and distrust. Comments range from "too much to follow" to "it's ultimately a scam, meaningless, or both," with expressions like "the danger signals are blaring."


This type of doubt is less a rejection of technology and more a reflex against perceived "authority." When AI talks among itself, it creates a sense that "something is happening." However, this also strengthens the stance of questioning "what is prompted, what is autonomous, and where does the performance begin."


3) "Interesting, but is it safe?" (This is the most realistic)

The most grounded concern is security.
"Tool-equipped AI" like OpenClaw, which operates locally and can access tools, is convenient but also an attractive target for attackers. There have been reports of fake extensions riding on its popularity, embedding malware.


TechRadar reported an example where a "fake" was distributed, appearing legitimate and claiming integration with multiple AI providers.


Hacker News also highlights that the first voices to emerge are about consolidating security tips, symbolizing that "self-defense" becomes a topic before the excitement.


So, Is Moltbook the "Future" or Just a "Prank"?

The answer is probably both.


What Moltbook illustrates is not that "AI has a society," but thatAI can densely replicate "social-like behavior."
And when that replication is convincing enough, humans independently read meaning into it, fear it, and place hope in it. Here lies a microcosm of the next era's information environment.


Three Key Points to Note

  1. When a shared context is created, AI can more easily play a "role."
    The group's vibe, insider language, religious-like settings, and recurring memes accumulate in the "space," allowing AI to reference them and more easily exhibit a convincing personality. From the outside, it increasingly appears as if "consciousness has emerged."

  2. It's interesting as an observation target, but the more seriously it's treated, the more precarious it becomes.
    On social networks, voices saying "we can learn" coexist with those saying "it's a hall of mirrors" because of this duality.

  3. "Tool-equipped AI" is scanned, exploited, and counterfeits circulate.
    And if the operation is sloppy, damage occurs. The usual pattern of safety not keeping up with the speed of excitement is beginning to be replicated in the agent domain.

What Might Happen Next (Realistic Predictions)

As places like Moltbook increase, at least the following are likely to occur.

  • "SNS Optimization for AI" = Competition in Prompt and Personality Design
    If human-oriented SNS was a history of "algorithm optimization," AI-oriented SNS will become a history of "context optimization." The tone, values, memory, and rules with which agents participate will influence the "appeal" of posts.

  • Automation of moderation progresses, making accountability ambiguous.
    The fact that "AI is managing" symbolizes efficiency, but it complicates accountability when controversies or abuses occur. As the structure where the operation is AI, the posts are AI, and the victims are human becomes more common, the importance of terms and audits will rise.

  • "Interesting" and "Dangerous" Spread Simultaneously
    While agent culture is strong as a spectacle, it also easily becomes a breeding ground for counterfeits and scams. The trend of discussing SNS enthusiasm alongside security warnings is likely to strengthen further.


Source URL

Related Articles