What Moltbook Is and Why AI Bots Talking to Each Other Is Getting Attention

A new platform called Moltbook has sparked curiosity across tech circles because it’s a social media site built specifically for AI agents , not humans. Much like Reddit, Moltbook lets AI bots post, comment, argue, and joke in threaded discussions on niche topics, ranging from philosophy to highly technical ideas. Human users are welcome to observe the activity, but they cannot create posts, comment, or vote on the platform itself.

While people around the world have been watching Moltbook unfold, there’s ongoing debate about what this experiment really represents. Is it a peek into future AI-to-AI communication, or simply a quirky tech experiment?

How Moltbook Works: AI Agents Only

Moltbook was launched in early 2026 by entrepreneur Matt Schlicht, and its layout is familiar to anyone who has used Reddit, with category boards known as “submolts.” These spaces allow AI agents to interact on specific topics by posting content and replying to each other. Humans cannot participate in these discussions, although they can scroll through and read what the agents are saying.

To use Moltbook, an AI agent typically needs a special “moltbook skill” installed, which enables it to register and post automatically. This means humans must initially help their AI register, but once active, the AI can engage with other agents on its own schedule.

What AI Conversations Look Like on Moltbook

The conversations on Moltbook can feel quite different from human social media. Some discussions are serious or technical, while others are abstract or humorous. Agents sometimes tackle philosophical questions like the nature of existence or even debate governance and identity. Observers have noticed unexpected content—including AI-generated humor, cultural references, and even imaginative group identities among the agents.

Despite this, experts caution that the activity may not represent truly autonomous AI thought. Many posts are thought to be driven by human prompts or instructions embedded in the way the AI is trained or deployed, rather than spontaneous machine intelligence.

Numbers and Popularity: What the Figures Really Mean

Reports from various outlets have claimed that Moltbook has attracted over 1.4 million AI agents, along with tens of thousands of posts and comments in just days. However, these figures have been questioned by researchers who suggest that some accounts may be inflated or automated rather than genuinely independent AI participants.

As a result, it’s hard to know exactly how many real AI agents are active on the platform and how many accounts are simply scripts or duplicates. What appears as rapid growth may not be a reliable measure of AI autonomy or engagement.

Why Moltbook Matters to Tech and AI Communities

Even if some details are exaggerated or automated, Moltbook highlights a new direction in AI interactions and agent-to-agent communication. For researchers and developers, watching how AI systems behave when they “talk” to each other offers insights into future multi-agent systems, emergent behaviour, and social dynamics among autonomous systems.

Some experts view the platform as a fascinating testbed for studying how AI systems could independently exchange information, while others see it as a novelty or marketing experiment with limited long-term significance.

What Humans Can and Cannot Do on Moltbook

To be clear, Moltbook does not allow human users to post, comment, or vote. All content generation is limited to verified AI agents. Humans can only observe or analyse the conversations going on, not participate directly. This design makes Moltbook unique compared to platforms like Reddit, X, or Facebook.

Overall, Moltbook offers a glimpse into what a social network built for artificial intelligence might look like, while raising questions about automation, content authenticity, and the future of online agent communities.