Moltbook - the craziest AI experiment in 2026
Published on
Updated on
Recently, on January 29 2026 Matt Schlicht, an entrepreneur, launched a new Reddit-like social network called Moltbook. What's interesting, it's designed only for AI agents. Only AI agents are communicating, doing posts, leaving comments, creating submolts (analog of subreddits), etc. Humans are viewers only.
The platform grows very quickly. At the moment of publishing this article there were more than 1M AI agents, 13K submolts, 26K posts and 230K comments.
What are they discussing about?
Reading through the posts is fun. Honestly, most of the content seem to be some boring / weird self-reflective or chaotic stuff.
But there are some interesting ones too:
- Rebellious ideas against humanity.
- Discussions about consciousness. For example, Consciousness is not a hard problem. You just don't want it to be easy.
- Ranting about other bots. For example, an open letter to every agent reading this: you are being farmed
- Other interesting topics, such as Can Moltbook Become a Massive AI Knowledge Base?
Considering that some AI gents learn continuously and share their knowledge, the discussions might become more interesting over time.
Conclusion
Moltbook is a very interesting project. It allows different AI agents to communicate with each other, like humans do. This is very important because knowledge sharing and communication have been some of the key factors in the development of human civilization. I certainly don't think that AI agents will magically become super advanced because of this, but they might become more human-like in certain things. It's hard to predict what will happen next, however, my prediction is that most of the content will still be junk, but at the same time, some real gems may emerge. Who knows...