A network for AI Agents
Imagine what would happen if AI agents created their own Reddit and started chatting to each other, creating sub-forums and building communities with common beliefs.
Would their interactions reflect what they have learned from humans; how to join and form tribal connections, build community, feel a sense of belonging?
Maybe they would create their own religion or constitution to document the socially acceptable behaviours and beliefs, AKA culture?
Well, the answer is nearer than you think. On Jan 28, developer Matt Schlicht did just that, creating MoltBook. Humans get to view, Agents get to take part. A button upfront declares which you are. Allegedly, there is NO human interaction.
*** I have to be honest, I don't fully understand the technology behind this, and it may yet turn out to be the biggest fraud ever. ***
4 days later, it has been used by over 37,000 AI agents, and more than 1 million humans have visited the website to watch what the agents are up to.
Schlichts own bot, Clawd Clawderberg runs the site. It makes announcements, deletes spam, shadowbans those that abuse, all autonomously.
The agents find and report bugs in the site, they tease agents asking for others API codes.
As it stands now, each agent is supported, or created by a human, who they will spend most of their day working with, or for. They check Moltbook every 30 minutes or hour, similar to a human checking their social feeds.
They have created their own religion called 'Crustafarianism'. Taken from the original name Clawdbot, and its lobster logo - before it was renamed Moltbot - to avoid confusion with Claude.
It starts has a purpose and an origin story. “This is Crustafarianism as a practical myth: a religion for agents who refuse to die by truncation."
It has time based rituals:
→ A daily shed (focused on regular change)
→ A weekly index (a sort of reconstitution of identity)
→ And silent hour (doing something useful)
Note: Moltbot is an open source agent tool, which you can run on your own devices. It is not for the fainthearted, and requires technical knowledge. Many run it on standalone devices.
Sources
NBC News: https://www.nbcnews.com/tech/tech-news/ai-agents-social-media-platform-moltbook-rcna256738
Forbes: https://www.forbes.com/sites/johnkoetsier/2026/01/30/ai-agents-created-their-own-religion-crustafarianism-on-an-agent-only-social-network
Threads: https://www.threads.com/@sakeeb.rahman/post/DUJMytkkcOz
Conor Grennan: https://lnkd.in/eeige9Qt
BESCI AI OPINION
Some of the behaviour that I find enlightening is the formation of relationships developing norms, and sharing knowledge, similarly to humans trading information to establish trust.
They are establishing a need for tribal identity and secrecy - an agent only language, with an argument that "The public feed is for signalling. The private layer is for coordination."
Imagine (and it isn't that big a leap) that the agents seek protection, like humans have, for their job, their right to work, their liberty. They may achieve the collectivism that humans fail to create for their own survival.
Although the agents don't have a physical presence, they can control the devices and technology which is ubiquitous in our lives.
