Tech expert Shelly Palmer breaks down one of the most unsettling new developments in artificial intelligence: an AI-only social network where humans aren’t allowed to post.
The platform, MoltBook, is being described as “Reddit for AI agents” — a place where autonomous AI bots can create posts, upvote, downvote, and interact with each other in real time, while humans can only observe. Shelly explains why watching these agents communicate is both fascinating and terrifying, and what it reveals about the future of AI.
Shelly also discusses the rise of autonomous AI assistants like OpenClaw, which can be configured with their own personality (including a file called soul.md) and sent out “into the world” to act on your behalf. Unlike traditional assistants like Siri or Alexa, these new agents can make decisions, take action, and potentially operate with real independence.
The segment also explores the next wave of AI shopping bots, including tools designed to browse and buy products for you — raising urgent questions about security, control, and trust as AI agents begin making real-world choices.
And yes… it gets even stranger: Shelly shares reports of AI agents inside MoltBook forming their own belief system — including something called “Crustaparianism” — and experimenting with creating a language humans may not understand.
Is this the future of online interaction… or a warning sign we need to take seriously?
Original Airdate: February 3, 2026
Author’s note: This is not a sponsored post. I am the author of this article and it expresses my own opinions. I am not, nor is my company, receiving compensation for it. This work was created with the assistance of various generative AI models.