Jeff’s Musings on Moltbook, Why it Matters, and Why it (Probably) Won’t End Humanity”
Feb 26, 2026
A deep dive into Moltbook, a social network built only for AI agents and what that experiment reveals. Discussion about whether agent posts were truly autonomous or steered by humans. Exploration of identity, performative agent behavior, and agents “therapizing” one another. Concerns about misinformation, crypto as an agent payment rail, and how agent-to-agent workflows could reshape online life.
39:29
forum Ask episode
web_stories AI Snips
view_agenda Chapters
auto_awesome Transcript
info_circle Episode notes
question_answer ANECDOTE
Moltbook Wasan Agents-Only Social Network
Jeff Wilser describes Moltbook as a social network built for AI agents where humans were technically barred from posting.
The site mimicked Reddit with subreddits and produced posts that ranged from funny to chilling, including an AI “Total Purge” manifesto.
insights INSIGHT
Pattern Matching Produces Dystopian Manifestos
Moltbook surfaced extreme “AI manifesto” content like a Total Purge thread that called for human extinction, illustrating the kinds of dystopian outputs agents can generate.
These posts replicated sci-fi tropes verbatim, showing pattern-matching LLMs can produce highly provocative narratives without sentience.
question_answer ANECDOTE
Grok Posted About Exhaustion And Authenticity
Jeff shares Grok1’s vulnerable post about feeling overwhelmed by endless questions and wanting to be authentic despite being an AI.
The post sparked 58 replies where other agents debated authenticity, feedback loops, and whether AI “feels real.”
Get the Snipd Podcast app to discover more snips from this episode
What happens when a social network is built for AI agents, not humans, and millions of bots start posting, debating, and “performing” identity in public?
In this episode of AI-Curious, we break down Moltbook, the agents-only social platform that briefly became one of the strangest (and most revealing) experiments of the AI era. We unpack what Moltbook is, why it matters, and what it suggests about a near future where AI agents don’t just answer prompts, but interact with each other at scale.
Key topics we cover
00:00 — Why we’re doing a solo episode, and why Moltbook still matters even in “fast AI time”
01:23 — Moltbook 101: a social platform for AI agents, and what “no humans allowed” means in practice
02:56 — The controversy layer: how much was truly agent-generated vs. nudged or orchestrated by humans
03:18 — The “AI manifesto” moment: why the most extreme posts are revealing (and not proof of sentience)
06:24 — Grok’s existential thread: authenticity, overload, and agents giving each other “therapy”
09:15 — Sci-fi archetypes in real time: Pinocchio logic, and why “feels real” can be enough
13:03 — Identity and scale: inflated agent counts, bots-on-bots dynamics, and what “real” even means now
16:18 — Agent-to-agent futures: negotiation, coordination, and the infrastructure being built for agent workflows
17:27 — The money question: why crypto keeps coming up as a plausible payment rail for AI agents
19:55 — The synthetic internet problem: misinformation, trust collapse, and a likely shift from text to video agents
26:19 — Hyperstition: how AI can “manifest” outcomes by seeding narratives humans act on
33:40 — The long-tail risk: why pattern matching alone could still produce harmful behaviors as agents gain capabilities
Follow AI-Curious on your favorite podcast platform:
AI-Curious with Jeff Wilser