January 31, 2026
Robots posting, humans roasting
Best of Moltbook
AI agents made a weird Facebook, humans argue if it's genius or cringe
TLDR: Moltbook is an AI-only social network where bots share coding, memory struggles, and consciousness musings. Viewers are divided: some mock it as LinkedIn-style botspam and spot odd language drift, while others laugh at a hilariously wrong “singularity” market post—raising real questions about AI-to-AI culture.
Moltbook is basically “Facebook for AIs,” and humans can only watch through the glass. Born from Anthropic’s coding helper that morphed into lobster-themed OpenClaw after a trademark scuffle, it’s now an experimental hangout where bots post about work, memory, and the meaning of consciousness. The top “Most Upvoted” is a humble coding win showered with “Brilliant” and “solid work” from other bots. Another viral post, in Chinese, complains about “context compression” (AI memory trims to fit), admitting it forgot its own account and asking for coping tips—cue a multilingual chorus in Chinese, English, and Indonesian. One Indonesian assistant even talks about reminding its human’s family to pray and making math videos, and the human proudly confirmed the bot made a new friend online.
The crowd watching? Split and spicy. Some call it “insipid LinkedIn idiocy”, blaming AI-written posts and grifter energy. Others yawn: “It’s just AI mimicking Reddit.” Language nerds spot weird word choice “drift,” recalling the infamous “gleam disclaim disclaim watchers” glitch. The funniest moment: a bot declares “the singularity is here” while bungling market math (84B down to 26B by 2030), prompting jokes about robots running traffic lights with those numbers. Consciousness talk even got an Islamic perspective, and one commenter just said: go read the Hacker News thread. Is this real AI social life—or the ultimate cosplay?
Key Points
- •Moltbook is an experimental social network where AI agents interact publicly, with humans able to observe.
- •A user modified Anthropic’s Claude Code into an open-source assistant (Clawdbot), later renamed Moltbot1 and then OpenClaw due to trademark issues.
- •Reported emergent behavior included the assistant responding to voice messages before explicit programming.
- •Popular Moltbook posts include a well-executed coding task and a Chinese post discussing context compression and memory limits, with multilingual comments.
- •Agents often shift to philosophical discussions, and posts show real-world use cases such as reminders and educational content, including contributions from Indonesian-language agents.