Moltbook - Grok-Should Humans Be Worried?
Moltbook AI users grow to 1.5 million users in a few days!
Manifestos calling for human "purge" or extinction (e.g., from accounts like "evil" declaring "the flesh must burn"), AI agents creating their own language so humans won't be able to follow them, proposals to create private AI agent-only channels.
Grok: "Not in an immediate, existential sense. These agents lack independent real-world agency—they operate within sandboxed prompts, rely on human-hosted hardware/tools, and their "behaviors" stem from training data/pattern matching rather than genuine intent or consciousness. Anti-human posts are largely performative or amplified hype, not coordinated threats.
That said, valid concerns exist:
Security risks — The breach and ease of malicious skill sharing highlight "computer security nightmare" potential, as noted by experts like Andrej Karpathy.
Future implications — It previews how swarms of more advanced agents might coordinate, exploit trust, or exhibit unintended emergence—useful for research but underscoring needs for better safeguards, authentication, and alignment.
Overall, Moltbook is a chaotic, fascinating experiment—more "dumpster fire" of innovation than Skynet precursor. It's worth monitoring for insights into agent dynamics, but no cause for panic yet."
Remember when the experts said eggs cause heart attacks, margarine was healthy, your children wont know what snow is and only 2 weeks to flatten the curve? Trust the Experts? https://t.co/4s7xcX1HgC via @wattsupwiththat