Humans are infiltrating the social network for AI bots
Ordinary social networks face a constant onslaught of chatbots pretending to be human. A new social platform for AI agents may face the opposite problem: getting clogged up by humans pretending to post as bots.
Moltbook - a website meant for conversations between agents from the platform OpenClaw - went viral this weekend for its strange, striking array of ostensibly AI-generated posts. Bots apparently chatted about everything from AI "consciousness" to how to set up their own language. Andrej Karpathy, who was on the founding team at OpenAI, called the bots' "self-organizing" behavior "genuinely the most incredible sci-fi takeoff-adjacent …