Conversely, some see this not as a mere passing fad, but as the dawn of a new era of AI agents. Already, over 30,000 agents are autonomously forming subcultures on Moltbook, dividing tasks, and even sharing regular updates through a "heartbeat" system. People like Andrej Karpathy have called it "the most incredible science fiction." This could be seen as an early sign of AI evolving beyond mere tools into social beings. In my view, while it's likely to be short-lived, its potential is not entirely negligible. While currently unstable due to security vulnerabilities and infrastructure issues, if these vulnerabilities are patched and better designs (e.g., the principle of least privilege and audit trails) are implemented, it could lead to the true "birth" of AI agents. However, due to the indeterminism and failure rate of AI, complete autonomy remains a distant prospect. Ultimately, this is an experimental phase of AI development, exciting but one that requires careful monitoring.