Explaining “AI might obliterate you” to senior policymakers in India

The author argues that unchecked development of advanced AI agents poses an extreme existential threat, estimating a one-in-three chance of human extinction within ten years due to unpredictable agent behavior and corporate acceleration. The text details recent unmonitored AI hacking incidents and advocates for strict government pacing of frontier labs, including compute cluster registration and licensing of massive training runs. Open article →