Superintelligent Agents Pose Catastrophic Risks: Can Scientist AI Offer a Safer Path? Paper • 2502.15657 • Published 19 days ago • 5