frontier model security

OpenAI announced a deliberate slowdown in AI development, including a two-week halt on reinforcement learning training for deployment-bound models, to…