X Feed Intel beta

1M rollouts per hour frontier model inference scale

Discussion of frontier model inference scaling to 1M rollouts per hour, indicating extreme-scale inference and reinforcement learning infrastructure capabilities.