Inference time scaling means whoever had the most compute has the highest intelligence model.