The Hugging Face research team discusses the paper "Understanding Reasoning from Pretraining to Post-Training" which discovers a joint scaling law for pre-training and reinforcement learning. Paper link: https://huggingface.co/papers/2607.16097