DeepSeek’s R1 and OpenAI’s Deep Research They’ll push companies to use techniques like distillation, supervised fine-tuning (SFT), reinforcement learning (RL) and retrieval-augmented generation (RAG) to build smarter, more specialized AI applications