OpenAI released the 'Reinforcement Learning from Text' (RFT) feature on Day 2 of the new product launch event. This feature allows users to use their own dataset to customize the training of AI models using reinforcement learning algorithms, significantly improving the accuracy and expertise of the models in specific domains. This feature can be applied in various fields such as scientific research, law, finance, insurance, engineering, etc., but it is currently in the preview stage and is expected to be released to the public next year. OpenAI has launched the RFT alpha program and invited universities, research institutions, and businesses to participate in testing this feature.