https://singularityfeed.com/reinforcement-learning-from-human-feedback-explained-simply/