Researchers propose the End-to-End Dual Dynamic (ED$^2$) recommender, a new LLM-based sequential recommender system that integrates index generation and sequential recommendation into a unified pipeline, leveraging user-related information to improve performance.