The recent advancements in recommender systems and language models (LLMs) have shown a significant shift towards enhancing both the ethical considerations and the personalization capabilities of these systems. Researchers are increasingly focusing on developing models that not only provide accurate recommendations but also align with democratic values and user sentiments. The integration of LLMs into recommender systems has opened new avenues for generating more personalized and explainable recommendations, addressing issues such as data sparsity and user privacy. Notably, there is a growing emphasis on moving beyond traditional user-centric models to post-userist approaches that consider broader stakeholder relationships and the ethical implications of recommendation algorithms. Additionally, the field is witnessing a push towards more transparent and fair evaluation methods, with a particular interest in pessimistic evaluation to ensure that systems perform well across all user segments, not just on average. These developments collectively aim to create more equitable and user-friendly recommendation environments, reflecting a deeper understanding of the societal impact of these technologies.