The recent developments in the research area of summarization and discourse relation annotation are pushing the boundaries of existing methodologies, particularly in leveraging large language models (LLMs) for more efficient and effective solutions. A notable trend is the creation and utilization of large-scale datasets, such as FineSumFact and RoLargeSum, which are enhancing the training and evaluation of summarization models, especially for non-English languages. These datasets are enabling more nuanced and context-aware summarization, which is crucial for diverse linguistic and cultural contexts. Additionally, there is a growing focus on event-centric summarization, exemplified by EventSum, which addresses the challenge of summarizing dynamic events across multiple documents, introducing new metrics to better evaluate the comprehensiveness of summaries. The field is also witnessing advancements in discourse relation annotation, with studies like the one on Crowdsourcing Task Design for Discourse Relation Annotation highlighting the impact of task design on annotation diversity and accuracy. Furthermore, frameworks like PerSphere are emerging to tackle the issue of echo chambers by facilitating multi-faceted perspective retrieval and summarization, emphasizing the importance of breaking free from information silos. In the realm of event relation extraction, innovative approaches such as LogicERE are being developed to enhance the logical coherence and reliability of event graphs through high-order reasoning and logical constraint modeling. Overall, these developments are collectively advancing the state-of-the-art in summarization and discourse analysis, paving the way for more sophisticated and contextually aware NLP applications.