The field of video understanding and retrieval is rapidly advancing with a focus on improving the accuracy and efficiency of video analysis. Recent developments have seen the introduction of new frameworks and models that enhance the ability to comprehend and retrieve video content. One of the key areas of research is the development of multimodal large language models (MLLMs) that can effectively integrate visual and textual information to improve video understanding. These models have shown significant improvements in tasks such as long video comprehension and action localization. Another area of focus is the development of more efficient and effective methods for video retrieval, including the use of temporal fusion and pivot-based approaches. Noteworthy papers in this area include the proposal of Nar-KFC, a plug-and-play module for effective and efficient long video perception, and DisTime, a lightweight framework that enhances temporal comprehension in Video-LLMs. Additionally, the introduction of new datasets such as VidEvent and TF-CoVR has provided valuable resources for researchers to develop and evaluate new models and methods. Overall, the field is moving towards more accurate and efficient video analysis, with a focus on developing models and methods that can effectively integrate multiple sources of information and handle complex video content.