The field of video anomaly detection is moving towards more complex and generalizable solutions. Researchers are exploring hybrid approaches that combine the strengths of different models and techniques to improve detection accuracy and robustness. Notably, the integration of large language models and self-supervised learning methods is showing promising results. Additionally, there is a growing interest in developing more interpretable and explainable models, with techniques such as fine-grained prompting and game-theoretic fusion methods being proposed. These advances have the potential to improve the reliability and effectiveness of video anomaly detection systems in various applications, including intelligent surveillance and autonomous driving. Noteworthy papers include: HyCoVAD, which introduces a hybrid SSL-LLM model that achieves state-of-the-art performance on complex video anomaly detection tasks. PANDA, which proposes an agentic AI engineer based on MLLMs that can automatically handle any scene and any anomaly types without training data or human involvement. Strategic Fusion of Vision Language Models, which presents a game-theoretic fusion method for multi-label understanding of ego-view dashcam video that improves the reliability and accuracy of vision-language models in autonomous driving pipelines. Unlocking Vision-Language Models for Video Anomaly Detection, which proposes a structured prompting framework that leverages action-centric knowledge to elicit more accurate and interpretable reasoning from frozen VLMs.