The field of reinforcement learning is moving towards integrating foundation models to improve sample efficiency and decision-making in complex environments. Recent research has focused on leveraging the prior knowledge and reasoning capabilities of foundation models to enhance reinforcement learning agents. This integration has shown promising results in various applications, including climate risk assessment and adaptive forest management. Notably, the use of foundation models has enabled the development of more effective exploration strategies and improved performance in sparse-reward settings. However, challenges remain, such as addressing the 'knowing-doing gap' in foundation models and developing more robust multi-objective reinforcement learning methods. Noteworthy papers include: Foundation Models as World Models, which evaluated the use of foundation world models and foundation agents in grid-world environments. BoreaRL introduced a multi-objective reinforcement learning environment for climate-adaptive boreal forest management, highlighting the challenges of optimizing forest management for both carbon sequestration and permafrost preservation.