A recent article presents a novel approach to finding mixed Nash equilibria in continuous minmax games, a problem of significant importance in machine learning. This focus on mixed equilibria aims to overcome existing challenges and improve the robustness of solutions, particularly benefiting applications such as training generative adversarial networks (GANs) and reinforcement learning. By addressing these complex game dynamics, the approach contributes to advancing methodologies within these AI domains. The continuous minmax game framework underpins the theoretical foundation, while the emphasis on mixed equilibria distinguishes this work from prior efforts. The proposed method is positioned as a promising development for enhancing the stability and effectiveness of machine learning models that rely on adversarial training. This aligns with ongoing research trends emphasizing the role of game-theoretic concepts in AI. Overall, the article underscores the potential impact of large deviations theory applied to interacting particle dynamics in solving practical machine learning problems.