The field of artificial intelligence is witnessing significant developments in multi-agent collaboration and large language models. Researchers are exploring innovative ways to enable efficient communication and coordination among agents, with a focus on scalability and reliability. The use of lightweight structure and schema-induced games is being investigated as a means to steer convention formation and improve agreement among agents. Meanwhile, the integration of large language models with embodied systems, such as self-driving cars and service robots, is being explored to enhance operational efficiency and maximize cooperative mechanisms. Noteworthy papers in this area include: Debate2Create, which introduces a framework for robot co-design via large language model debates, yielding diverse and specialized morphologies. The Geometry of Dialogue, which proposes an interaction-centric framework for automatic team composition that discovers synergistic model clusters and outperforms random baselines on downstream benchmarks.