The "classic" BDI model supports communication (and thus a form of coordination) between agents based on message exchange, but what if agents could share their beliefs, in sich a way that whenever an agent perceives something, this information becomes immediately available to the rest of the agents in the system? In which way this would improve the system's performances, and which new choices would become available for its design?
Hive-BDI is an extension to the classic BDI model which provides agents with this capability, allowing the MAS to act as if it were a single entity, agents being minor entities serving a collective mind.