Understanding the Connection Between AI and Historical Governance
The article explores the intriguing parallels between the structure of modern AI models and historical governance systems, particularly the Consell de Cent in Catalonia. This council was a governing body that advised on local matters. The discussion highlights how contemporary AI models, especially those using Mixture-of-Experts (MoE), are increasingly composed of distinct parts that work together, much like a council of experts.
Key Insights
- The Consell de Cent, established in the late medieval period, managed governance in Catalonia until the early 1700s.
- Modern AI models are becoming larger and more complex, resembling the partitioning of hardware into virtual machines.
- Mixture-of-Experts models use multiple sub-networks, where only a few are activated for specific tasks, enhancing efficiency.
- The concept of a “society of the mind” introduced by Marvin Minsky parallels the collaborative nature of these AI models, suggesting that intelligence arises from the interaction of specialized components.
The Importance of This Evolution
This evolution in AI is significant as it mirrors historical methods of organization and decision-making. By drawing from the past, AI can enhance its efficiency and performance. The ability to utilize a mixture of specialized models allows for better resource management and reduces computational costs. Understanding these historical connections can provide valuable insights into the future of AI development and its potential impact on decision-making processes.
