Modern Feudalism with AGI – Part I: The History of AI and why regulation is needed

The document explores the evolution of Artificial Intelligence (AI) from its origins in neuroscience to its current state, emphasizing the need for regulation in the age of Artificial General Intelligence (AGI). ​ It traces AI’s history, starting from the 1956 Dartmouth Summer Research Project funded by the Rockefeller Foundation, which marked the formal beginning of AI as an academic discipline. ​ Early developments like the perceptron and explanation-based AI (EBAI) faced technological and algorithmic limitations, leading to periods of stagnation known as AI winters. ​ However, breakthroughs such as backpropagation, deep learning revival by Geoffrey Hinton, and the advent of GPUs by Nvidia enabled the modern era of AI, characterized by large-scale models like AlexNet and transformers. ​ The document highlights the importance of data, compute, and algorithms in AI development, with data being the most critical factor for model performance. ​

The text also discusses the global race for AI dominance, with countries like Canada, India, and Japan investing in sovereign AI solutions to maintain control over data and infrastructure. It contrasts foundational models, which are general-purpose and resource-intensive, with vertical AI, which focuses on domain-specific applications for faster returns. ​ The document emphasizes the potential of AGI to revolutionize productivity by decoupling output from human input, leading to exponential growth and a shift in economic dynamics. ​ However, it warns of the risks of wealth concentration and diminishing labor value in an AGI-driven economy, likening the future to modern feudalism. ​ The need for AI regulation is underscored, with compliance tools suggested as a way to monitor and manage the ethical and economic implications of AI advancements. ​