ENTR 810 | Session 11 | AI: Von Neumann to Altman Hardware Evolution

Introduction:

The history of Artificial Intelligence (AI) ranges from its foundational theories to its present role as a transformative technology, illustrating the close connection between software models and the hardware that supports them. Early pioneers like Alan Turing, who helped crack the Enigma code and introduced the Turing Test for machine intelligence, and John McCarthy, who coined the term "Artificial Intelligence" at the 1956 Dartmouth Workshop, laid the groundwork for the field. John von Neumann also made significant contributions by imagining adaptive systems and defining the "von Neumann Architecture" for computer design, which remained the standard into the 21st century. After challenging periods such as the "AI Winter," the field saw a resurgence fueled by deep learning—a method that uses deep neural networks and improves predictably as more compute power and data become available. Today's AI boom relies heavily on massive computational infrastructure, with GPUs (Graphics Processing Units) serving as the "heartbeat of AI." Nvidia capitalized on this shift, transitioning from a gaming company to the key provider of AI acceleration hardware, offering GPUs and CUDA libraries. Breakthroughs in software, especially the 2017 introduction of Transformers by Google Brain, enabled models to learn context and track relationships in sequential data, leading to the development of the Generative Pre-Trained Transformer (GPT) and the landmark release of ChatGPT in 2022. Although this progress is expected to usher in an "Age of Intelligence" characterized by widespread prosperity, personalized assistants, and solutions to complex problems, experts also raise serious concerns about ethics, potential disruptions to labor markets, and the fundamental risk that AI goals could diverge from human values, or that humans might weaponize the technology.