4.0

Average Course Rating

Ghost in the Machine traces the history of one of humanity’s most powerful and most consequential ideas: that humans could build machines that surpass themselves, in all faculties, and thus outwit our maker, ultimately cheating death and disease. This idea has become the best-funded futurist project of our time. We will investigate how mythical thinking about intelligent machines has evolved since World War II, what effects such thinking has had on the development of intelligent machinery, and how breakthroughs in neural networks, deep learning, and transformers over seventy years have shaped our capacity to foresee and manage the future of artificial—and human—intelligence. We will trace seven critical ideas: extinction, salvation, the intelligence explosion, training, black boxes, agents, and a curious vanishing act. The detailed historical reconstruction of innovation reveals how the technology actually works: students will acquire familiarity with some of the core concepts of artificial intelligence, such as neural networks, gradient descent, false peaks, vectorization, weights, backpropagation, self-attention, transformers, the scaling laws, interpretability, feature extraction, circuits, superposition, and alignment trends. Students will also gain appreciation for the key thinkers, innovators, and engineers who have shaped AI — its technology, its ideology, and its politics. No math or technical background is required. Rid will use this class as an experimental space to demo cutting edge AI-tools that enhance and accelerate the research and learning process — as they are getting updated and released.