Tom and Nate discuss some of the most dominant metaphors in machine learning these days -- alchemy and deep learning's roots, the Oppenheimer film and a modern "Manhattan Project for AI", and of course, a sprinkle of AGI.