On this week’s episode, I’m joined by Nate Soares to talk about his new book, cowritten with Eliezer Yudkowsky, If Anyone Builds It, Everyone Dies: Why Superhuman AI Would Kill Us All. It’s a fascinating book—some will say fearmongering and sensationalist; I, frankly, think they’re overly optimistic about our ability to constrain the development of general intelligence in AI—in large part because of how it’s structured. Each chapter is preceded by a fable of sorts about the nature of intelligence and the desires of intelligent beings that look and think very differently from humans. The point in each of these passages is less that AI will want to eliminate humanity and more that it might do so incidentally, through natural processes of resource acquisition. 

This made me think about how AI is typically portrayed in film; it is all too often a Terminator-style scenario, where the intelligence is antagonistic in human ways and for human reasons. We talked some about how storytellers could do a better job of thinking about AI as it might actually exist versus how it might be like us; Ex Machina is a movie that came in for special discussion due to the thoughtful nature of the treatment of its robotic antagonist’s desires. If this episode made you think, I hope you share it with a friend!