Uploads from Ontology Explained: Philosophy and AI

Currently Playing: An Honest Breakdown of Eliezer Yudkowsky's AI Doom Book

Playlist link: https://www.youtube.com/playlist?list=PLSjZcAM1g_Us In this first episode of my walkthrough of *If Anyone Builds It, Everyone Dies* by Eliezer Yudkowsky and Nate Soares, I lay out the goals and methodology for the series before diving into the book's Introduction: *Hard Calls and Easy Calls*. This isn't a reaction video or a hit piece. It's an attempt to reconstruct the strongest version of the authors' arguments and evaluate them as charitably and rigorously as possible. I've spent years reading Yudkowsky's work and following discussions around AI alignment and existential risk. While I'm deeply skeptical of many of the book's arguments, I also believe that serious claims deserve serious engagement. If the authors are right, the stakes couldn't be much higher. If they're wrong, it's important to understand exactly where the arguments fail. In this episode, we examine: • How the Introduction establishes the stakes of the AI alignment debate. • The book's definitions of Artificial Superintelligence (ASI) and the Alignment Problem. • The analogies and framing devices used to shape the reader's understanding of the problem. • Which claims are relatively uncontroversial—and which are already asking us to make significant philosophical leaps. My goal throughout this series is to be thorough, charitable, and honest. I'll reconstruct the arguments as strongly as I can, but I won't hesitate to point out where I think they succeed or fail. Next episode: Chapter 1 – *Humanity's Special Power*. If you're reading along, now's a great time to read Chapter 1 before the next installment. If you enjoy careful philosophical analysis of AI, alignment, rationality, and argumentation, consider liking the video and subscribing for the rest of the series.