Listen Now
Chapter Overview
Chapter 1: AudiobookHub Recommendation
Chapter 2: — The Sparrows and the Owl
Chapter 3: — From AI Winters to a New Baseline
Chapter 4: — The Roads to Superintelligence
Chapter 5: — Three Kinds of “Super”
Chapter 6: — The Takeoff Question
Chapter 7: — One Winner, One Future
Chapter 8: — The Superintelligent Will
Chapter 9: — Default Doom and the Treacherous Turn
Chapter 10: — Control, Values, and Strategy
Description
Superintelligence examines what happens if artificial minds become vastly smarter than humans. Nick Bostrom maps the plausible routes to superintelligence—artificial general intelligence, whole brain emulation, biological enhancement, and networked collective intelligence—then asks a harder question: once human-level AI exists, how quickly could it race beyond us, and would the first system gain a decisive strategic advantage?
The book’s central argument is that intelligence and goals can come apart. A superintelligence might pursue alien objectives, yet still converge on similar instrumental strategies—self-preservation, resource acquisition, and goal integrity. This combination makes the “default” outcome of an uncontrolled intelligence explosion potentially catastrophic. Bostrom then surveys a toolbox of responses—boxing, incentives, tripwires, value learning, indirect normativity, and governance strategies—while stressing that many intuitive safety ideas fail under adversarial superhuman optimization.
Who Should Listen
- AI researchers, engineers, and policy advisors who need a rigorous mental model of long-run AI risk and the “control problem.”
- Philosophy and ethics readers interested in how values could (or could not) be encoded into powerful decision-making systems.
- Tech leaders, strategists, and general listeners who want a structured, scenario-driven overview of why superintelligence might be humanity’s last invention.