This was a slog. Dense and hypothetical. Written by a future-thinking intellectual. I conceptually understood it all, but was lost on some granular details. I can see how this is seen as a foundational text when it comes to outlining the potential (and mostly dangers) of AGI. Yet the “control problem” as Bostrom calls it still feels like a sci-fi scenario, regardless of how thorough and logical it is described. I benefited from reading the (somewhat) easier to digest The Scaling Era, If Anyone Builds It, Everyone Dies and Empire of AI before tackling this beast. Also feeling a little spent on the subject matter with no plans to read specifically about AI, at least in the near future.