AI safety researcher maps the two worlds where we don't all die from AI

DavidSKrueger · x · 2026-09-12

Cambridge researcher David Krueger quotes Tyler John on p(doom): "doom" literally means every human killed — the most extreme case. If you put 10% on that, you probably put 30% on catastrophes killing billions and destroying the US government.

Krueger argues the worlds where we don't all die come in only two flavors: a superintelligent AI takes over but happens to be nice, or society gets its act together and stops letting us build dangerous systems. Notably, the societal response is baked into both scenarios.

Original post →

More from AGI Musings

AGI Musings channel →