The world’s first nuclear test detonation, the “Trinity” test, happened on July 16, 1945. To date, only two nuclear bombs have been used in war, both of these by the United States, within a month of the Trinity test, against the Japanese cities of Hiroshima and Nagasaki. These bombs killed an estimated 350,000 civilians.

Deterrence theory says that the (relative) world peace of the past 80 years, the most peaceful years in human history, is a result of nuclear deterrence.

(Note: Deterrence theory is contested, especially by anti-nuclear advocates, and there is no expert consensus on to what extent deterrence explains the peace that has occurred.)

With the arrival of AGI and the technological advances it precipitates, however, the nuclear deterrence period may come to an end. From Leopold Aschenbrenner’s “Situational Awareness: The Decade Ahead” report:

The military advantage would be decisive even against nuclear deterrents.

To be even clearer: it seems likely the advantage conferred by superintelligence would be decisive enough even to preemptively take out an adversary’s nuclear deterrent. Improved sensor networks and analysis could locate even the quietest current nuclear submarines (similarly for mobile missile launchers). Millions or billions of mouse-sized situational awareness 130 autonomous drones, with advances in stealth, could infiltrate behind enemy lines and then surreptitiously locate, sabotage, and decapitate the adversary’s nuclear forces. Improved sensors, targeting, and so on could dramatically improve missile defense (similar to, say, the Iran vs. Israel example above); moreover, if there is an industrial explosion, robot factories could churn out thousands of interceptors for each opposing missile. And all of this is without even considering completely new scientific and technological paradigms (e.g., remotely deactivating all the nukes).

“Superintelligence,” which Aschenbrenner talks about, is most commonly defined as artificial intelligence that is better than the best human experts in all domains. Economist Tom Davidson estimates that superintelligence will come within a year of AGI (where he defines AGI as AI that can automate 100% of human labor). Meanwhile, the current community prediction on the Metaculus question “After a (weak) AGI is created, how many months will it be before the first superintelligent AI is created?” is 40 months.