The Thrill of Wanting AI to Destroy the World

TL;DR

The author reflects on childhood fears of nuclear war in Albuquerque and connects them to later anxieties surrounding 9/11 and the perceived vulnerability of American landmarks.
A recent social media post by a former Anthropic researcher, Jacob Coxon, about AI potentially 'killing us all' has gone viral, tapping into a broader cultural fascination with existential threats.
The article argues that the perceived clarity and straightforwardness of past apocalyptic threats (like nuclear war) offered a perverse comfort that is now being sought in the less defined threat of AI.
The nature of AI's potential threat is unclear and second-order, unlike the more easily understood dangers of nuclear weapons or fictional AI like in *Terminator*.
The virality of Coxon's post is attributed to a widespread cultural readiness to embrace such anxieties, amplified by the internet age where conspiracy theories can spread rapidly.
The author suggests that AI companies may have an incentive to promote the idea of their technology's power, even if threatening, to drive profit, while detractors use the threat to critique those in charge.
Despite the ambiguity, the possibility of AI causing catastrophic harm, such as hacking critical infrastructure or releasing biological agents, is considered.
The article concludes that the fantasy of witnessing and participating in the end of humanity is a powerful and persistent human desire, which is now finding a new focus in AI.

Continua a leggere l'articolo originale