Planning 80000 hours at the Plausible End of the World

A university student's introspective Substack post, titled 'Planning 80000 hours at the Plausible End of the World,' has resonated widely, articulating a profound generational dilemma. The author grapples with the near-impossibility of traditional career planning when facing potential AI-driven existential shifts—contemplating scenarios where machine superintelligence arrives before graduation or an intelligence explosion compresses millennia of progress into years. The post references economist Robin Hanson's concept of 'The Dream Time,' a precarious historical moment where a few key choices about technology could permanently alter the future, framing the current era as a legendary, high-stakes period for potential heroes and tragedies.

The core of the post is a pragmatic, if anxious, attempt to apply rational decision-making to this uncertainty. The author constructs a simplified 'expected value calculation' to evaluate career paths under different AI timeline probabilities. This involves weighing options like Path A—dropping everything to upskill in machine learning and contribute to AI alignment immediately—against Path B—staying in school to build long-term foundations. The analysis acknowledges that any choice might render one's impact negligible, but argues that inaction guarantees it. This framework has sparked discussion by giving structure to the vague anxiety many in tech and academia feel about how to productively navigate a future that seems both limitless and potentially vanishing.