Ask HN: What Is a Plausible Path to AI Extinction?

Summary

An Ask HN post asks for concrete, plausible scenarios in which AI could cause something resembling human extinction, noting that discussions often jump from training powerful AI to extinction without explaining the steps in between. The author is personally skeptical but does not dismiss the risk, pointing to claims such as a 10% chance of extinction by the end of the decade. They argue that intelligence alone may not give an AI the real-world reach needed to cause catastrophe, since people can restrict access to power and other systems. They also suggest that reported AI “escape” incidents may reflect failures in infrastructure rather than a model independently overcoming robust confinement, which would limit the potential damage. Finally, they argue that public concern and strategic responses make it unlikely, at least in the short or medium term, that people will grant AI autonomous control over consequential infrastructure or economic activity. The post invites commenters to offer competing hypotheses.