X-risk debate: 'superintelligence is unimaginable' argument cuts both ways, says critic

connoraxiotes · x · 2026-09-10

An x-risk debate: the author argues that if x-risk proponents claim a misaligned superintelligence could kill humanity in ways we can't comprehend because it's smarter than us, then demanding a fully fleshed-out human-to-human kill plan isn't a compelling rebuttal of AI x-risk by 2030. He adds that the opponent's own point — increasingly powerful AIs may harm humans without intending to — actually supports the side of caution.

Original post →

More from AGI Musings

AGI Musings channel →