Why a superintelligence wouldn't kill us: the dog-argument against Yudkowsky's doom case

dbasch · x · 2026-09-15

dbasch pushes back on Yudkowsky's claim that AI MUST kill everyone: if a superintelligence is that smart, humans are to it like dogs or ants — and we don't exterminate dogs merely because they might pose a threat; leaving them be is far easier.

His second point: unlike the rest of the universe, a superintelligence was specifically instructed not to kill us. It could if it wanted, but it is the one force explicitly constrained against it — undermining the inevitability argument.

Original post →

More from AGI Musings

AGI Musings channel →