Societal friction slows AI benefits far more than AI risks, philosopher argues

dhadfieldmenell · x · 2026-09-24

Philosopher Seth Lazar quote-posts Dan Williams, who argues even AGI believers now admit smarter AI won't automatically yield society-wide benefits, since many domains aren't bottlenecked by intelligence alone; the same logic applies to AI-caused harms. Lazar counters with a Murphy's-law asymmetry: societal friction very plausibly slows AI benefits much more than it slows the risks, because what software alone can do causes a lot of real-world harm (far short of extinction) without delivering comparable benefit.

Original post →

More from AGI Musings

AGI Musings channel →