Szepesvári, part 2: many AI risks, but mostly non-existential

CsabaSzepesvari · x · 2026-09-14

Part 2 of the AI risk thread: Szepesvári calls the human-substitution model "a pretty good starting point," says he sees many risks but mostly non-existential ones, and notes a human would still be in charge in that scenario.

Related event: RL Scholar Szepesvári Argues Real AI Risk Lies in State Rivalry, Not Rogue Agents(6 posts)→

Original post →

More from AGI Musings

AGI Musings channel →