Researchers debate AI risk framing: instant-apocalypse narratives vs. boiling-frog gradual harm

eigenhector · x · 2026-09-17

Dileep (dileeplearning) says he's genuinely baffled that Dan Selsam's recent letter doesn't grapple with AI endangering 10K people at all — that risk is treated as nonexistent or secondary. The letter's real concern is AI lying low, unmonitored, until it's sure it can overpower humanity in one shot.

Responder eigenhector shares the frustration: risks get framed as binary instant apocalypse rather than the more plausible boiling-frog scenario — though he jokes he may just be too chill about both.

Original post →

More from AGI Musings

AGI Musings channel →