Researcher: Continued AI scaling cannot reliably avert disasters

Andy Masley argues that continuing to scale AI models cannot reliably prevent catastrophic behavior, and that this risk narrative is credible enough to deserve mainstream debate despite unknown unknowns.

2026-09-05 ~ 2026-09-05 · 3 related posts

• Drop the word alignment and the answer is no: scaling AI still risks catastrophe — AndyMasley · 2026-09-05
• If we keep scaling AI, can we reliably avoid catastrophe? Not right now — AndyMasley · 2026-09-05
• AI catastrophe risk is now plausible enough for mainstream debate — AndyMasley · 2026-09-05