AI safety and 'normal technology' camps converge on policy despite worldview gaps

dhadfieldmenell · x · 2026-09-17

Nathan Calvin notes growing convergence between the LawAI and 'AI as a normal technology' camps around figuring out which traditional governance tools work in novel circumstances. The quoted discussion praises Sayash Kapoor and Arvind Narayanan's epistemic humility, highlighting the line 'we do not need consensus on worldviews to have agreement on policy' — such consensus is likely impossible, but many policies are robust across assumptions because they focus on managing uncertainty and building capacity.

Original post →

More from AGI Musings

AGI Musings channel →