The AI Safety Schism: Why Theorists Resent Pragmatists Like Anthropic

sebkrier · x · 2026-08-10

The post outlines a fundamental divide in AI safety: one camp believes alignment requires theoretical breakthroughs we lack, while the other treats it as an ordinary ML problem solved iteratively by building capabilities.

Because the pragmatic camp (e.g., Anthropic) advances capabilities to solve safety, they face intense hostility from theorists like Yudkowsky, who argue that Effective Altruists have caused immense harm through overconfidence and poor strategy.

Original post →

More from AGI Musings

AGI Musings channel →