Author slams Anthropic's 'pacing' framing as cover for unconsented labor-replacement race
GarrisonLovely · x · 2026-11-19
AI-risk author Garrison Lovely sharply criticized Anthropic's "Pacing" role description. The attached screenshot defines Pacing as Anthropic's work ensuring that "as AI capabilities advance," enough time and effort goes into making them safe.
Lovely's core objections:
- The passive phrase "as AI capabilities advance" hides the actor — these companies are advancing capabilities themselves, a euphemism for "trying to build universal labor-replacing machines without the consent of the public or any guarantee the work will be done safely";
- The definition is logically incoherent and leaves no opening for actually slowing capabilities;
- No company should be counted on to do the right thing; stopping the industry is up to the public. He lays out plans in long and short versions, tied to his book Obsolete (an FT Best Economics Book of 2026), which frames the AI industry as a trillion-dollar race to replace human labor.
More from AGI Musings
- MIT CSAIL paper: frozen LLM agents design their own self-improvement without being told how — omarsar0 · 2026-11-19
- Anthropic Says Claude Solved Two Open Problems in Black Hole Physics — ghumare64 · 2026-11-19
- Reuters: China distills US model outputs to close AI gap despite chip export controls — r0ck3t23 · 2026-11-19
- Gary Marcus amplifies warning of an AI bubble chain reaction into a financial crisis — GaryMarcus · 2026-11-19
- Read More Facts and Fewer Theories: Why Real Experts Memorize Specifics — panickssery · 2026-11-19
- Dario Amodei calls open-source model path dangerous; critics slam it as a control grab — adamamcbride · 2026-11-19