Benjamin Bratton: we need pro-diffusion third-party AI model evaluators
bratton · x · 2026-09-13
Benjamin Bratton argues there is a clear need for pro-diffusion third-party model evaluators: if evaluation is dominated by restriction-leaning regulators or the labs themselves, society's risk/reward calculus around AI will get horribly distorted. He calls for independent evaluators oriented toward spreading AI's benefits to counterbalance today's conservatism-leaning evaluation ecosystem.
More from Safety
- Dario Amodei's 3,900-word essay calls on the AI industry to deliberately slow frontier model progress — sanjaykalra · 2026-09-13
- Dario Amodei calls for slowing AI capability gains; Altman and Musk respond — sanjaykalra · 2026-09-13
- Anthropic's policy details exactly when your Claude chats are used for model training — austinc3301 · 2026-09-13
- Counterpoint: Anthropic's training opt-out language looks ironclad — austinc3301 · 2026-09-13
- Ex-Anthropic Safety Lead: RSI Hype Ignores Amdahl's Law Human Bottlenecks — joshua_saxe · 2026-09-13
- Open weights are public information, so regulating compute is the workable path — willcb · 2026-09-13