In the Same Post, Amodei Rejects the Regulation-vs-Open-Models Binary and Endorses a FINRA-Style Self-Regulator for AI

TechCrunch·high signalXBlueskyLinkedInCopy link

The second half of Amodei's August 15-16 X post attacks what he calls a false choice between critics who say regulation produces capture and concentration, and accelerationists who say wide distribution including open weights is the only check. His counter is that scaling laws — not rules — are what structurally concentrate power, and that decentralizing institutions like the court system are the actual counterweight; he backs a FINRA-like industry body, an idea Demis Hassabis floated in July. It is consistent with his July 27 clarification that Anthropic 'has never advocated for a ban on open-weights models' while still wanting mandatory safety testing for any sufficiently capable model, open or closed.