How Private Institutions Can Solve AI Tail Risk | Brendan McCord

McCord argues private liability, mutual insurance, and narrow public rules can govern AI safety faster and better than regulatory "pacing."

• Critiquing Bostrom on pacing: McCord says Nick Bostrom is now "not sure of the sign" of his own intervention in founding AI safety thinking, yet still treats pacing proposals as unambiguously good.
• Pacing's core flaw: McCord argues pacing is ill-defined (compute today, algorithms tomorrow) and any "pacer" capable of stopping state/non-state defectors would need coercive power with no defined limits.

Themes, sentiment and reasons are AI-written paraphrases, checked so they never quote an episode. They may contain errors. Not investment advice. Episodes remain the property of their creators.