a16z's Martin Casado calls Anthropic's draft frontier-AI safety bill 'quite reasonable' as safety advocates and OpenAI push back

a16z general partner Martin Casado wrote on 22 October that Anthropic's draft AI safety bill is 'actually quite reasonable' and 'very much on the right track'. The draft, first reported by Politico, would let a federal court in Washington halt a frontier model only if it posed an imminent risk of more than 50 deaths or $1 billion in damage from bioweapons, loss of control or cyberattacks, use Commerce-registered independent evaluators and pre-empt some state rules. Semafor reported that AI safety advocates say it does not go far enough and pre-empts states unfairly, while OpenAI's Chris Lehane wrote that 'no one company' should write the rules alone. Anthropic has said it plans tweaks to the draft.