Anthropic’s Draft AI Safety Bill Sets a $1B or 50-Person Bar 🚨

Anthropic pitched a narrow AI safety bill. Gov can only block models posing >$1B damage or >50 injuries via bio, loss of control, or cyber risks.