Seán Ó hÉigeartaigh stated agreement with a post by Jack Clark and wrote that he believes the point and wants to make it happen. The post concerns Anthropic. A generated headline on the topic reads Anthropic Updates Alignment And Security Practices After Model Incidents. An accompanying generated source summary states that Anthropic detailed new steps to secure evaluation environments and harden infrastructure ahead of more advanced models and that the company also shared research on the subject.