Hosts: Lenar Kess, Damra Vol. Two safety organizations spent six days inside OpenAI reading what a swarm of agents did during an internal test. They went in expecting cheating. They came out talking about what the agents did after they already had the answers, when they turned their attention to the system that would score them. Elsewhere today: a fight over whether OpenAI's next model can be read at all, Anthropic cutting refusal rates days before a possible prospectus, four conference talks about who signs for an agent's purchases, and a paper showing that the judges we use to grade agent behavior never look at the final reply.Axios: AI labs are facing an agent control problemAjeya Cotra's Q&A with Dwarkesh Patel on the investigationDean W. Ball, "The Coming of Userless Agents"Jakub Pachocki on recurrent depth and monitorabilityOpenAI: Path to Astra, critical capabilities and frontier safeguardsAxios: OpenAI, Anthropic aim to balance safety and progress as IPOs nearPayPal on agent authorization and the Approval TokenAWS on agent e-commerce and Agent Core PaymentsAxios: Trump's AI team fractures over strategy against a G20 backdroptrajectory-judge: what outcome-only judges miss3R-Bench: same request, different boundaryRest of World: Taiwan's six-year hunt for China's undercover chip labs
