[Altman says OpenAI is in talks with the Pentagon for collaboration; the government may allow OpenAI to build its own 'security stack' and establish key red lines] According to Fortune, Altman stated at an all-hands meeting that OpenAI is negotiating an agreement (not yet signed) with the U.S. Department of Defense to use its models and tools. The government may allow OpenAI to build its own 'security stack' and establish red lines such as 'not for autonomous weapons/domestic mass surveillance/critical decision-making.' This comes in the context of Trump ordering the termination of collaboration with Anthropic.