The White House is finalizing an AI framework and classifying it, with a meeting today attended by most major AI companies. The government aims to balance regulatory oversight of frontier models—addressing cybersecurity and US-China AI policy—against industry concerns, particularly regarding US companies' use of Chinese open-source models. Nvidia assembled over forty US companies, including OpenAI and Google, to sign on, with Anthropic as the exception, advocating for regulation on cybersecurity, safety, and China. The executive order designates the benchmarking process and coverage thresholds as classified, meaning the rules exist but the tests and bar are secret, a bar only the largest labs can afford to aim at. This friction is likely to slow the US down on AI globally over the longer term. Separately, Apple and OpenAI are trading arguments in public, and Elon Musk faces a decision on merging SpaceXAI and Tesla. In all three cases, the people with the most at stake have the least ability to settle matters on their own. The government is turning what was an Autobahn for the US tech industry into places with occasional traffic lights and thirty mile an hour school zones, as seen during The Blip 2.0, when both Anthropic and OpenAI were paused from releasing their latest models for almost three weeks before the pause was lifted.