The U.S. government is poised to exert unprecedented control over AI model releases, potentially halting Anthropic's Fable and Mythos models indefinitely and limiting OpenAI's new GPT 5.6 to a customer-by-customer preview.

This stringent oversight could stifle the rapid development of costly new systems, particularly as AI labs struggle to boost their bottom lines. A slowdown in model creation could have far-reaching implications for the industry's growth and innovation.

The tech world is divided, with some accusing Anthropic of regulatory capture and others alleging OpenAI’s cozy relationship with Trump. However, these issues are now overshadowed by a broader concern: how to establish a sensible release process without stifling progress or public access.

As Dean Ball argues, the current approach lacks clear safety assurances and tests beyond the government's expertise. The real challenge is finding a balance that addresses genuine concerns while allowing the industry to thrive. Collaboration and trust in independent groups will be key to navigating this complex landscape.