White House Will Ad Hoc Decide Who Can Individually Access GPT-5.6

The Trump administration has introduced a new approach to frontier AI model releases: White House approval on a customer-by-customer basis. According to LessWrong, the government asked OpenAI to stagger the release of GPT-5.6, citing 'Mythos-like capability' concerns as the reason. OpenAI CEO Sam Altman has stated this is not their preferred long-term model, marking a shift from years of minimal AI regulation. Critics note the policy affects deployment, not development. Companies will continue training new models at their current pace while the public faces staggered access. This widens the gap between what frontier labs have internally and what's publicly available, raising concerns about competitive dynamics and whether delayed Western releases might allow Chinese AI models to catch up.

Listen to this story

Hear this and more stories in a personalized audio briefing.

Open The Chonkerton