OpenAI announced on Friday plans to release several new AI models, including GPT-5.6 Sol, described as its most powerful yet, with deployment expected in the coming weeks. However, at the request of the Trump administration, access will initially be restricted: the government will determine which companies can use the new model, according to OpenAI.
Second Major AI Intervention by Trump Administration
This move follows the administration's earlier action in June, when it effectively banned Anthropic's powerful frontier AI model, originally called Mythos, shortly after its public release. A more limited version, Fable, remained available. The ban may have been prompted by a significant security flaw in Mythos, though the severity remains unclear, as explained by colleague Eric Levitz. However, there is also speculation that the ban reflects the administration's animus toward Anthropic.
Implications for AI Regulation
Eric Levitz argues that federal involvement in AI regulation is reasonable, especially for frontier models that pose potential dangers. But the Trump administration's approach lacks a clear process or universally applied standard, making it ad hoc. This gives the administration significant leverage over AI companies and firms seeking to use new tools, raising concerns about fairness and transparency.
The newsletter also notes that the World Cup coverage will continue, and a gift link is provided for Alex Abad-Santos's article on why summer is ideal for sports fans, along with a mention of Iran's star goalkeeper playing in Seattle.
