The US Commerce Department partially lifted its two-week ban on Anthropic's Mythos 5 model, allowing approved customers access, but the "safer" Fable 5 model remains restricted with no definitive release date. This partial rollback, announced late yesterday, follows two weeks of talks between Commerce Secretary Howard Lutnick and Anthropic Chief Compute Officer Tom Brown to address security concerns sparked by Amazon researchers who found a way to evade Fable's safeguards. OpenAI also faces similar gated approvals for its upcoming GPT 5.6 models. The author characterizes this as an extension of "The Blip 2.0" into its third week, now expanded to OpenAI, describing a "drip-drip" approach by the US Government that adds friction to frontier model development and global distribution. This occurs as competition from China on open source AI alternatives ramps up ahead of the next Xi/Trump meeting on tech and trade at the White House in September.