Sacks lost his AI czar role as the Trump White House reversed course on AI deregulation, driven by Anthropic’s Mythos leak and geopolitical pressure.
Key Takeaways
Anthropic’s unreleased Mythos model, capable of finding critical cybersecurity vulnerabilities, alarmed Treasury, Commerce, and the NSC into demanding pre-deployment oversight.
Commerce Department designated CAISI (under NIST) to conduct pre-release testing on frontier AI models; xAI, Microsoft, and Google DeepMind have already signed agreements.
NIST was gutted by DOGE in 2024 but is now actively rehiring for technical positions as its authority is restored.
Sacks twice failed to pass a congressional moratorium on state AI laws, then tried an executive order workaround that alienated GOP state lawmakers who went on record to the WSJ.
Iran’s drone strikes on two AWS data centers in the UAE, with damage severe enough to disrupt Middle East operations for months, reframed AI infrastructure as geopolitical target.
