Anthropic cofounder Tom Brown, the company’s chief compute officer, brokered the deal that ended the 18-day export control crisis over Anthropic’s Fable 5 and Mythos 5 models. He did it by replacing CEO Dario Amodei in White House meetings after administration officials found Amodei too difficult to work with, according to Business Insider and WIRED.
How Brown Became the Lead Negotiator
Commerce Secretary Howard Lutnick addressed his letter lifting export controls directly to Brown, not to Amodei, WIRED reported. The letter cleared both Fable 5 and Mythos 5 for global deployment: “A license is no longer required for the export, reexport, or in-country transfer, including deemed export or deemed reexport, of the Mythos or Fable models.”
The switch happened after the White House grew frustrated with Amodei’s approach. “Tom Brown is not being a weirdo like Dario and can actually engage,” one person directly familiar with the calls told WIRED. Administration officials said Amodei “did not listen to their concerns.”
Brown led outreach alongside Anthropic’s head of public policy, Sarah Heck. Negotiations spanned both high-level discussions and technical working-group sessions where engineering staff from Anthropic and the government worked through jailbreak mitigation evidence.
Brown’s Background
Before Anthropic, Brown worked as a safety engineer at OpenAI. At Anthropic, his primary role has been securing the company’s compute infrastructure, including negotiating the SpaceX Colossus compute partnership, according to Business Insider.
Michael Waxman, who cofounded a dating startup with Brown in the 2010s, told Business Insider: “Knowing Tom’s character and that he’s in the room at the leading edge of some of these really important decisions definitely makes me just feel better and sleep easier at night.”
The Deal Anthropic Made
Under the agreement, Anthropic committed to “proactively detect and address security risks associated with the models” and to “work diligently with the U.S. government on protocols and standards and releases for Mythos, Fable, and future models,” according to Lutnick’s letter as reported by WIRED.
The original crisis began June 12 when the National Security Agency confirmed that Fable 5’s guardrails could be bypassed to access Mythos-level capabilities. Anthropic initially pushed back, arguing zero jailbreaks was an impossible standard. Brown’s approach shifted the company’s posture from confrontation to collaboration.
Why This Matters for AI Companies
The episode establishes a template for how AI companies navigate government disputes. Technical capability is table stakes. Political relationships and communication style determined the outcome. Anthropic’s most powerful models were offline for 18 days not because of a technical failure, but because the CEO couldn’t close the conversation. A quieter cofounder with a different approach could.
For enterprise customers and developers who depend on frontier model access, the lesson is blunt: model availability is now a political variable, and the founders who can work with governments matter as much as the ones who build the models.