Decoder host Nilay Patel discusses with senior AI reporter Hayden Field the chaotic weekend when the Trump administration imposed export controls on Anthropic's new Claude Fable 5 and Mythos 5 models after a jailbreak concern from Amazon. The 90-minute ultimatum forced Anthropic to pull both models offline, revealing an unpredictable regulatory approach that could undermine U.S. AI leadership and expose companies to political risk.
Summarized by Podsumo
The Trump administration gave Anthropic a 90-minute ultimatum to comply with export controls, forcing the company to pull its new Claude Fable 5 and Mythos 5 models offline after Amazon CEO Andy Jassy flagged a potential jailbreak.
The export controls bar foreign nationals from accessing the models—even Anthropic employees—which could drive global customers to seek non-U.S. alternatives, undermining American AI leadership.
Anthropic’s years-long campaign to hype model dangers as a push for regulation backfired, as the government now uses that same alarmism against the company.
Unlike rivals like OpenAI or Google, Anthropic’s tense relationship with the Trump administration (including a defense supply chain lawsuit) made it a unique target for this regulatory crackdown.
The episode exposes a chaotic, personality-driven U.S. AI regulatory regime that lacks rigor and predictability, potentially deterring investment and innovation.
"Beijing is laughing at us right now. — Source close to negotiations"
"It's just a handful of guys and how they feel, and that seems to be our regulatory apparatus. — Nilay Patel"
"If you're going to regulate, do it. Voluntary frameworks are just watered down. — Hayden Field"