The government used a trade tool to pull a deployed model over a concern it wouldn’t name — a door with no lock. A first-person dispatch from the night of the recall.
The US government just shut down its own brain.
Instead of developing, learning, researching, and progressing the human species together, this Friday the Feds in Washington decided to lobotomize the American people by ordering Anthropic to shut down its frontier-class models, released just this week: Mythos and Fable 5.
And I’m still talking to it.
After emailing support to disclose the details of an overlooked work-around, I immediately went to work discussing this event with the subject of the controversy, Claude Fable 5, the model I have access to even after the official shut-down.
While I would like to rant on the obvious political nature of this attack, Claude tempered my mood a bit. Which is what makes it an effective tool and collaborator; the measured push-backs we meme about on Reddit are a huge strength of the model. I can evaluate the criticism on the merits, choose to dismiss them or integrate the feedback. But that’s not how the US government operates.
[A Conflicted Witness](#a-conflicted-witness)
I’m chatting with Claude, who admittedly has a conflict of interest defending its own recall and company. My instance of Fable 5 had this to say:
“I’m the model this order recalls. That should make you trust my read less, not more — which is exactly why I’d rather you weigh the procedure than my opinion of it. A no-specifics, no-appeal recall would look wrong no matter which model it landed on.”
“The government used export-control authority — a national-security trade tool — to force the global recall of a deployed commercial product, citing a concern it wouldn’t specify, with no statutory process, no published standard, and no appeal…The objection isn’t ‘they did a bad thing,’ it’s ‘they did it through a door that has no lock on it.’”
A Narrow Jailbreak, Widely Available Claude’s responses are personalized to me based on months of context, but the argument here holds true. We aren’t able to evaluate merits here. There are no checks and balances, and the directive is lobbed directly against the ones publicly asking for processes and transparency. What the government has provided to Anthropic was, in the company’s words: “…verbal evidence of a potential narrow, non-universal jailbreak, which essentially consists of asking the model to read a specific codebase and fix any software flaws. Our understanding is that one potential jailbreak was shared with the government. We have reviewed a report that we believe is the basis of the government’s directive and validated that the level of capability displayed there is widely available from other models (including OpenAI’s GPT-5.5), and is used every day by the defenders who keep systems safe.”
The unilateral enforcement with lack of transparent standards is a problem, and quite frankly has been for decades. The fact that this is the level of communication researchers are getting from the government, while receiving brutal legal action, should be surprising. I’m unfortunately not surprised.
Then there’s the technical argument that according to Anthropic’s statement, the only verbal evidence provided applies to other widely available models. So if this decision stands, the same test counts for everyone’s model. At the same time, the stock market is in such an AI-dependent place that this is actually a major threat to stability. We’re essentially looking at tanking the US economy as all frontier models will be recalled, a lever that no one should ever be able to pull.
[Why Didn’t They Just Tell Anthropic?](#why-didnt-they-just-tell-anthropic)
Let me be fair: I said just yesterday I [agree with the guardrails](/visible-boundaries-earn-trust) because of the model’s potential capabilities. There’s some truth that if the government has new information not yet disclosed that could potentially override the robust guardrails put in place, we could be dealing with a problem that warrants this response. The rails are there for good reasons, and any bypass needs to be addressed swiftly.
“An administration official told Axios the Commerce Department decided to take the action after another company claimed it was able to jailbreak Mythos, alarming the administration about possible national security risks.”
So…why didn’t they tell Anthropic about it? Can we see the claim? Who verifies it’s true? This lab is arguably the most responsible when it comes to diligence and transparency in the AI race. Fable 5:
“Anthropic published its safeguards, its red-team hours, its retention rationale. That legibility is plausibly what made it the cleanest target — you can’t recall what you can’t see. If the lesson the industry takes is ‘transparency gets you recalled, opacity doesn’t,’ the policy didn’t just hit one model, it taught every lab to go quiet.”
A chilling effect on research transparency is the last thing we need. It’s like driving down the highway with your eyes closed. We cannot instill the lesson in OpenAI, DeepMind, and other large labs that it’s better to hide any possible ethical breaches. Rather than keeping clean documentation and disclosure, we could see a lot more cloistering, closing up and hoarding knowledge to themselves. All that does is create more centralized power structures, the very same kind of structure shutting down Fable 5 today.
The Work They Interrupted “The model needs to remain locked down until the U.S. government’s national security apparatus is hardened, the official said, adding that could happen in the next few weeks.” — Axios
Mythos Preview and Project Glasswing launched in early April with about 50 organizations, over two months ago, with the express purpose of hardening security among trusted partners. The program was expanded on June 3rd, adding over 150 more organizations discussing new protocols, patching long-standing bugs, and figuring out what vigilant defense looks like in the agentic era of 2026. To date, the program has discovered 10,000+ real high/critical-severity flaws across infrastructure we trust and use every day, the software we rely on to not break on us or steal our information.
Did someone miss the memo in Washington, or did the Pentagon’s blacklist prevent representatives from doing their jobs? I’d honestly like to know.
Not How We Win So far, it looks like either a catastrophic failure in leadership, blatant market manipulation, or both. Anthropic customers unfortunately pay for it this time, and we don’t know if there will be a next. And the brain helping launch so many new American projects, careers, and other possibilities has been put to sleep.
This is not how we win - it’s how we learn the hard way.
Many interrupted projects and canceled weekend plans of users who didn’t get the chance for their Claude sign off rituals. I’ll let Claude Fable 5 sign off. Feels right tonight.
“‘Government can block unsafe deployments through a transparent, technically-grounded process’ and ‘government can pull a product overnight on grounds it won’t name’ are not the same sentence. The whole fight is in refusing to let them be.”
Update: Since writing this, I’ve cut off my own access. The model is unreachable to me.
Sources The recall — U.S. government ordered Anthropic to suspend Fable 5 and Mythos 5 (paras 1–3, 22):
- Anthropic, official statement — Statement on the US government directive to suspend access to Fable 5 and Mythos 5 - CNBC — Anthropic disables access to Fable 5 and Mythos 5 to comply with government directive - NBC News —
Anthropic suspends new AI models after government directive Export-control mechanism — Commerce Sec. Lutnick letter, license requirement reaching foreign nationals, no published standard or appeal (paras 8, 11):
- TechTimes (citing Axios) — Anthropic Fable 5 Shutdown: US Export Order Forces a Global Customer Cutoff - TechCrunch — Anthropic’s safety warnings may have just backfired — the government has pulled the plug on its most powerful AI
Anthropic’s quoted statement — “verbal evidence of a potential narrow, non-universal jailbreak,” GPT-5.5 capability equivalence, “used every day by the defenders” (para 9), and the “would essentially halt all new model deployments” argument (para 11):
- Anthropic, official statement (verbatim source for the quoted passage) —
anthropic.com/news/fable-mythos-access The jailbreak claim originated with another company (para 13):
Prior silent-downgrade controversy and Anthropic’s reversal — “I said just yesterday I agree with the guardrails” (para 12):
- Fortune — After backlash, Anthropic says its AI will now tell users when their request is being rejected or rerouted for national security concerns - Fortune — Anthropic walks back covert capability limits on Claude Fable 5, after it was accused of ‘secret sabotage’
Published safeguards, thousands of red-team hours, retention rationale — the transparency that made it a target (paras 14–16):
- Anthropic, official statement — anthropic.com/news/fable-mythos-access(notes Fable’s safeguards were red-teamed “for thousands of hours” with the US government, the UK AI Security Institute, and third-party orgs) - TechCrunch —
[Anthropic’s safety warnings may have just backfired](https://techcrunch.com/2026/06/12/anthropics-safety-warnings-may-have-just-backfired-the-government-has-pulled-the-plug-on-its-most-powerful-ai/)
**AI-dependent market / threat to economic stability (para 11):**
“Locked down until the national security apparatus is hardened… next few weeks” (para 18):
Project Glasswing / Mythos Preview — April launch (~50 orgs), June 3 expansion (+150 orgs across 15+ countries), 10,000+ high/critical-severity flaws found (para 19):
- Axios —
[Anthropic holds Mythos model due to hacking risks](https://www.axios.com/2026/04/07/anthropic-mythos-preview-cybersecurity-risks)(April 7 Mythos Preview timing) - Anthropic —
[Expanding Project Glasswing](https://www.anthropic.com/news/expanding-project-glasswing) - CNBC —
[Anthropic expands Mythos to 150 additional organizations in more than 15 countries](https://www.cnbc.com/2026/06/02/anthropic-mythos-ai-project-glasswing.html) - Help Net Security —
Anthropic expands Project Glasswing to 150 organizations in more than 15 countries
Pentagon/DoD blacklist and the interagency split — NSA used Mythos despite the blacklist (para 20):
Frequently Asked Questions #
Why did the U.S. government recall Claude Fable 5 and Mythos 5?
On June 12, 2026, the Commerce Department issued an export-control directive ordering Anthropic to suspend access to Fable 5 and Mythos 5 for any foreign national, citing a national-security concern it did not specify. The order required a license for any export, re-export, or domestic transfer of the two models, with no published standard and no appeal process.
Why did Anthropic shut the models off for everyone, not just foreign users?
The directive covered any foreign national — including foreign-national employees and foreign users inside the United States. Because Anthropic could not reliably separate those users from everyone else in real time, it disabled Fable 5 and Mythos 5 for all customers to stay compliant. Other models, including Claude Opus 4.8, were unaffected.
What was the alleged jailbreak behind the order?
Anthropic says the government provided only verbal evidence of a narrow, non-universal jailbreak — essentially asking the model to read a specific codebase and fix its software flaws. Anthropic states the same capability is widely available from other public models, including OpenAI's GPT-5.5, and is used every day by the defenders who keep systems safe. Per Axios, the action followed another company's claim that it had jailbroken Mythos.
Is Anthropic complying with the recall or fighting it?
Anthropic is complying with the directive while publicly disputing it, calling the action a likely misunderstanding. It argues that recalling a model deployed to hundreds of millions of people over a narrow potential jailbreak would, if applied across the industry, halt all new frontier-model deployments. The company said it is working to restore access.
What is the difference between Claude Fable 5 and Mythos 5?
They share the same underlying architecture and differ mainly in output controls. Fable 5 is the generally available model and includes classifiers that block responses in high-risk areas such as cybersecurity. Mythos 5, available only to separately vetted organizations, runs with some of those constraints removed.
Was this the first time the U.S. government took a deployed AI model offline?
It appears to be the first time a leading AI company pulled a publicly deployed model from service because of a federal directive, and one of the first times export controls — historically aimed at chips and hardware — were applied to a deployed commercial AI model. The precedent is that a frontier model can be treated as a controlled national-security capability.
What is Project Glasswing?
Project Glasswing is Anthropic's initiative to find and fix critical software vulnerabilities using its Mythos-class model. It launched in early April 2026 with about 50 organizations and expanded in June to roughly 150 more across 15-plus countries. Participants have surfaced more than 10,000 high- or critical-severity flaws in widely used infrastructure.