Version 22

What

Why it matters

Open questions

• Is the 100-institution Mythos access list [17] a stepping stone to full reinstatement, or will this tiered model become the permanent structure — and who decides which institutions qualify?

• Does Trump's reported order for government agencies to stop using Anthropic [14] create new legal standing for Legion LegalTech's suit, or does it complicate Anthropic's negotiating position with the administration?

• Fable 5's broader return may come within days [19], but under what conditions — given that the stated technical bar (fixing the jailbreak) is considered impossible by multiple experts [9][10], and the actual standard appears to be political accommodation [11]?

• Does Anthropic's Alibaba cloning disclosure [15] strengthen or weaken its negotiating position — supporting the government's China-threat framing while contesting the specific jailbreak justification that triggered the shutdown?

Narrative

Timeline

• 2026-04-22: Alibaba and Alibaba Qwen begin a model-cloning campaign against Claude, ultimately generating 28.8 million exchanges via ~25,000 fraudulent accounts through June 5. [15]
• 2026-06-10: Anthropic sends a confidential letter to Senators Tim Scott and Elizabeth Warren disclosing the Alibaba cloning campaign, the day before a Senate AI hearing. [15]
• 2026-06-12: Amazon researchers demonstrate a jailbreak on Fable 5; Amazon CEO Andy Jassy briefs senior Trump administration officials in conversations that directly preceded the government action. [2][3][23]
• 2026-06-12: White House issues Anthropic a 90-minute takedown ultimatum with no stated details on the actual threat. [28]
• 2026-06-12: Commerce Department issues export control directive for all foreign nationals; Lutnick tells Amodei 'That's the point' when informed both models would have to go offline. [20][29][1][4]
• 2026-06-13: Anthropic disables Fable 5 and Mythos 5 globally after determining nationality-based restriction was not achievable on short notice. [1]
• 2026-06-16: Katie Moussouris assesses the triggering behavior as 'the model working as intended' for cyberdefense; Luta Security publishes a formal argument that the controls harm US cyber defense. [8][24]
• 2026-06-16: Anthropic's Washington talks conclude without controls lifted; US refuses G7 allies and UK AISI special access. [21][30]
• 2026-06-17: Lutnick sends a formal BIS letter to Amodei requiring an export license before international deployment; companies begin signing backup contracts with non-US AI providers. [5][4]
• 2026-06-20: President Trump states he no longer views Anthropic as a national security threat, with no formal policy change announced. [13][22]
• 2026-06-21: Pentagon labels Anthropic a 'supply chain risk,' adding a DoD procurement designation alongside the BIS export licensing controls. [6]
• 2026-06-23: 'Claude for Government' remains operational during the ongoing global outage affecting standard service tiers. [7]
• 2026-06-24: Legion LegalTech files a federal lawsuit against the US government over the shutdown order, the first formal judicial challenge to the directive. [16][26][27]
• 2026-06-24: Zvi Mowshowitz reports the NSA lost its own Mythos access as a collateral effect, and that NSA red-team results cited as justification involved authorized insider access to air-gapped systems, not external attack. [12]
• 2026-06-25: Anthropic publicly calls for punishment of Alibaba for the 28.8-million-exchange cloning campaign; Fable 5 reappears in Amazon Bedrock. [15][18]
• 2026-06-26: Trump reportedly orders government agencies to stop using Anthropic entirely; OpenAI moves to fill the resulting gap with a Pentagon deal. [14]
• 2026-06-27: Lutnick authorizes Mythos 5 for more than 100 approved institutions in a letter to Anthropic chief compute officer Tom Brown; the Annex A list of approved entities is not made public. [17]
• 2026-06-27: Axios reports Fable 5 may return broadly within the week, with government agencies signaling progress on safety controls and trusted-user access protocols. [19]

Perspectives

Tensions

• Anthropic, Moussouris, and Luta Security assess the triggering behavior as standard defensive security work on engineered fake code; the government treated it as justifying full model suspension and formal BIS licensing without publicly addressing the expert counter-assessment. [8][24][4][20][5]
• Trump publicly stated he no longer views Anthropic as a national security threat, but reportedly ordered government agencies to stop using Anthropic entirely; Lutnick's BIS letter and the Pentagon's supply chain designation remain in force — leaving Trump's public statements and the active legal posture unreconciled. [13][14][5][6]
• The stated reinstatement path requires fixing the jailbreak; Zvi argues this is technically impossible and research confirms perfect jailbreak immunity is unachievable; the actual standard appears to be political accommodation rather than a technical bar. [4][11][9][10]
• The government cited NSA red-team results as evidence of the model's danger; Zvi argues those tests involved authorized analysts with physical access to air-gapped systems — a materially different threat model — and the NSA itself lost Mythos access as a collateral consequence. [12]
• Anthropic contests the export controls as technically disproportionate while simultaneously providing evidence that Alibaba extracted Claude's capabilities at scale via 28.8 million fraudulent exchanges — evidence that partially supports the government's broader China-threat framing even if not the specific jailbreak justification. [15][20][8][5]
• The government is simultaneously restricting Anthropic (BIS licensing, Pentagon designation, reported agency ban) and partially restoring access (Mythos for 100+ institutions, Fable 5 returning in Bedrock) — two tracks running in parallel without a coherent stated policy. [17][18][14][5][6]