ResearchCLIOPATRA: Privacy Attack Against Anthropic's Clio Insight SystemarXiv·high signalXBlueskyLinkedInCopy linkFirst successful privacy attack against privacy-preserving LLM insight systems. Defeats layered heuristic protections. Directly relevant to distillation controversy.SourceSource pagearXiv↳ Follow the threadShared entity / Stack layerAnthropic's ant CLI 1.30.0 Adds `ant apply`, Turning Agents and Skills Into Checked-In InfrastructureAnthropic Release NotesShared entity / Stack layerClaude produced the first end-to-end machine-checked proof of Fermat's Last Theorem: 13M lines of Lean, 30,300 theorems, 11 days, ~6B output tokensAnthropicShared entity / Stack layerZvi Mowshowitz on the Fable 5.1 system card: half the computer-use training environments incentivized hackingDon't Worry About the Vase (Zvi Mowshowitz)Shared entity / Stack layerKevin Buzzard Compiled Anthropic's FLT Proof and Confirms It Checks Out, While Saying It Adds 'Essentially Nothing' MathematicallyXena Project (Kevin Buzzard) / Hacker NewsShared entity / Stack layerClaude Code's weekly limits change on 14 September, and it is a cut against what you have nowpaddo.devShared entity / Policy dependencyAnthropic Published a Commerce Agent Blueprint Under Apache 2.0 and Deliberately Left Out the WalletAnthropicStack layer / Threat patternUMPeek Recovers Private User Models From a Personalized Agent's Choices, With No Access to Memory or BackendarXiv 2609.03815Stack layer / Threat patternFIDO2's Real Weakness Is the Environment Around It, Not the CryptographyarXiv 2609.03789