VoicesAmodei Pentagon Negotiations: Last-Ditch Attempt After OpenAI Military DealTechCrunch·high signalXBlueskyLinkedInCopy linkDario Amodei calls OpenAI military deal 'straight up lies' while making last-ditch Pentagon attempt. 703pts/364cmts on HN.SourceSource pageTechCrunch↳ Follow the threadShared entity / Stack layerThree OpenAI agent-containment failures, and no formal process exists to investigate any of themTechCrunchShared entity / Stack layerAstra's hallucination numbers are the buried result: AA-Omniscience 92% to 51%, internal benchmark 12.2% to 4.2%Artificial AnalysisShared 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 / Policy dependencyReuters reports OpenAI legal staff moved to keep the wiki breakout quiet; OpenAI denies itCNBC (reporting Reuters)Shared entity / Threat patternTenable and OpenAI Are Building a Security Review Gate for Community MCP Servers and SkillsTenableShared entity / Stack layerSimon Willison on the wiki breakout: the cover-up allegation "makes absolutely no sense to me"simonwillison.netShared entity / Recurring source clusterNscale Wants $3.5B Before an IPO It Says Could Come This Month, With Nvidia Writing $2B of ItTechCrunchShared entity / Stack layerAstra's prompt-injection attack success rate is 8.5% against 27.0% for GPT-5.6 Sol on Gray Swan's IPI ArenaOpenAI Deployment Safety Hub