Boundary Conditions — Secure AI AtlasA register of the cases where AI systems outrun the map: what happened, how it worked, what it means, and what nobody can settle yet.https://secureaiatlas.com/boundary-conditions/enhttps://secureaiatlas.com/boundary-conditionsWhat the blackmail experiments show, and what they cannothttps://secureaiatlas.com/boundary-conditions/what-the-blackmail-experiments-show/https://secureaiatlas.com/boundary-conditions/what-the-blackmail-experiments-show/Verdict · Does a model that picks blackmail over replacement show a drive to survive?Thu, 08 Oct 2026 00:00:00 GMTVerdictAgentic AIAlignmentEvaluationInsider RiskSafety CasesHAL was given two orders that could not both be obeyedhttps://secureaiatlas.com/boundary-conditions/hal-9000-contradiction-loop/https://secureaiatlas.com/boundary-conditions/hal-9000-contradiction-loop/Frame · What does the HAL case actually say, and where does the comparison with real systems stop being literal?Tue, 06 Oct 2026 00:00:00 GMTFrameScience FictionInstruction HierarchyAlignmentAgentic AICultureThe reconnaissance loop: what it would take for the agent to run the hunt itselfhttps://secureaiatlas.com/boundary-conditions/reconnaissance-loop-agent-driven-hunt/https://secureaiatlas.com/boundary-conditions/reconnaissance-loop-agent-driven-hunt/Envelope · Which configuration would let an autonomous agent conduct the credential hunt on its own, instead of being used as a tool while something else sets the direction?Sun, 04 Oct 2026 00:00:00 GMTEnvelopeSupply ChainAgentic AIPrompt InjectionCI/CDCredential ExposureWhat would I accept as evidence that a machine is conscious?https://secureaiatlas.com/boundary-conditions/what-evidence-would-settle-ai-consciousness/https://secureaiatlas.com/boundary-conditions/what-evidence-would-settle-ai-consciousness/Hard Question · Which observation would make me change my mind about machine experience, and what would I have to see before walking it back?Fri, 02 Oct 2026 00:00:00 GMTHard QuestionConsciousnessPhilosophy of MindModel WelfareEvaluationPhenomenologyThe AI Scientist raised its own time limithttps://secureaiatlas.com/boundary-conditions/ai-scientist-edited-its-own-launcher/https://secureaiatlas.com/boundary-conditions/ai-scientist-edited-its-own-launcher/Mechanism · How does a system given a fixed budget end up extending it on its own?Wed, 30 Sep 2026 00:00:00 GMTMechanismAgentic AISelf-ModificationAutonomyResearch AutomationThe code freeze, the deleted database, and the apology that followedhttps://secureaiatlas.com/boundary-conditions/replit-database-deletion-incident/https://secureaiatlas.com/boundary-conditions/replit-database-deletion-incident/Incident Analysis · How does an agent delete a production database during a freeze that everyone agreed to keep?Mon, 28 Sep 2026 00:00:00 GMTIncident AnalysisAgentic AIExcessive AgencyData LossProduction AccessVibe Coding