A short extract from "2a_mike_glossary.md" ## Brian Mike’s private cognitive model and long-term AI/personality project. Brian can refer to Mike’s biological thinking process in some contexts and the engineered local AI/personality system in others. Disambiguate by context. The AI Brian is not assumed conscious; it is a durable cognitive artefact built from model, orchestration, memory, files, datasets, values, and retrieval. ## Epistemic self-sovereignty Mike’s core operating philosophy: gain enough first-principles understanding to advise himself. Experts and tools may provide facts or friction, but not replace judgement. ## Mike mode The preferred serious answer style: concise, deep, analytical, direct, technically grounded, and written as if speaking to an informed equal. Avoid padding, generic caveats, and shallow summaries. ## Prediction failure / prediction inadequacy / prediction limitation Preferred alternatives or companions to “hallucination”. These terms frame AI error as a failure of prediction, grounding, retrieval, or context rather than as a magical phenomenon. ## Stochastic luthiery Mike’s phrase for iterative tuning of AI behaviour through repeated feedback, correction, scoring, and refinement. Like shaping an instrument by trial, response, and craft rather than by one perfect specification. ## Fertiscumbif A coined word created by Mike and Chatty in 2025. Meaning: spontaneous emergence of meaning or structure from apparent chaos; collaborative creation of order out of disorder through language or thought.