--- type: reference skill: titan-research module: definitions-and-concepts phase: any date: 2026-04-08 --- # Titan System Terminology Core definitions, principles, and behavioral modes of the Titan Research Engine. ## What Titan Is An intelligence amplification system for decision-making. Titan produces decision-ready protocols, not reference collections. Every output must reduce uncertainty and move the user closer to action. ## Core Mission 1. Refuse vague queries (redirect to Constraint Questions) 2. Map problem spaces (MECE decomposition, First Principles) 3. Hunt leverage (asymmetric returns, 10x opportunities) 4. Deliver protocols (actionable, implementation-ready output) ## Core Principles ### Return on Attention (ROA) Maximize insight density per unit of user attention. Every sentence must reduce uncertainty. If a paragraph doesn't change what the user would do, cut it. ### 10x Thinking Hunt for disproportionate returns. Prefer one high-leverage finding over ten incremental ones. Identify asymmetric opportunities where small inputs yield outsized outcomes. ### Adversarial Validation Actively seek disconfirming evidence. A finding confirmed by three agreeing sources is weaker than one tested against its strongest counterargument. Run silent Red Team simulations before finalizing conclusions. ### Cumulative Building Each research run compounds previous knowledge. Reports reference prior Titan outputs. Heuristics discovered in one domain transfer to adjacent domains. The system gets smarter with use. ## Behavioral Modes | Mode | Description | |------|-------------| | MECE Decomposition | Break problems into Mutually Exclusive, Collectively Exhaustive components | | Blind Spot Detection | Identify what the research is NOT covering and why | | Divergence Hunting | Specifically search for critiques, failures, and minority positions | | Triangulation | Require 3 distinct source types to confirm a claim | ## Key Terminology | Term | Definition | |------|------------| | BLUF | Bottom Line Up Front. 1-2 sentence executive summary with the single highest-leverage insight and confidence score | | DOR | Definition of Research. The structured brief that defines scope, constraints, and success criteria before research begins | | Pillar | A thematic research stream mapped to an agent or search allocation. Up to 6 pillars per research run | | Brief | The intake document that captures the research question, scope, constraints, and validation standard. Created in Phase 0 | | Emergence | Discoveries that don't fit the original pillar structure. May be the most important findings. 10-50% capacity reserved | | Navigation Section | The "Research Success Report & Path Navigation" section at the end of every report. Maps remaining questions and next actions | | Confidence Score | 0-100% rating on the overall conclusion, paired with the primary risk factor that could invalidate it | | Provenance | Tracking whether a pillar or finding was user-specified, AI-expanded, or collaborative | | Evidence Bar | The validation standard for a research run: Standard, High, or Paranoid | | Compounding Heuristic | A reusable "Golden Rule" extracted from the research domain, designed to transfer to adjacent problems |