Concepts
152 concept pages — the canonical spine.
- Abstract-to-Concrete Grounding
- Adaptation Strategy Ladder
- Adversarial Planning Council
- Advisor Mode
- Agent Audition
- Agent Communication Topology
- Agent-Computer Interface (ACI)
- Agent Gateway
- Agent Loop
- Agent-Mediated Software
- Agent-Native Cron
- Agent Rituals
- Agent-Shape Triage
- Agent Supervision
- Agent Task Graph
- Agentic Distribution
- Agentic Simplicity
- Agentic UI Testing
- Agentic Workflow Patterns
- AI Completion Asymptote
- AI Second Brain
- Annoyances vs Constraints
- Attention Budget
- The Augmented LLM
- Authenticity Collapse
- Authority-Independent Verification
- Bounded Fan-Out
- Bounded Negotiation with Fallback
- BYO Inference Endpoint
- Capability Overhang
- Capture-Organize-Distill-Express (CODE)
- Capture-Storage-Retrieval Pipeline
- Check Gaming
- Chinchilla Scaling Law
- CLI Tools over MCP Servers
- Cognitive Offload Cost
- Cognitive Scaffolding
- Compiled Knowledge Base
- Concise Prompting
- Context Compression
- Context-Independent Review
- Context Routing
- Context Smart Zone
- Context Substrate
- Contextual Retrieval
- Cross-Model Independence
- Curatorial-Voice Learning
- Data-Centric AI
- Decision Log
- Decoding & Sampling Controls
- Dynamic Retrieval
- Error Analysis
- Eval-Driven Development
- Evergreen vs Volatile Context
- Evidence-Gated Completion
- Execution Commoditization
- Extended Mind Criteria
- Extended Mind
- Falsification-First Questioning
- Fidelity-Raising Prototype
- Frontier Scouting
- Golden Templates
- Harness / Model Fit
- Harness / Model Separation
- HTML as Native Output
- Hybrid Retrieval
- Imagination Constraint
- In-House AI Diagnostician
- Incremental Indexing
- Incremental vs Discontinuous Tasks
- Inference Batching
- Inference Bottlenecks
- Inference Latency Metrics
- Input & Output Guardrails
- Intent Context
- Knowledge-Graph Lint
- Knowledge Graph Retrieval
- Knowledge Work as Code
- KV Cache Fragmentation
- KV Cache
- Perplexity & Cross-Entropy
- Layered Agent Memory
- Layered Agentic Architecture
- Leading Words
- Levels of Evaluation
- LLM-as-Judge
- LLM as Resource Router
- Loop Engineering
- Loop Training Mode
- Low-Blast-Radius First
- Managed Agent
- Map-First Planning
- Meta-Prompt
- Minimal Skill Surface
- Model Distillation
- Model Merging
- Model Parallelism
- Model Quantization
- Model-Tier Routing
- Motivated Exposition
- Multimodal Video Ingestion
- Negative Prompting
- Offline Consolidation Cycle
- Open-Weight Capability Gap
- Orchestrator Unknowns (Finding Your Unknowns)
- Organize by Actionability
- Paged Attention
- Parameter-Efficient Fine-Tuning (PEFT / LoRA)
- Parity Principle
- Post-Training Alignment
- Pre-Deployment Validation
- Prefill and Decode Phases
- Prompt Attack Surface
- Prompt Caching
- Public Skill Adoption
- Pure Agent Application
- Query-Shaped Storage
- Rank Fusion
- Reasoning Effort Control
- Recency-Grounded Research
- Repeated-Sampling Scaling
- Repo-Local Capability Binding
- Reranking
- Resolution-Typed Tasks
- Retrieval Maturity Levels
- Reusable Workflow Library
- Role-Typed Agent Roster
- Safety Routing Fallback
- Scene-Change Frame Sampling
- Search-Then-Get
- Self-Improving System
- Semantic Retrieval
- Shared Review Surface
- Skill Artifact Transfer
- Skill Authoring Checklist
- Skill-Driven Loop Development
- Skill Extraction
- Skill Invocation Trigger
- Skill Pruning
- Spec-Driven Development
- Speculative Decoding
- Stakeholder Clone
- Step Isolation
- Synthetic Data Generation
- Tacit Capability Awareness
- Team-Forming Constraints
- Text-Space Optimization
- Total Cost of Inference
- Validation-Gated Update
- Vibe Coding
- Wargaming (Adversarial Contingency Planning)
- Workflows vs Agents