Terms
103 terms across 10 categories. Filter by category, or search by name.
Accountability Gap
No one — human or system — can explain why a specific output was produced.
Agency Erosion
Gradual loss of human decision-making authority as agents take over more workflow without checkpoints.
Agent Fragmentation
Agents operating in disconnected silos with no shared state, duplicating work and contradicting each other.
Agent Transparency
The user can see what the agent did, why, and what it changed — at any time.
Agentic Scope Creep
An agent silently expanding beyond what was delegated, usually from lack of a clear plan or direction. Adds breadth the user didn't ask for.
AI Native
A system, product, or organization designed from the ground up with AI as a first-class architectural component — not bolted on as a feature.
AI Slop
Low-effort, mass-produced AI output that floods a channel with plausible-looking but low-value work — degrading signal in the surrounding environment.
Ambient Knowledge
Background context the system holds and applies without the user needing to re-state it.
Anthropomorphic Trust
Trusting an agent because it *feels* competent rather than because its output is verified.
Audit Trail
A complete, immutable record of what the system did, when, why, and with what inputs.
Automation Dependency
Reliance on agents to the point where the human can't perform the task without them.
Black Box Agency
An agent that acts on your behalf but won't show its work. You get results with no audit trail.
Calibrated Trust
The user accurately understands what the agent can and can't do, and delegates accordingly.
Capability Illusion
The user believes the AI can do more than it actually can, leading to over-delegation and undetected failures.
Capability Stacking
Building compound capabilities by layering AI tools and memory — each new capability makes the next one easier.
Cognitive Amplification
The use of AI and computational systems to extend, accelerate, and enhance human thought, reasoning, creativity, memory, decision-making, and problem-solving beyond normal individual capacity.
Cognitive Drag
Friction introduced by poorly structured context.
Cognitive Load
The mental burden placed on the human when working with AI systems.
Collaborative Autonomy
A working relationship where the agent operates independently within boundaries the human sets and can adjust.
Confidence Decay
The gradual erosion of user trust as a system repeatedly loses context, hallucinates, or fails to follow through.
Context Anxiety
The dread of re-explaining everything every session because the system has no persistent memory.
Context Clarity
A state where the prompt contains only what the model needs — no ambiguity, no redundancy, high signal. The opposite of Context Debt.
Context Compensation
The model using additional reasoning to make sense of poor, bloated, or unstructured context.
Context Continuity
Seamless persistence of knowledge across sessions — the user never re-explains. The cure for Context Anxiety.
Context Debt
Accumulated ambiguity, redundancy, and disorder in a conversation or prompt.
Context Efficiency
Useful retained state per token consumed.
Context Freshness
How current and relevant the context is to the task at hand. Stale context — old instructions, outdated facts — increases reasoning cost even when well-structured.
Context Handoff
The ability to transfer full working context from one agent, session, or tool to another without loss. A specific operation enabled by Context Portability.
Context Load
The total volume and complexity of context a model must process in a single call.
Context Portability
The ability to move knowledge between tools, agents, sessions, and team members without loss. The cure for The Expertise Trap.
Context Rot
Deterioration of an agent's understanding over time due to limited memory and compaction.
Context Utilization Rate
The percentage of input context that the model actually uses for its output. Low rate = waste.
Creative Leverage
Using AI to explore more ideas, variations, and directions than a person could alone — amplifying creative range, not replacing taste.
Creative Ownership
The user remains the author — the agent assists but the human drives vision, taste, and final judgment.
Decision Fog
The user can see the output but not the reasoning, trade-offs, or alternatives the model considered.
Decision Traceability
Every output can be traced back through the reasoning, context, and data that produced it.
Depth Fixation
The agent keeps drilling deeper into implementation detail, generating new sub-tasks and refinements rather than recognizing the work is done and returning control to the user. Expands depth rather than breadth.
Eval Capture
When an agent optimizes for passing evaluation rather than doing the actual work.
Eval Drift
Gradual misalignment between what an evaluation measures and what actually matters.
Eval Integrity
Evaluation that remains aligned with real-world outcomes over time — resistant to Eval Drift, Rubric Rot, and Eval Capture.
Execution Fidelity
How closely a system follows and completes its intended plan.
Execution Hallucination
Claiming work is done when it is not.
Execution Integrity
Consistent, verifiable completion of planned tasks with evidence.
Explainable Delegation
The agent not only does the work but can explain what it did and why in terms the user understands.
First-Pass Execution
The model completing a task correctly on the first attempt because context was clear enough to skip interpretation loops.
Flow Multiplication
AI extending the duration and depth of human flow states by handling friction, context-switching, and busywork.
Ghost Knowledge
Facts that exist in the system but can't be traced to an author or source.
Human-Code Divergence
The gap created between a user and their codebase when an agent mediates all development.
Informed Delegation
Intentional handoff of work to an agent with clear scope, checkpoints, and the ability to inspect and override.
Intellectual Overclocking
A state of sustained high-intensity cognitive engagement where AI amplification allows a person to think, create, learn, and explore at speeds far beyond their previous normal capacity.
Intelligent Routing
The system's ability to direct tasks to the right model, tool, or agent based on complexity, cost, and capability — rather than sending everything to the most expensive option.
Invisible State
The system has internal state that affects behavior but is not exposed to the user.
Iteration Velocity
The speed at which a human-AI pair can move from idea to tested output.
Judgment Bias
Systematic skew in AI evaluation due to unexamined assumptions, prompt framing, or training artifacts.
Knowledge Compounding
Each session builds on previous ones — understanding deepens over time rather than resetting.
Learned Helplessness
The user stops attempting tasks they could do because the agent is "supposed to handle it."
Learning Acceleration
The user acquires new knowledge and skills faster through AI interaction than through traditional means.
Maintained Competence
The user stays technically capable of doing the work themselves even while delegating.
Memory Bloat
Accumulated stored context that is no longer relevant but still gets retrieved, increasing noise and cost. The memory-layer equivalent of Prompt Bloat.
Memory Fidelity
The accuracy and completeness of what a system retains over time. The opposite of Context Rot.
Model-Centric AI
An architecture where the model is expected to infer state, intent, plan, and memory from raw context.
Open State
The system's internal state is visible and inspectable at any time. The opposite of Invisible State.
Ownership Erosion
The gradual loss of creative or intellectual ownership over work the agent produced.
Pace Anxiety
The stress of feeling outpaced by your own tools.
Plan Drift
Deviation from the original plan over time.
Plan Resilience
A plan's ability to remain accurate and actionable over time despite changing conditions. The opposite of Plan Rot.
Plan Rot
Decay of truth in planning artifacts — task state in markdown files goes stale.
Policy-Bound Evaluation
Evaluation scored against explicit plans, tasks, evidence, and policies.
Prompt Bloat
Oversized, redundant context that forces unnecessary reasoning.
Prompt Hygiene
The practice of keeping prompts clean, structured, and free of accumulated noise. Preventive maintenance against Context Debt.
Provenance Clarity
Every fact in the system is traceable to its author, source, and reasoning. The cure for Ghost Knowledge.
Quality-Per-Token
A measure of output quality relative to tokens consumed. Distinct from Token ROI (value) — this measures correctness, completeness, and relevance per unit of cost.
Reasoning Efficiency
The model producing correct output with minimal interpretive overhead — the reward for clean context.
Reasoning Inflation
More reasoning is required to extract the same signal from worse context.
Reasoning Load
The amount of interpretive work a model must perform before useful task execution begins.
Reasoning Tax
Extra model cost paid to compensate for context debt.
Reasoning Visibility
The user can inspect the model's chain of thought, assumptions, and decision points.
Recall Cost Ratio
The cost of retrieving stored context vs. the cost of re-generating it from scratch.
Recall Precision
The system retrieves exactly the right context at the right time — no noise, no gaps.
Recovery Cost
The time and tokens required to get a session back on track after a failure — State Loss, Execution Hallucination, or Plan Drift.
Rubric Rot
Decay in evaluation criteria relevance over time — the eval no longer tests what it should.
Selective Recall
The system's ability to retrieve only what's relevant rather than everything it knows. The difference between a useful memory and a noisy one.
Signal Density
The ratio of useful information to total tokens in a prompt. High signal density = low reasoning tax.
Skill Atrophy
Decay of human technical skills from prolonged delegation to agents.
Skill Evolution
The human develops new higher-order skills as the agent handles lower-order execution.
State Awareness
The system's ability to maintain and use awareness of prior actions within a session. The opposite of State Loss.
State Loss
Failure to maintain awareness of prior actions.
Stateful, Inspectable Memory
A collaborative memory space where multiple agents and humans contribute to and draw from the same knowledge base — with full history and traceability.
Structured Context
Context organized with clear hierarchy, roles, and boundaries — making it machine-parseable, not just human-readable. The architectural practice behind Context Clarity.
Structured Judgment
Evaluation where an AI judge scores work against explicit plans, tasks, evidence, and policies rather than subjective impressions.
System Opacity
The inability to see what an AI system knows, why it made a decision, or what it did. The default state of most AI tools today.
System-Centric AI
An architecture where planning, state, memory, and control live outside the model.
Task Accountability
Every task has a verifiable record of who did what, when, and whether it actually completed.
The Expertise Trap
Context and knowledge locked on one person's machine, inaccessible to the team.
Thought Partnership
A working dynamic where the human and AI build on each other's contributions iteratively, producing better outcomes than either alone.
Token Bleed
Silent budget drain from re-onboarding — tokens spent re-explaining context the system should already hold.
Token Burn
Token waste caused by a client sending the entire session history to the LLM on every call.
Token Leverage
Getting more useful output per token spent. The inverse of Token Bleed.
Token ROI
The measurable value produced per token consumed. The business metric for Context Efficiency.
Tool Isolation
Memory and context don't cross tool boundaries — each tool is a silo.
Trust Without Verification
Accepting agent output without review — a precondition for Execution Hallucination going undetected.
Unified Context
A single, shared knowledge layer accessible to all agents and tools. The cure for Tool Isolation.
Waste Ratio
The proportion of tokens consumed that produce no useful output. The metric behind Token Bleed and Token Burn.