Principles
2,201 principles · page 5 of 46
Track agent outcome accuracy separately from execution…
Track agent outcome accuracy separately from execution speed, classifying errors as systematic bias or random noise to determine appropriate correction mechanisms.
When agents reach optimization ceilings within their current
When agents reach optimization ceilings within their current framework, replace the framework entirely rather than continuing incremental refinement, as further optimization produces negligible returns while framework replacement enables order-of-magnitude improvements.
Allocate cognitive surplus to exploring alternative…
Allocate cognitive surplus to exploring alternative frameworks when capacity is high, and exploit proven frameworks when capacity is constrained, rather than maintaining fixed exploration rates regardless of context.
Deliberately accept temporary performance decreases…
Deliberately accept temporary performance decreases by exploring different frameworks when stuck at local optima, as hill-climbing within current constraints cannot reach higher peaks elsewhere in the solution space.
Execute continuous incremental improvement (kaizen) within…
Execute continuous incremental improvement (kaizen) within frameworks while maintaining meta-awareness to recognize when frameworks themselves need replacement (kaikaku), as the two modes serve complementary rather than competing functions.
Name the implicit framework governing each agent to make…
Name the implicit framework governing each agent to make architectural assumptions visible for evaluation, as unexamined frameworks feel like reality rather than choices.
Design your attention's choice architecture by changing…
Design your attention's choice architecture by changing defaults, removing triggers, and restructuring physical and digital spaces rather than relying on willpower to resist distractions in unchanged environments.
Dedicate focused, time-boxed sessions to improving one…
Dedicate focused, time-boxed sessions to improving one specific agent rather than attempting continuous unfocused improvement across all systems simultaneously.
Design optimization sessions to create moderate activation…
Design optimization sessions to create moderate activation pressure through deadlines—too little produces apathy, too much produces anxiety, optimal performance occurs between these extremes.
Hold measurement protocol constant across before-and-after…
Hold measurement protocol constant across before-and-after conditions while varying only the system being optimized—changing both simultaneously eliminates your ability to attribute effects to causes.
Benchmark efficiency, accuracy, and quality dimensions…
Benchmark efficiency, accuracy, and quality dimensions together rather than only measuring what is easy—optimization that improves measurable proxies while degrading unmeasured outcomes is regression disguised as progress.
Record your rationale for each optimization before observing
Record your rationale for each optimization before observing its outcome—this pre-commitment defeats hindsight bias and preserves your actual reasoning for later calibration.
Log failures with equal rigor to successes—changes…
Log failures with equal rigor to successes—changes that produced no effect or made things worse constrain your search space more than successes that might be luck.
Resist optimizing systems that have not yet generated enough
Resist optimizing systems that have not yet generated enough performance data to reveal their actual bottlenecks—premature optimization adds complexity to components that may not matter.
Adjust systems when environmental context shifts even…
Adjust systems when environmental context shifts even if nothing has broken—what was optimal under yesterday's conditions becomes suboptimal as the world changes, independent of system defects.
Distribute optimization responsibility throughout the system
Distribute optimization responsibility throughout the system rather than treating it as a separate specialized role—every operator should simultaneously use and refine their processes.
Design agents with explicit retirement criteria at creation…
Design agents with explicit retirement criteria at creation time — specify in advance what conditions would indicate the agent should be revised or replaced.
Pre-commit each day's attention allocation by deciding…
Pre-commit each day's attention allocation by deciding your most important task before opening any device or application, overriding the default mode network's tendency toward reactive stimulus response.
Scale new behaviors down to their minimum viable execution…
Scale new behaviors down to their minimum viable execution during initial deployment, then expand scope only after the activation circuit is reliable.
Design agents for the version of yourself who will encounter
Design agents for the version of yourself who will encounter the trigger under actual conditions (tired, distracted, stressed) rather than the idealized version making the plan.
Encode new behaviors in environmental structure through…
Encode new behaviors in environmental structure through specific physical cues, timing, and symbolic markers rather than relying on intention alone.
Expect and plan for an implementation dip where performance…
Expect and plan for an implementation dip where performance temporarily degrades when deploying a new agent, and continue execution through this phase rather than abandoning the agent.
Treat missing a single execution of a new habit as data…
Treat missing a single execution of a new habit as data rather than failure, and resume the next day without restarting the deployment timeline.
Schedule maintenance reviews at frequencies matched…
Schedule maintenance reviews at frequencies matched to the rate of environmental change in each agent's domain rather than using a uniform review cadence.
Inspect agents during their stable operational phase…
Inspect agents during their stable operational phase specifically because they appear to be working, not after they fail — early detection of drift is cheaper than repair after collapse.
Separate performance maintenance (is the agent doing what…
Separate performance maintenance (is the agent doing what it's designed to do) from alignment maintenance (should the agent still exist) and run them at different frequencies.
During maintenance reviews, strip accumulated cruft…
During maintenance reviews, strip accumulated cruft by removing steps, exceptions, and scope expansions that have made the agent more complex than its essential function requires.
When modifying an agent, evolution is preferable…
When modifying an agent, evolution is preferable when the core architecture remains sound and the gap between current and needed state is small; replacement is preferable when core assumptions have shifted or the architecture cannot represent new requirements.
Batch similar shallow tasks together to minimize…
Batch similar shallow tasks together to minimize the cognitive overhead of task-type switching and reduce attention residue accumulation.
Test for sunk cost contamination in agent retirement…
Test for sunk cost contamination in agent retirement decisions by asking whether you would design anything resembling the current agent if starting fresh today with the same need.
Expect an extinction burst when retiring an agent…
Expect an extinction burst when retiring an agent—a temporary increase in the urge to perform the retired behavior that occurs before decline—and do not interpret this as evidence the retirement was wrong.
When building a successor agent, inherit the predecessor's…
When building a successor agent, inherit the predecessor's accumulated knowledge by documenting edge cases, implicit dependencies, and temporal adjustments rather than starting from a blank slate.
When designing cognitive agents, examine the full pattern…
When designing cognitive agents, examine the full pattern across multiple past attempts rather than treating each as an isolated failure, as recurring design assumptions reveal systematic blind spots that single-instance analysis cannot detect.
Design cognitive systems for your actual operating…
Design cognitive systems for your actual operating environment including its variability and constraints, not for idealized conditions you wish you had, as agents designed for optimal conditions fail when reality reasserts itself.
Build cognitive agents with multiple independent trigger…
Build cognitive agents with multiple independent trigger mechanisms and resources rather than single points of dependency, as redundancy enables the system to survive environmental disruption that would otherwise cause cascading failure.
Evaluate cognitive agents by how they contribute…
Evaluate cognitive agents by how they contribute to the portfolio's overall properties (diversification, correlation, coverage) rather than by individual performance alone, as portfolio-level risk emerges from relationships between components.
Maintain most cognitive agents as stable, low-maintenance…
Maintain most cognitive agents as stable, low-maintenance foundations while reserving a small portion for high-variance experimental practices, as this barbell structure provides both reliability and the capacity for transformative breakthroughs.
Diversify cognitive agents across life domains rather…
Diversify cognitive agents across life domains rather than concentrating them in one area, as domain diversity ensures system-level functioning continues when any single domain experiences disruption.
Identify and eliminate shared dependencies across multiple…
Identify and eliminate shared dependencies across multiple agents during portfolio review, as correlated failure points create systemic fragility that individual agent optimization cannot address.
Make context switching costs visible through deliberate…
Make context switching costs visible through deliberate tracking before attempting to reduce them, because the tax is invisible in subjective experience and most people dramatically underestimate how frequently they switch.
Assess agents using frequency-times-behavioral-delta (how…
Assess agents using frequency-times-behavioral-delta (how often they fire × how much they change your actions) rather than perceived importance, as most perceived-important agents prove to be high-frequency noise generators upon measurement.
Retire cognitive agents based on current value production…
Retire cognitive agents based on current value production rather than sunk investment, as maintaining low-impact agents due to past effort compounds resource drain through the endowment effect and status quo bias.
Regular audits of recurring cognitive processes identify…
Regular audits of recurring cognitive processes identify agents that persist through inertia rather than current utility, enabling resource reallocation before accumulation costs compound.
Articulating the specific outcome an agent produces…
Articulating the specific outcome an agent produces and when you last used that outcome distinguishes active value-generators from legacy processes running on autopilot.
Estimate agent costs in three currencies—time, attention…
Estimate agent costs in three currencies—time, attention, and opportunity—because the attention cost typically exceeds the time cost and determines actual system load.
Documentation must be updated in the same commit or editing…
Documentation must be updated in the same commit or editing session as behavioral changes, not as a separate follow-up task, because manual synchronization steps that can be skipped will be skipped.
Documentation should disclose constraints and failure modes…
Documentation should disclose constraints and failure modes, not just capabilities, because users discover limitations through failure when documentation omits them.
Coordination costs between agents grow quadratically…
Coordination costs between agents grow quadratically (n(n-1)/2) while value grows linearly, establishing a natural upper bound on portfolio size before overhead exceeds benefit.