Principles
2,201 principles · page 3 of 46
Write compressed interpretations during conversations rather
Write compressed interpretations during conversations rather than verbatim transcripts, as the compression process itself creates stronger memory encoding than passive recording.
Write about emotional experiences with increasing use…
Write about emotional experiences with increasing use of causal and insight words (because, realize, understand) rather than staying purely in feeling-description to maximize psychological and physical health benefits.
Resist interpreting emotions during initial capture; record…
Resist interpreting emotions during initial capture; record raw affective state and defer causal explanation to review when multiple entries enable pattern recognition.
Capture every instance of surprise (prediction-expectation…
Capture every instance of surprise (prediction-expectation mismatch) immediately, regardless of magnitude, because small surprises reveal systematic model blind spots that large surprises obscure.
When delegating to AI systems, maintain human capability…
When delegating to AI systems, maintain human capability to evaluate output quality independent of the AI, as AI assistance creates illusions of understanding that operate without the performer's awareness.
Identify tasks where you are a single point of failure (bus…
Identify tasks where you are a single point of failure (bus factor of one) and systematically transfer ownership to create resilient systems that function without your presence.
Delegate tasks at 70% of your quality level when adequate…
Delegate tasks at 70% of your quality level when adequate work done reliably by others creates more system value than excellent work that depends on your availability.
Explicitly name the level of autonomy you are granting…
Explicitly name the level of autonomy you are granting in each delegation to prevent mismatches between intended and assumed authority.
Match delegation level to the combination of task…
Match delegation level to the combination of task reversibility, delegate's task-specific competence, and the trajectory of trust rather than using a fixed level per person.
Practice delegation sub-skills deliberately—specification…
Practice delegation sub-skills deliberately—specification clarity, calibration accuracy, feedback delivery, and discomfort tolerance—through structured repetition with clear goals rather than merely accumulating delegation experience.
Design distributed regulatory systems with requisite variety
Design distributed regulatory systems with requisite variety matching the complexity of what they regulate rather than attempting personal control of all decisions.
Control outcomes by designing the system within which…
Control outcomes by designing the system within which decisions are made—boundaries, feedback loops, standards, escalation criteria—rather than controlling individual decisions directly.
Invest the initial cost of clear specification, capable…
Invest the initial cost of clear specification, capable delegate selection, and verification loops to create delegation that produces multiplicative rather than additive returns.
Delegate the building of systems that handle recurring task…
Delegate the building of systems that handle recurring task classes rather than delegating individual task instances to create compounding capacity.
Convert captured surprises into precisely-formed open…
Convert captured surprises into precisely-formed open questions that activate as persistent attentional filters, directing future information-gathering toward resolving the revealed model gap.
Before delegating a task, determine whether the task should…
Before delegating a task, determine whether the task should exist at all to avoid optimizing execution of low-value work.
Route work to the appropriate agent—human, system, tool…
Route work to the appropriate agent—human, system, tool, or environment—based on who can handle it at equivalent or better quality rather than defaulting to personal execution.
Invest reclaimed time from delegation into higher-leverage…
Invest reclaimed time from delegation into higher-leverage work—system design, strategic thinking, relationship building—rather than filling it with more execution tasks to achieve compound returns.
Design systems that route incoming work to appropriate…
Design systems that route incoming work to appropriate agents without your direct involvement for each decision to achieve scalable delegation.
Instrument systems with the minimum number of metrics…
Instrument systems with the minimum number of metrics that change behavior when they move, because observation itself consumes the cognitive resources needed to act on what you observe.
Define success operationally as specific measurement…
Define success operationally as specific measurement procedures rather than abstract qualities, because unmeasurable definitions prevent detection of the gap between felt understanding and actual performance.
Pair quantitative metrics with qualitative judgment…
Pair quantitative metrics with qualitative judgment to prevent surrogation where the measure psychologically replaces the construct it was designed to represent.
Match monitoring frequency to the rate at which meaningful…
Match monitoring frequency to the rate at which meaningful change occurs in each system, sampling at least twice as fast as the fastest change you need to detect to avoid aliasing real patterns.
Use event-driven monitoring for acute failures that require…
Use event-driven monitoring for acute failures that require immediate response and scheduled reviews for trend analysis, because neither alone provides sufficient signal across different failure timescales.
Consolidate all agent status information onto a single…
Consolidate all agent status information onto a single surface visible without navigation, because distributed monitoring increases the cognitive cost of checking below the threshold where checking actually occurs.
Evaluate decisions by process quality rather than outcome…
Evaluate decisions by process quality rather than outcome, using documented reasoning and predictions to separate signal (decision skill) from noise (outcome luck).
Include temporal trend information alongside point-in-time…
Include temporal trend information alongside point-in-time status, because current state on a declining trend is more alarming than degraded state on a recovery trend.
Establish an error budget that defines acceptable…
Establish an error budget that defines acceptable unreliability as a resource to spend on experimentation rather than a failure to eliminate, because pursuing perfect reliability prevents the risk-taking required for growth.
Calculate mean time between failures (MTBF) and mean time…
Calculate mean time between failures (MTBF) and mean time to recovery (MTTR) separately, because agents with the same failure rate can have dramatically different impact depending on recovery speed.
Monitor multiple dimensions of effectiveness simultaneously…
Monitor multiple dimensions of effectiveness simultaneously to prevent metric gaming — single metrics become corrupted when they become targets.
Measure time-to-activation as a core performance metric…
Measure time-to-activation as a core performance metric — speed of response often determines whether a correct response is useful.
Use percentile distributions rather than averages to measure
Use percentile distributions rather than averages to measure agent performance under varying conditions — P50, P95, and P99 expose different failure modes.
Reduce the number of alternatives an agent must evaluate…
Reduce the number of alternatives an agent must evaluate to decrease response latency — discrimination time scales logarithmically with choice count.
Implement if-then trigger-response structures to achieve…
Implement if-then trigger-response structures to achieve strategic automaticity — explicit trigger specification creates faster activation than goal-alone approaches.
Add debouncing (requiring conditions to persist…
Add debouncing (requiring conditions to persist for a duration) to reduce false positive rates without tightening triggers — time filters noise from signal.
Schedule explicit negative audits asking 'what should…
Schedule explicit negative audits asking 'what should have triggered but didn't' rather than waiting for agents to fire — false negatives don't announce themselves.
Make capture tools visible and accessible at eye-level…
Make capture tools visible and accessible at eye-level or top-of-stack in every context, because visibility determines attention and proximity determines behavior independent of conscious intention.
Externalize agent specifications in written form…
Externalize agent specifications in written form as the reference for drift detection — comparison requires a fixed baseline outside memory.
Implement scheduled automated comparisons between agent…
Implement scheduled automated comparisons between agent specification and actual behavior — humans do not spontaneously notice gradual degradation.
Distinguish activity monitoring from purpose monitoring…
Distinguish activity monitoring from purpose monitoring — an agent can fire reliably while producing output that no longer serves its original goal.
Rotate measurement targets periodically to prevent…
Rotate measurement targets periodically to prevent optimization for the metric displacing optimization for the outcome.
Define 'decision enabled' as the success criterion for each…
Define 'decision enabled' as the success criterion for each monitoring activity—if no decision changed in the past 30 days, eliminate the monitoring.
Separate monitoring into quantifiable metrics (automate)…
Separate monitoring into quantifiable metrics (automate), pattern recognition (semi-automate), and meaning-making (manual reflection).
Establish dynamic baselines from actual performance data…
Establish dynamic baselines from actual performance data rather than setting thresholds from aspirational targets.
Schedule periodic meta-monitoring reviews to verify…
Schedule periodic meta-monitoring reviews to verify that automated monitoring systems remain calibrated to changing reality.
Review accumulated journal entries analytically rather…
Review accumulated journal entries analytically rather than chronologically to detect patterns invisible in single observations.
Include both successes and failures in monitoring logs…
Include both successes and failures in monitoring logs to prevent perception bias toward system dysfunction.
Increase ability and improve prompts rather than relying…
Increase ability and improve prompts rather than relying on motivation to sustain capture behavior, as motivation fluctuates unpredictably while environmental triggers remain stable.