Definitionv1
Benchmark: a measurable baseline measurement of system
Benchmark: a measurable baseline measurement of system performance taken before any optimization changes are applied, used to determine whether subsequent changes actually improved performance
Why This Is a Definition
This definition precisely establishes the semantic boundary of 'benchmark' by identifying it as a specific type of measurement (measurable baseline) taken at a particular time (before optimization) with a specific purpose (determining if changes improved performance). It distinguishes this from mere measurement or observation and emphasizes the comparative nature of benchmarking as a logical necessity for determining actual improvement.
Connections
Defines (58)
AxiomExponential Information DecayAxiomExtended Cognition ThesisAxiomDirected Attention as Depletable ResourceAxiomHindsight Bias and Calibration NecessityAxiomTwo-Level Metacognitive ArchitectureAxiomConversational Memory Asymmetry From Production PlanningAxiomUltradian and Circadian Cognitive RhythmsAxiomPatterns Exist in Hierarchical Logical LevelsAxiomEmotional Hijacking of JudgmentAxiomSystematic Overconfidence TaxonomyAxiomEmotion as Systematic Cognitive ModulatorAxiomNatural Frequency Format AdvantageAxiomCognition Operates Through Dual Processing SystemsAxiomValid Intuition Requires Stable Regularities and FeedbackAxiomMental States Are Cognitively ImputableAxiomCognitive and Affective Empathy Are DistinctAxiomAutomatic Pattern PerceptionAxiomDunbar's Number Limits Stable RelationshipsAxiomBasic-Level Category PrivilegeAxiomConstrual Level Effects on PerceptionAxiomPiagetian Equilibration Through Schema DynamicsAxiomFlexible Context-Dependent CategorizationAxiomPeople interpret failure as either evidence about theirAxiomEvery detection system faces a fundamental tradeoff betweenAxiomTask switching between different types of cognitive workAxiomWhen estimating future task duration, people naturally adoptAxiomExpert performance in complex domains requires deliberateAxiomCognitive performance follows an inverted-U relationshipPrincipleDesign capture and retrieval systems to minimize frictionPrincipleApply the same tags to notes from different domains whenPrincipleProcess inbox items in two distinct passes—first clarifyingPrincipleInstrument systems with the minimum number of metrics thatPrincipleMatch monitoring frequency to the rate at which meaningfulPrincipleDedicate focused, time-boxed sessions to improving onePrincipleDesign optimization sessions to create moderate activationPrincipleHold measurement protocol constant across before-and-afterPrincipleBenchmark efficiency, accuracy, and quality dimensionsPrincipleIdentify and eliminate shared dependencies across multiplePrincipleBlock your measured peak attention hours on your calendar asPrincipleWhen large language models express 90%+ linguisticPrincipleIdentify your information pipeline bottleneck by testingPrincipleShift focus from event reflection to system reflection whenPrincipleUse goals to set directional targets but rely on systems ofPrincipleFor complex cognitive tasks that resist starting, designPrincipleDesign micro-chain final links to produce one minimal unitPrincipleDesign social chain links to specify your own behaviorPrincipleIdentify the essential function each routine serves andPrincipleWeight emotional data more heavily in domains where you havePrincipleWhen holding more power in a relationship (formal authority,PrincipleReview accumulated notes in batches spanning weeks or monthsPrincipleUse satisficing decision rules (define 'good enough'PrincipleAggregate predictions by confidence level and compare statedPrincipleDesign physical and digital workspaces to afford only thePrincipleDocument not only what tools you use but the completePrincipleTest each candidate classification dimension by askingPrincipleWhen presenting complex information to diverse audiences,PrincipleScore tasks on three dimensions — irreversibility (cost ofPrincipleSpecify every delegation with five components: concrete