The live instrument can describe those relative preferences under one frozen administration.
Method & evidence
Useful interpretation without false certainty.
Operator Instinct is built to compare six useful human operating responses fairly, preserve a clear research record, and say only what the current evidence can support.
The construct
Which useful move do you reach for first?
The instrument does not ask whether a person agrees with admirable statements. It places two credible operating responses in tension and observes the relative pattern across many choices.
45 paired situations
Every scoring screen presents two useful responses to a short machine-assisted work situation.
15 direct comparisons
Every one of the six instincts is compared with every other instinct three times.
Equal exposure
Every instinct appears exactly 15 times, so no instinct receives more opportunity than another.
Server-side scoring
Choice keys, item mappings, and scoring logic do not ship in participant-facing browser assets.
What the result means
A relative pattern, not six independent abilities.
The current score describes how often each instinct was selected when offered. Because each choice increases one part of a fixed profile, the dimensions are relative to one another within this administration.
01
A named identity interprets the leading pair; it is not a fixed diagnosis.
02
The six-sided signature visualizes the distribution; it is not a physical phase law.
03
Timing and uncertainty are stored as research evidence, not shown as a measure of human value.
Evidence ladder
Three claims. Three different tests.
Test-retest work, cognitive interviews, item calibration, and subgroup analysis must establish this.
That requires standardized work tasks and blinded outcome ratings beyond self-report.
Current claim boundary
What we say. What we refuse to imply.
Appropriate now
Maps the useful operating moves you reach for first; shows a relative six-instinct pattern; produces a private operating identity and practical guidance.
Not appropriate now
Scientifically identifies the best AI workers; predicts ROI; measures innate aptitude; establishes a percentile; validates hiring or promotion decisions.
Organization use
Developmental interpretation, shared language, cohort coverage, and responsible activation. No named employee ranking by default.
Validation path
More sophistication comes from evidence, not decorative math.
Comprehension
Cognitive interviews, reading-level review, ambiguity detection, and desirability testing.
Calibration
Real cohorts, item choice rates, side effects, uncertainty patterns, and hierarchical paired-comparison shadow scoring.
Stability & fairness
Retest studies, identity recovery, subgroup behavior, accessibility, and differential-functioning analysis.
Criterion validity
Blinded machine-assisted work simulations testing whether the six-instinct profile adds information beyond AI experience and domain expertise.
Research context
The problem is supported. This instrument must earn its own validation.
- McKinsey, The State of AI 2025 — broad adoption with materially lower scaling.
- IBM CEO Study 2025 — a gap between AI initiatives, expected ROI, and enterprise-wide scale.
- Anthropic, Claude Code Expertise 2026 — coding-specific evidence that expertise changes how people direct capable models.
These sources motivate the research question. They do not validate Operator Instinct or its identities.
Read the product through its limits
Serious science starts by refusing the easy overclaim.
Every production administration preserves its version boundary, private report, and explicit use limits while contributing to the evidence needed for future calibration.