The irreducible epistemic atoms underlying the curriculum. 4,828 atoms across 8 types and 2 molecules
Design decision rules with 3-5 criteria rather than complex multi-variable models, as simple heuristics are more robust across variable conditions.
Build complex behavioral systems through incremental extension of simple working systems rather than top-down comprehensive design.
Scope each behavioral agent to a single trigger-action pair to prevent cascading failures and enable independent refinement of system components.
External documentation of cognitive processes converts them from degradable tacit knowledge into stable, improvable explicit knowledge that can be systematically reviewed and refined.
Under conditions of stress, fatigue, or emotional pressure, humans execute degraded versions of cognitive processes without awareness—making externalized checklists necessary for consistency when stakes are highest.
Attempting to articulate a process in writing surfaces logical gaps, unstated assumptions, and missing edge cases that remain invisible during internal rehearsal.
Prospective hindsight—imagining a future failure has already occurred—increases identification of causal factors by 30% compared to forward-looking prediction by shifting cognition from advocacy to explanation mode.
High-confidence errors produce stronger learning when corrected than low-confidence errors because prediction violation mobilizes additional attentional and encoding resources.
Interpreting failure as information about design rather than evidence about capability enables systematic improvement, while interpreting failure as identity-threat triggers avoidance and prevents learning.
Blameless analysis of failures generates more accurate diagnostic information than blame-based analysis because fear of punishment suppresses error reporting and distorts causal attribution.
Cognitive distortions in underlying schemas cause agents to execute flawlessly while producing systematically wrong outputs, making efficiency without accuracy high-speed drift.
Automated processes stop being examined, causing schemas to drift from reality through assimilation without accommodation—making periodic schema validation necessary for reliable automation.
Social interactions benefit from designed response protocols (social agents) rather than improvisation because they recur predictably, trigger emotion before cognition, and have consistent failure modes shaped by historical reinforcement.
Separating observation from evaluation before emotional response fully forms enables cognitive reappraisal, which is more effective than post-response suppression for emotion regulation.
Converting emotional responses into underlying needs redirects attention from other-person behavior to self-requirements, generating response options that reactive scripts exclude.
Structured feedback protocols that separate situation-behavior-impact prevent character judgment while maintaining specificity, enabling feedback delivery without triggering defensive responses.
Social agents produce emotional regulation as a byproduct of following structured procedures rather than requiring trait-level calmness, making effective social behavior accessible through process rather than personality change.
Every decision automation system must include explicit override conditions that specify when to suspend the automated rule and return to full deliberation.
Before entering a difficult conversation, explicitly articulate both your perspective and the other person's perspective, then identify where the two stories diverge as the locus of actual dialogue.
Pre-solve communication structure through templates so cognitive resources can focus on content rather than form.
Design stress management responses around simple physiological protocols that can execute under cognitive load, not complex cognitive strategies that require intact executive function.
Establish a strict priority ordering between agents so that when conflicts occur, the higher-priority agent's instruction is automatically implemented without real-time deliberation.
Build feedback loops into agent systems through regular review of whether agents fired, whether they produced intended outcomes, and whether conditions have changed that invalidate their design.
Decompose complex recurring behaviors into narrow-scope agents with clear interfaces rather than building multi-purpose agents, as complexity should live in agent interactions not in individual agent logic.