Skip to content
How to ThinkIn the Age of AI

Concepts

The irreducible epistemic atoms underlying the curriculum. 4,828 atoms across 8 types

Rule

Update decision agents after every activation based…

After each decision agent activation, update the agent's criteria based on whether they produced a good outcome, treating the agent as a living heuristic that improves with each use rather than a permanent law.

1 lessondecision-makingbehavior-changecontinuous-improvement
Rule

Add 'Did I use the right framework?' to every decision…

In post-decision review, explicitly add the meta-question 'Did I use the right framework for this decision?' and note framework-decision mismatches (comprehensive analysis on trivial reversible choices, satisficing on irreversible high-stakes decisions) to build your personal routing table.

1 lessondecision-makingmeta-decisionreview
Rule

After recurring activities, spend 60 seconds recording…

Add a 60-second structured observation step immediately after recurring activities, recording one sentence about output and one about potential changes, to convert open-loop repetition into closed-loop learning.

1 lessonfeedback-loopscontinuous-improvementreflection
Rule

A complete feedback loop needs three elements

For any recurring activity, explicitly define three elements—the specific output being measured, the standard for comparison, and the adjustment rule triggered by deviation—to create a complete minimal feedback loop.

1 lessonfeedback-loopssystems-thinkingcontinuous-improvement
Rule

Translate every discrepancy into a specific behavioral…

When a feedback loop identifies a discrepancy between current and target state, translate the evaluation into a specific behavioral adjustment for the next cycle rather than stopping at awareness, because learning occurs during adjustment not observation.

1 lessonfeedback-loopscontinuous-improvementaction-bias
Rule

After 10+ post-action reviews, analyze in aggregate…

After accumulating 10+ post-action reviews, analyze them in aggregate to identify structural causes appearing across multiple unrelated tasks—these recurring patterns indicate systemic tendencies requiring architectural fixes not isolated corrections.

1 lessonretrospectivespattern-recognitionmeta-analysis
Rule

After one cycle, review the corrector itself — did…

After deploying a self-correcting mechanism for one cycle period, add a meta-correction review asking whether the correction actually prevented the target error, adjusting the corrector itself if it failed.

1 lessonmeta-systemserror-correctionvalidation