Concepts
The irreducible epistemic atoms underlying the curriculum. 4,828 atoms across 8 types
Run simplified systems for two full cycles before…
Validate simplified systems by running them for at least two full cycles before declaring success, as one cycle cannot distinguish between successful simplification and lucky conditions.
Keep separate Candidate and Confirmed pattern lists…
Maintain two separate lists—'Pattern Candidates' and 'Confirmed Patterns'—promoting candidates only after they survive three independent observations, one alternative-explanation check, and one successful prediction.
Zero disconfirmations in a validation log =…
Treat a validation log with no disconfirmations as a warning signal of selective documentation rather than validation success, because unbiased testing inevitably produces some surprises.
Before finalizing a schema update, verify the new version…
Before finalizing any schema update, list five situations where the old schema produced accurate results and verify the new schema handles all five to ensure backwards compatibility.
Validate new feedback mechanisms in the first cycle
For each feedback mechanism you build, verify within the first cycle that the data reveals something unknown, that measurement effort is sustainable, and that you can specify one concrete adjustment based on results—if any component fails, redesign before continuing.
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.
Review new agents 7-14 days after deployment — remove…
Schedule agent addition reviews 7-14 days after deployment to verify whether actual coordination cost matches pre-addition estimates, removing the agent if cost exceeds estimate by more than 50%.
Validate your values hierarchy against 3 past decisions…
Test your stated values hierarchy against three past difficult decisions to verify whether it produces the choices you actually made—if it doesn't, either the hierarchy is miscalibrated or the past choices violated your values, and both discoveries are actionable.