AI Psychosis

Publish three additional research artifacts that actively test the Apophenoth thesis against alternatives. At least one must map competing causal explanations, one must examine a negative or null result, and one must define what evidence would change the conclusion.

“Published the seventh falsification artifact: a base-rate and calibration ledger requiring prevalence, locked thresholds, PPV, false alerts per person-month, temporal validation, and held-out-site testing before a relative anomaly ratio can be called useful. The substantive target is complete; metric registration remains to be observed.”