Duke Power used IQ tests to open higher departments to blacks, but Supreme Court required job-relatedness due to disparate minority impact. Meta-analysis finds mean IQ-job performance correlation of 0.53. Defenses confirm true correlation 0.5-0.6 after scrutiny. Military meta-analysis (N=10,088) shows g predicts performance stably over experience, adjusted r=0.39. Another Army study (N=4,039) finds cognitive ability correlates 0.63 with technical proficiency. Uncorrected for IQ reliability, real-world correlations are 0.37-0.60. Job performance unreliability understates true IQ correlation. High IQ yields large organizational externalities beyond individual output. Affirmative action in management inflicts massive costs from lower-IQ hires. IQ tests have high predictive validity for job performance despite construct validity debates. Publication bias ignored but priors suggest minimal impact. Corrections for range restriction and reliabilities are valid and necessary. Decreasing correlation over time claim is p-hackable and untrusted. Rater reliability corrections use inter-rater appropriately to account for imperfect assessments. Cognitive ability outperforms personality or interests for task proficiency. Vocational interests predict combat job proficiency but less than IQ overall. g-loaded tests like AFQT maintain validity across 31 military occupations.