This invited review by VanRaden et al. evaluates the reliability of genomic predictions for North American Holstein bulls. The study utilized genotypes from 38,416 markers and genetic evaluations from 3,576 Holstein bulls born before 1999 to predict daughter deviations for 1,759 bulls born from 1999 to 2002. Genomic predictions were computed using both linear and nonlinear models, with the latter accounting for major genes. The combined predictions, which included official parent averages and genomic predictions, were more accurate than official parent averages for all 27 traits, with coefficients of determination (R²) ranging from 0.05 to 0.38 higher than those from parent averages alone. The greatest benefits were observed for fat percentage due to a known gene with a large effect. The realized reliabilities of combined genomic predictions were 23% higher than those of parent averages, equivalent to 11 additional daughter records. The study also found that doubling the number of bulls genotyped provided more benefits in reliability than doubling the number of markers. Genomic prediction improves reliability by tracing the inheritance of genes, even those with small effects. The review concludes that genomic methods are effective in determining which genes animals share, and the implementation of genomic predictions will enhance genetic progress in dairy cattle breeding.This invited review by VanRaden et al. evaluates the reliability of genomic predictions for North American Holstein bulls. The study utilized genotypes from 38,416 markers and genetic evaluations from 3,576 Holstein bulls born before 1999 to predict daughter deviations for 1,759 bulls born from 1999 to 2002. Genomic predictions were computed using both linear and nonlinear models, with the latter accounting for major genes. The combined predictions, which included official parent averages and genomic predictions, were more accurate than official parent averages for all 27 traits, with coefficients of determination (R²) ranging from 0.05 to 0.38 higher than those from parent averages alone. The greatest benefits were observed for fat percentage due to a known gene with a large effect. The realized reliabilities of combined genomic predictions were 23% higher than those of parent averages, equivalent to 11 additional daughter records. The study also found that doubling the number of bulls genotyped provided more benefits in reliability than doubling the number of markers. Genomic prediction improves reliability by tracing the inheritance of genes, even those with small effects. The review concludes that genomic methods are effective in determining which genes animals share, and the implementation of genomic predictions will enhance genetic progress in dairy cattle breeding.