When you think of the words “optimal human performance”, what scene comes to mind? Maybe an Olympic athlete sprinting 100m in just barely longer than 9 seconds? Or a child prodigy playing a complex concerto on piano or violin? Now think of the military: do you picture an Army Ranger or Navy SEAL? In those two examples, it is not difficult to imagine a soldier in peak physical condition, but it is critical to acknowledge that every human has differences in mental and physical performance, and some could respond less optimally to certain aspects of training or to specific interventions, including changes in diet, exercise, or brain stimulation. What this means is that following general methods and paradigms in training—specifically those based on “group averages”—might be unproductive or even harmful to individuals that for some reason would appear to be “outliers” relative to others. And while the term “outlier” does not infer a negative predisposition for an individual, it does underscore the concept that their performance optimization might not fit precisely with the other members of a group in training that respond well to the “average” performance-improvement regimen. This is where DARPA is aiming to make a difference by searching for alternatives to the traditional “one-size-fits-all” approach, with their Teaching AI to Leverage Overlooked Residuals (TAILOR) program. By using an alternative AI approach—“third wave AI”—they want performers to focus on at least one area of human performance and provide malleable solutions that are able to adapt to altered circumstances, environments, and related information, in order to permit abstraction that would not normally be achieved from rigid training and performance optimization paradigms. The program solicitation states: “Capabilities sought under TAILOR are likely to require moving beyond brute-force statistical learning and black box predictions and instead begin to use contextual reasoning to make counterfactual predictions (i.e., ‘if person X had been given intervention Y, then they would have had outcome Z’). By incorporating biological, psychological, and social factors that give rise to individual differences, TAILOR hypothesizes that AI can transform these factors to better predict individualized human performance optimization outcomes and adapt if those factors change.” So, once provided data and a goal for a specific individual or team, a developed tool would be used to evaluate how much that individual or team would benefit from a given intervention, but it should also be capable of accounting for counterfactuals—examples that would respond differently to an intervention—that arise from the analysis. While the public might have a ubiquitous idea of the “optimal” soldier, the TAILOR program might well be key to helping unlock the best performance capabilities from every brave individual that has stepped forward to serve their country in close combat.