Next-generation sequencing is now deeply integrated in clinical diagnostics of various genetic disorders, such as inborn errors of immunity. A central diagnostic challenge however is that most genetic variants found through sequencing cannot be classified as either benign or pathogenic and are labeled variants of uncertain significance (VUS). To begin not only reading a person's genome, but also attributing meaning to variants, we established massively scalable CRISPR-dependent base editing tools, performed directly in primary human T cells, which enable generation and mapping to phenotypes of interest. These approaches enable accelerated diagnosis and treatment implementation, as we have shown for Activated PI3K-delta syndrome, and provide a blueprint for annotating other gene-disease pairs in a similar manner. Furthermore, we reason that learning from these "experiments of nature" will inform targeted modifications to improve T cell function and therefore the building block for cellular cancer therapies. This work is funded by and NIH/NIAID R01 grant (R01AI199896), a CRI grant, and through additional foundations.