Combining big data with personalized medicine is an unprecedented opportunity. It will probably be cheaper than current practices in the long term, particularly given the questionable effectiveness of many medications (see Nature 517, 540; 2015). Success in this endeavour will depend on training the next generation of clinicians and data scientists to deploy terabytes of data to select from a range of diagnosis and treatment options. Undergraduate and graduate bioinformatics programmes need to embrace data-analytics courses geared towards generating a new type of medical specialist—one who no longer needs to see patients, just their data.