Researchers have used machine learning to crack the code governing charge transfer and color emission in chains of molecules, with applications in data storage, security inks, organic light-emitting diodes (OLEDs), and solar energy.