In the modern world, computer applications operate in an environment of increasing complexity. Ubiquitously connected, they rely on continuously changing interfaces with the external world to drive computations based on patterns in big data. In this video, MIRA MEZINI considers how programming technology can optimally respond to this radically changed operating environment. Arguing that accidental complexity can be overcome by the invention of new programming abstractions that directly express notions in the world, Mezini employs mathematical modeling and empirical benchmarking to validate her hypotheses. Making computer systems easier and cheaper to evolve and maintain by reducing their complexity, the solutions proposed here are also key enablers for the next generation of data-driven applications (e.g. software for self-driving cars and environmental modeling).