The use of qubits to construct computers opens the door to dramatic computational speed ups for certain problems. The maturity of modern quantum computers has moved the field from being predominantly a quantum device-focused research area to also include practical quantum application focused research. In this talk I will discuss a new experimental result on a foundational aspect of how to program quantum computers. A central principle of classical computer programming is the equivalence between data and instructions about which operations to apply to that data. In quantum computers this equivalence is broken: Classical hardware is used to generate the sequence of instructions to be executed on the quantum data stored in the quantum computer. Our experiment shows for the first time how the instruction-data symmetry can be restored to quantum computers. We use superconducting qubits as a platform to implement high-fidelity quantum operations (>99.7% fidelity CPHASE gate) enabling the so-called Density Matrix Exponentiation algorithm, to generate these quantum instructions. This algorithm provides large quantum speedup for a family of other quantum algorithms, which I will briefly discuss.