Quantum computing is currently taking off as a research field and quantum supremacy has been claimed recently. Starting from its potentially disruptive applications I will review the three quantum computing approaches of near-term uncorrected algorithms, fault tolerant computing, and quantum annealing. I will describe how the European flagship project OpenSuperQ approaches building a 50-100 qubit quantum computers as a research infrastructure. I will describe the view that the way how to control and calibrate a quantum computer is a crucial component and will ultimately allow to connect the three paradigms into the most effective way of using a quantum machine.