The efforts towards building a quantum computer range from the experimental demonstration of quantum algorithms to the miniaturisation driven by integrated photonics. Now superconducting circuits mark a significant step in fault-tolerant quantum computation, with researchers from the University of California and the University of Melbourne achieving repetitive error detection in a one-dimensional superconducting quantum device.

While there is an ongoing debate as to what platform is best suited to build a quantum computer, the key requirements are clear-cut: preserving coherence and ensuring scalability are on top of the list. Superconducting circuits constitute a scalable platform which ensures good control over individual quantum bits (or qubits). However, coherence is a fragile resource in these solid-state systems, and qubits can ‘lose’ their initial properties. For example, energy relaxation is thought to cause bit-flip errors (where a one flips to a zero or vice versa).

Such detrimental effects are intrinsically induced by the environment – be it in the form of material defects or coupling to external wiring – and can arise at any space-time point in the circuit. It is therefore crucial to devise strategies which preserve quantum states throughout error-prone computations. Reporting in *Nature*, Kelly *et al.* were able to protect a logical state (a sequence of zeroes or ones) from multiple, environment-induced errors by performing quantum error correction (QEC) in a superconducting circuit with long energy relaxation times.

The circuit appears as a linear array of nine qubits, a subset of which is measured in order to monitor the untouched data qubits carrying the precious information elaborated during a computation. This design allowed Kelly and coworkers to test a repetition code correcting for bit-flip errors – a one-dimensional version of surface codes for fault-tolerant quantum computing. With the data qubits initially prepared in a non-classical state, the authors showed that their QEC code preserves the ‘quantumness’ of the state. Most notably, the repetition code allows for the correction of errors occurring on a classical logical state distributed over the data qubits as the size of the circuit (and hence its complexity) increases.

The QEC strategy adopted by Kelly *et al.* brings together aspects previously addressed as individual issues. Their results suggest that the basic physical operations for successful error detection are already within experimental reach. An open challenge is the extension to two-dimensional qubit arrays as well as to many-cycle correction schemes.

*Nature* **519**, 66-69 (2015)