Quantum computers get faster by doing one step instead of a thousand
I read something this week that I keep coming back to. Researchers at Chalmers University in Sweden found a way to run certain quantum operations in a single control cycle instead of thousands. For some operations that works out to more than a thousand times faster.
Speed matters here more than it sounds like it should. Time is the main problem for a quantum computer. The information sits in a fragile state, and every extra microsecond it spends exposed to heat, vibration and stray fields is another chance for it to get scrambled. So a slow operation isn't just slow. It's also more likely to come out wrong. Cut the time and you cut the errors.
The work is on bosonic codes. Instead of putting one bit of quantum information in one physical qubit, you store it in the microwave field inside a resonator, spread across many photon states. That spreading gives you some built-in protection against errors. The catch has always been that manipulating those states took long sequences of carefully timed drive pulses. The Chalmers method, which they call quantum lattice gates, does the job within one period of the drive.
The paper is "Single-Period Floquet Control of Bosonic Codes with Quantum Lattice Gates", published in Physical Review Letters, with Lei Du and Tangyou Huang leading the work.
I should be clear about what this is. It's a control technique, not a machine. Nobody is shipping a fault-tolerant quantum computer because of it. But errors are the central problem in this field right now, and shortening the operations goes at that problem directly instead of piling more correction on top. That seems like the right place to be working.
Me llamó la atención que el método de Chalmers logra ejecutar una puerta cuántica en un solo ciclo, más de mil veces más rápido que antes; la diferencia se nota en la reducción de errores por decoherencia. Esto es genial para los bosonic codes porque simplifica mucho la corrección práctica. 🚀