Quantum computing promises to be a revolutionary tool, making short work of equations that classical computers would struggle to ever complete. Yet the workhorse of the quantum device, known as a qubit, is a delicate object prone to collapsing.

Keeping enough qubits in their ideal state long enough for computations has so far proved a challenge.

In a new experiment, scientists were able to keep a qubit in that state for twice as long as normal. Along the way, they demonstrated the practicality of quantum error correction (QEC), a process that keeps quantum information intact for longer by introducing room for redundancy and error removal.

The idea of QEC has been around since the mid-90s, but it's now been shown to work in real time. Part of the reason for the experiment's success was the introduction of machine learning AI algorithms to tweak the error correction routine.

"For the first time, we have shown that making the system more redundant and actively detecting and correcting quantum errors provided a gain in the resilience of quantum information," says physicist Michel Devoret, from Yale University in Connecticut.

Qubits are objects as they exist in a mix of quantum states. Where classic objects can have absolute states, a qubit's version of the same state would be best described using probability. As a qubit interacts with other qubits, their probabilities become entangled in computationally useful ways.

Unfortunately, it's not just other qubits that can entwine their states with a non-decided object. Everything in the environment acts as 'noise', potentially influencing those delicate probabilities and making room for errors.

Part of the reason scientists have struggled to implement QEC is because it can introduce errors of its own. The extra space afforded for error correction can make the qubit even more vulnerable to interference from the surrounding environment.

Like many quantum physics experiments, this one was run at ultra-cold temperatures – a hundred times colder than outer space, in this case. The setup has to be carefully controlled in order to protect the qubit as much as possible.

The error-corrected qubit lasted for 1.8 milliseconds – only a blink as we might experience it, but an impressive span for a qubit operating on the quantum level. Now the research team will be able to refine the process further.

"Our experiment shows that quantum error correction is a real practical tool," says Devoret. "It's more than just a proof-of-principle demonstration."

While scientists are making significant strides forward in the development of quantum computers – and there are rudimentary quantum computers in use now – there's still a long way to go before the full potential of the technology is realized.

Reducing noise, improving stability, and upgrading error correction are all going to play a big part in getting closer towards full-scale, practical quantum computers that everyone can use.

In this case the breakthrough was down to several different factors, rather than one change. The QEC code was actually one from 2001, but improvements to it as well as upgrades to the quantum circuit fabrication process made a difference.

"There is no single breakthrough that enabled this result," says Volodymyr Sivak, a research scientist at Google and formerly at Yale University. "It's actually a combination of a whole bunch of different technologies that were developed in the past few years, which we combined in this experiment."

"Our experiment validates a cornerstone assumption of quantum computing, and this makes me very excited about the future of this field."

The research has been published in Nature.