Google's Quantum AI Lab has recently announced a significant milestone. The company revealed that its latest developed quantum computing chip, "Willow," can solve a computational challenge in under five minutes. In contrast, Google states that the world's fastest supercomputer would take approximately 1020 years to complete the same task, a timeframe vastly exceeding the age of the universe.
This marks a substantial advancement compared to Google's 2019 announcement, where their quantum processor solved a mathematical equation in three minutes, whereas a supercomputer would require ten thousand years. At that time, IBM challenged this claim.
Beyond performance enhancements, researchers have also identified a method to reduce errors, which Google regards as "one of the greatest challenges facing quantum computing." Unlike traditional bits that represent 0 or 1, quantum computing utilizes quantum bits (qubits), which can exist simultaneously in multiple states—such as 1, 0, or any value in between.
Google points out that qubits are prone to errors because they "tend to rapidly exchange information with their environment." However, Google's researchers have discovered that by introducing more qubits into the system and performing real-time error correction, the error rate can be reduced. These research findings have been published in the journal Nature.
Hartmut Neven, founder of Google Quantum AI, wrote in the company's blog that this historic achievement is termed "below the threshold"—reducing error rates while increasing the number of qubits. "To demonstrate genuine progress in error correction, it is essential to prove below-threshold performance. Since Peter Shor introduced quantum error correction in 1995, this has remained a significant challenge."
According to Neven, the "Willow" chip, equipped with 105 qubits, "now boasts state-of-the-art performance." Microsoft, Amazon, and IBM are also developing their own quantum computing systems.
Google's next goal is to perform the first "classically useful" computation—one that is "relevant to practical applications" and unattainable by traditional computers. In the future, Neven indicates that quantum technology will be "indispensable" for gathering artificial intelligence training data, ultimately aiding in "discovering new drugs, designing more efficient electric vehicle batteries, and accelerating advancements in fusion and alternative energy technologies."
