IBM: Google didn't achieve Quantum Supremacy
What's the story
Last month, a leaked paper from Google made the claim of having achieved quantum supremacy, a major benchmark in the world of quantum computing. The publication drew immediate attention from tech enthusiasts around the globe but was soon taken down. Now, a group of researchers from IBM, which is another player in the quantum computing arena, has disputed that claim. Here's why.
Meaning
First, you should know what is quantum supremacy
For years, leading tech giants, including Google and IBM, have been building quantum computers - machines leveraging the phenomena of quantum mechanics to deliver exceptional computing power. However, it is one thing to develop a quantum computer and completely another to achieve quantum supremacy. The latter, in particular, is the breakthrough point at which a quantum computer solves problems that a classical computer cannot.
Google's claim
Google claimed to have achieved quantum supremacy
In the last week of September, a Google researcher claimed that the tech giant has achieved quantum supremacy with its own 53-qubit quantum computer - the Sycamore. He said, in the leaked paper, that the quantum machine took just 3 minutes and 20 seconds to perform a calculation that IBM Summit, the world's most powerful classical computer, would handle in 10,000 years.
IBM's stance
IBM's quantum research head, however, disputes the paper
As Google's paper drew attention, Dario Gil, the head of quantum research at IBM, described the news as indefensible and misleading. "Quantum computers are not 'supreme' against classical computers because of a laboratory experiment designed to essentially implement one very specific quantum sampling procedure with no practical applications," he had said while refuting Google's achievement.
Proof
More evidence released by IBM
Now, backing-up Gil's point, IBM has released a paper proving Google made an error in estimating that the calculation they nailed with their quantum computer would have been handled in 10,000 years by a classical machine. "An ideal simulation of the same task can be performed on a classical system in 2.5 days," the company said, adding that 2.5 days is the worst-case scenario.
Error
Google made an error in estimating RAM consumption
The IBM report added Google made an error in estimating the RAM requirement for running a quantum simulation in a classical computer. It assumed that the task would eat more RAM, ultimately getting to the overstated 10,000-year estimate. On the contrary, IBM used both RAM and hard drive for the task as well as other performance optimization techniques to come to the 2.5-day result.