Comment on this story
comment
The hype surrounding the rise of ChatGPT and the supposed ground Google is giving to Microsoft Corp. in the search wars. and OpenAI has overshadowed more important developments in computing, advances that will have far greater implications than which website offers better tax advice.
Quantum computing is the holy grail of scientists and researchers, but it’s still decades away from reality. However, Google’s parent company Alphabet Inc. tipped the ball last month with the news that it had found ways to improve one of the burgeoning field’s biggest problems: accuracy.
To date, all calculations are performed on a binary scale. A piece of information is stored as either 1 or 0, and these binary units (bits) are grouped together for further computation. For example, we need 4 bits to store the number 8 (1000 binary). It’s slow and clunky, but at least it’s simple and accurate. For nearly seven decades, silicon chips have been holding and processing bits.
Quantum bits – qubits – can store data in more than two forms (they can be 1 and 0 at the same time). This means that larger blocks of information can be processed in a given time. One of the many disadvantages is that the physical manifestation of a qubit requires supercold temperatures – just above 0 degrees Kelvin – and is susceptible to even the smallest perturbations, such as light. They are also error-prone, which is a big problem in computing.
In an article published in Nature last month, Google claims to have made a major breakthrough in an important subfield called quantum error correction. The approach is very simple. Rather than relying on individual physical qubits, scientists store information across many physical qubits, but then view that collection as a single one (called a logical qubit).
Google had theorized that combining a larger number of physical qubits into a single logical qubit would reduce the error rate. In their research, outlined in a blog post by Chief Executive Officer Sundar Pichai, the team found that a logical qubit made up of 49 physical qubits actually outperforms one made up of 17.
In reality, using 49 qubits to handle just a single logical one sounds inefficient and even overkill. Imagine storing your photos across 49 hard drives only to ensure that a single hard drive is overall error-free. But given the enormous potential of quantum computing, even such small steps represent significant progress.
More importantly, it gives the broader scientific community a foundation on which to build on this knowledge to further advance related fields such as materials science, mathematics and electrical engineering, all of which are needed to make an actual quantum computer a reality. The hope of building a system that can solve a problem that no current machine could feasibly solve is called quantum supremacy.(1)
Four years ago, Google said it completed a test in 200 seconds for a task that would take a conventional supercomputer thousands of years to complete, proof we are on the road to quantum supremacy.
But as with artificial intelligence tools like ChatGPT, proving that they work is only part of the puzzle. High accuracy and low error rates — something newer chatbots are prone to — remain elusive. Improvements on this front are a key goal for developers of both technologies, with OpenAI saying this week that its new GPT-4 is 40% more likely to deliver factual results than its predecessor.
Unfortunately, a cold computer processing data isn’t as fun as a digital assistant that can write limericks or compose a school essay. But in the future, these breakthroughs will be as comparable as the entertainment value of television is to the world-changing achievement of landing a man on the moon.
More from the Bloomberg Opinion:
• The race is on against a threat that doesn’t exist: Tim Culpan
• Google faces a serious threat from ChatGPT: Parmy Olson
• US Chip Curbs highlights cracks in China’s AI strategy: Tim Culpan
(1) Feasible is a nebulous term, but generally means completion in a reasonable timeframe, such as minutes or days rather than years.
This column does not necessarily represent the opinion of the editors or of Bloomberg LP and its owners.
Tim Culpan is a columnist for Bloomberg Opinion covering technology in Asia. He was previously a technology reporter for Bloomberg News.
For more stories like this, visit bloomberg.com/opinion
