Quantum computing edges toward mainstream

Quantum computer with cooling elements hanging from ceiling
IBM "Q" quantum computer on display in San Francisco. Photo: Scott Rosenberg/Axios

Quantum computing will enter the mainstream faster than most of us realize, a panel of experts told a San Francisco crowd earlier this week — with some important real-world applications emerging within five years.

Why it matters: Quantum computers won't replace the semiconductor-based electronic computers we live with today, but they might speed up the solving of fiendishly difficult problems in fields like molecular imaging, cryptography, probability and artificial intelligence. Once they do that, they will make fortunes, disrupt businesses — and open the door to a host of potential new problems.

"Within 5 years, we're going to see something that makes everyone look up and say, 'Wow, how is this possible?'" said Arvind Krishna, director of research at IBM, at a Tuesday event hosted by the Churchill Club.

How they work: Quantum computers exploit characteristics of subatomic particles — superposition and entanglement — to perform computations using quantum bits or qubits. Conventional bits are binary — they're either 1 or 0, "on" or "off." A qubit can be on, off, or both at the same time.

That enables quantum computers to achieve "a tremendous speedup in the parallel-ness of computation," said Kam Moler, Stanford professor of physics. While Google's 72-qubit computer is the biggest yet, experts say 100-qubit machines are on the horizon.

But, but, but:

  • Quantum computers are still really hard to build. Qubits need to operate at extreme colds, near absolute zero, to keep errors from creeping into their work. Even there, "decoherence" eventually kicks in — random information replaces the data you input.
  • The high error rates in quantum computing can be corrected for in programming, but even so, they make it less suited for calculations that require exactness and better for those involving probability.
  • Creating software for quantum computers requires programmers to rethink basic principles that have held sway in "classical computing" for 75 years.

The crypt-apocalypse: One realm that quantum computing could disrupt sooner rather than later is cryptography. Today's encryption systems typically involve factoring large numbers, and quantum computers can do that much faster than regular computers.

That means once they get good enough, quantum machines will be able to crack most of today's encryption techniques. By then experts hope to have new techniques in hand for protecting contemporary data, but the data you securely store today will become insecure at that time.

"If you want to keep data safe for 10 years, you should seriously consider moving to alternate encryption now," Krishna said.

Where they'll be: Near-absolute zero isn't coming to your desktop. Quantum computers will live in special facilities for the foreseeable future, but they'll be accessible, and programmable, from the cloud. According to Krishna, 80,000 people have already taken runs at writing programs for IBM's Q quantum computer from the web.