The Real Deal On IBM’s Latest Quantum Computing Announcement
Today, IBM made another splash in the quantum computing world by announcing its latest “Eagle” chip, which boasts 127 qubits. This is part of IBM’s public roadmap to larger machines over the next few years: “Osprey” in 2022 with 433 qubits and a “Condor” chip with 1,121 qubits. This is a big announcement, but it contains the potential to push the “quantum hype meter” into overdrive. Consider the following attention-grabbing headlines posted by the media online just today:
“IBM debuts quantum machine it says no standard computer can match.”
Or here is one that is a little more nuanced:
“The company says it will be close to what is called ‘quantum advantage’ — the point at which quantum computers can beat classical computers.”
The media is great at getting information out fast. But often their statements, while factually correct, lack proper context. This leads to a swirl of client questions in my inbox.
Unraveling The Announcement
Here’s what you need to know so you don’t go chasing your tail:
- Quantum computers are and will remain specialized problem solvers. Keep in mind that quantum computers will never replace today’s computers. Therefore, you must understand what phrases like “can beat” actually mean. Quantum computers are meant to take problems that classical computers struggle with and potentially solve them several to many times faster. This will let us optimize logistics and transportation routes, retail product assortments, or lending decisions. It will also let us eventually design new materials and drugs. There are many other esoteric problems in math and science that could lead to new discoveries as well as a revolution in the way we communicate information. But all this is going to happen modestly at first, and today’s digital computers will still do most of the day-to-day work.
- The power of quantum computers is hard to measure and compare across vendors. We have a mental model for computing power: number of bits, number of “cores,” speed of clock cycles, and memory. This lets us laypeople judge, for example, that a 64-bit, 3.3 GHz, 8-core, 2GB machine is better than one with lower specs. Quantum computers have no such equivalent. They have qubits, but not all qubits are equal across different types of machines; the quality of the qubits matters a lot. Also, there is no such thing as quantum memory, nor can quantum chips be connected in parallel today. Lastly, clock cycles must be measured on quantum circuit operations and not on binary register operations per second. All this means that there is no standard model yet for figuring out how fast the state of the art across vendors is progressing. Therefore, making a call that a certain number of qubits equals an advantage is misleading if not understood in context.
- Quantum advantage won’t come at once across all types of problems. The problem with statements like the ones above is that most people tend to read them and assume that quantum advantage — that is, when a quantum computer can do something more cost effectively than a classical computer — will be achieved all at once. This is not true. Quantum advantage will be achieved one tiny problem in one business or scientific domain at a time. Chemistry is a good guess as to where it will happen first because molecular modeling and simulation is inherently quantum in nature already, so quantum computers are a natural fit. IBM and most experts still think it will be 10 to 20 more years before large-scale quantum computers are doing really scary things like breaking crypto. And by then, we will likely have quantum-resistant or safe algorithms and architecture to protect data.
So, What Now?
What should you think of IBM’s announcement? Just as I’ve posted on other quantum computing announcements like this, it’s a milestone. It’s a significant, concrete commitment to fabricating quantum computers that are increasing in power along a trajectory toward quantum advantage. IBM deeply understands the three points above. Therefore, their announcement of probable quantum advantage by Condor is credible, but not a given.
What does that mean for you? It’s time to ensure that you have enough knowledge to make smart bets on where and when quantum advantage is likely to be reached. You need to be ready.