“Quantum computers are just too slow”

0
29

James S. Clarke has been director of the quantum hardware research department at the chipmaker since 2015 Intel in Portland, Oregon. He has co-authored over 50 scientific papers and holds various patents. Today he is working on products that will lead Intel into the next century – including quantum computers.

Mr. Clarke, why do we need quantum computers?
This is a technology that we routinely solve problems that we can only solve with supercomputers – or even with them.

Is this all a buzzword or reality?
It's both. The hype about quantum computers is a bit ahead of reality. But real progress is being made. The question is when the technology will be ready. At Intel, we believe that it may take ten years for quantum computers to change all our lives. If you look at the history of the semiconductor industry, you see that no real progress has become reality in less than ten years.

Are there any real quantum computers?
Right all-round quantum computers do not yet exist – those that can solve all sorts of problems. Today's machines are designed for very specific tasks.

Why are the governments of the world particularly interested in quantum computers?
This also has something to do with encryption and decryption of data. To crack an RSA encryption, you need a universal, full-fledged quantum computer.

James S. Clarke

“Artificial intelligence will be much more common in everyday life than quantum computers.”

(Photo: Twitter screenshot)

Where is Intel in the race?
We're researching a “super-conducting” computer, as well IBM. Google and others, but also a so-called “spin qubit in silicon”.

What is that please?
This is something like a transistor where only one electron defines the state of the qubit, zero or one. We are working on a special form of a single-electron transistor. Why is that so sexy? Because of the great similarity to normal transistors. Intel delivers 400 quadrillion transistors per year. That's one transistor per person per day per person on the planet. If we can build a quantum computer based on transistor technology, we believe we will have a big advantage.

Why exactly?
At the end of the day, everything is qubits. But a spin qubit would be a million times smaller than a traditional qubit. And the lifetime of a qubit, the time to complete a task, could be longer. If you switch a qubit from zero to unity today, it will drop back to zero in microseconds. By comparison, saving something on a USB flash drive is still there after a year.

Quantum computers must be cooled down to absolute zero, which is minus 273 degrees Celsius. Would that be necessary for the spin qubits you described?
Spin qubits could be run at a slightly higher temperature. This would allow us to bring additional components closer to the qubits, which would otherwise not survive the cold.

What would be the benefits?
It would enable us to better control qubits, increase lifetime, and build larger systems.

How realistic is it all to-day if you say that it will be another ten years before we have universal quantum computers?
Today's computers are a few qubits up to several tens of qubits. It's fun to play with. But what we need are computers with thousands or millions of qubits to do the things that are impossible today. It is not unthinkable that we will have chips with a million qubits that you can hold in your hand like today's Intel chips for ordinary machines.

What is the biggest problem?
Looking at the huge quantum “refrigerators” you need today is very impressive. But it takes days to cool them down, basically they are slow. I fear that even if we had hundreds of them, we would not be able to go through enough innovation cycles in ten years to come to a finished result, the universal quantum computer. We just do not learn fast enough.

What should corporate leaders do who wonder if they will need quantum computers in the future? Wait ten years and then get started?
It's not like nothing happens for nine years, and then it's suddenly there. It's a technology that is coming, and I should not be in the back seat as a business leader when it starts.

How can that be done?
You can look for an academic partner, for example the TU Delft in Stanford. There are so many good colleges doing research on this topic. They all work to train the quantum specialists of the future.

How can companies prepare for this development?
Let's say the company is involved in research and already works with supercomputers today. Then the question is how and which IT problems should be transferred to quantum computers. It can start with a handful of people and strategic partnerships with hardware manufacturers or universities. Can a quantum computer help the company in the future? To clarify this question, only a relatively small upfront investment is needed.

Everyone is talking about Artificial Intelligence today. What is more important for companies? AI or quantum computer?
Both have their place. But artificial intelligence will be much more commonplace in everyday life than quantum computers. However, it is not an either-or-decision. Both is important. Just do not let ourselves be dazzled by giant, shiny quantum “refrigerators” and believe they solve all our problems.

Mr. Clarke, thank you for the interview.

More: In the research field of quantum computers, Google could have achieved a breakthrough. Experts expect a boost for the future technology.

Quantum Computer (t) Intel (t) James S. Clarke (t) Digital Revolution (t) Qubit (t) Spin Qubit (t) IBM (t) Google (t) Computer Engineering Supercomputer (t) Interview (t ) Innovation (t) Digitalization (t) Artificial Intelligence KI (t) Delft University of Technology (t) Semiconductor (t) Components

LEAVE A REPLY

Please enter your comment!
Please enter your name here

This site uses Akismet to reduce spam. Learn how your comment data is processed.