快猫短视频

The past decade has seen significant advances and investment in quantum computing, and yet the devices we have today essentially have no practical purpose. That is down to two main reasons 鈥� the first being that the qubits, or quantum bits, that make up today鈥檚 machines still struggle with noise, or errors, that we are only just learning to correct. The second is that devices that could solve practical problems are expected to require many more qubits than even the biggest quantum computers currently have.

In 2018, at the California Institute of Technology coined the phrase 鈥渘oisy intermediate-scale quantum鈥�, or NISQ, to describe this current era of quantum computing 鈥� devices that are promising but imperfect, essentially a stepping stone towards larger machines that will crackle with fewer errors, eventually becoming 鈥渇ault tolerant鈥�. Now, he tells 快猫短视频 how he is setting his sights on the next era of quantum computing: the 鈥渕egaquop machine鈥�.

Quantum computers have finally arrived, but will they ever be useful?

Hundreds of quantum computing firms around the world are racing to commercialise these once-exotic devices, but the jury is still out on who is going to pull ahead and produce a machine that actually does something useful

Karmela Padavic-Callaghan: Your idea of the 鈥淣ISQ era鈥� has been fully embraced by the quantum computing field 鈥� was that a surprise?

John Preskill: I didn鈥檛 expect the term to be so widely embraced, but I did feel like we needed a word to succinctly express the idea that we were entering the era where we had machines that could perform at least some tasks that were very hard to simulate with conventional computers.

I was trying to emphasise the message that we鈥檙e probably going to need fault tolerance to really run applications that most people will care about. But in the meantime, we had an opportunity to experiment with these NISQ machines and maybe find some things that they can do that are useful, at least to scientists and possibly to businesses.

You have now named a new frontier for the quantum computing community, a device that can perform a million quantum operations, or the 鈥渕egaquop machine鈥�. Why?

Free newsletter

Sign up to The Daily

Sign up to newsletter

I think it鈥檚 important to have goals like the megaquop machine, something to shoot for. Why did I choose a million operations? First of all, because I don鈥檛 think we鈥檒l be able to get into that regime without using error correction and achieving fault tolerance. It is significantly beyond the quantum circuits that we can run without error correction, and it鈥檚 kind of on the edge of where we can start to do simulations of matter that I think at least scientists will find significantly more informative than what we can do in the NISQ era. Right now, people have run circuits with about 13,000 operations but you鈥檝e got to do a lot of error mitigation and the amount of physics you can get out is rather modest.

Does the megaquop machine have a counterpart in the history of traditional computing?

Computers were developed in the late 1940s and early 50s, largely motivated by wanting to use them to simulate physical systems. That will be interesting with machines at the megaquop scale. Those will be the most important applications, with practical implications for chemistry and material science. There鈥檚 something analogous to conventional computing there in that we鈥檙e going to start by using these machines to do science, and not necessarily for applications that affect the average person directly.

My other thought is, people often say that with quantum computing we don鈥檛 have the transistor yet, we鈥檙e still in the vacuum tube era or something, so there could be a technological shift that occurs when we come up with better quantum platforms that we can scale more easily.

How PsiQuantum plans to build world鈥檚 largest quantum computer by 2027

With an investment of AU$1 billion, PsiQuantum is planning to build a photonic quantum computer with a million qubits, far larger than any in existence today - and the firm says it will be ready in just two years

You鈥檝e said that determining what the megaquop machine will be useful for is a 鈥渃ompelling challenge for the quantum community鈥�. How can the growing number of quantum computing companies be part of meeting that challenge?

People should be thinking about, okay, what can I now do that I that I couldn鈥檛 do before? We should continue to try to apply some fresh thinking to what the applications are.

I have the concern that in the industry, there are overly optimistic expectations about the economic impact of quantum computing in terms of the time scale for reaching it. And that was true when we were talking about NISQ devices some few years back, and I think it鈥檚 going to be true of early fault tolerance as well. I think we really have a long road ahead of us to get to real economic value.

If you woke up tomorrow and someone handed you a megaquop machine, what would you do with it?

I have an interest in quantum field theory and the new types of phenomena that occur in strongly coupled field theories. We have pretty good tools for conventional computers for simulating these theories in one dimension, but the tools in two dimensions are not very good at all for conventional computers. So, for me, that鈥檚 an opportunity to do something interesting. The megaquop machine would maybe be a little short of what we鈥檒l need, but a good start for studying some phenomena that are just beyond the reach of conventional computers.