Disruption is coming with quantum computing | Hacker News Reader

Disruption is coming with quantum computing

singularityhub.com

59 pointsisaiahg29 comments

The phrase "quantum supremacy" sounds click-baity but it's actually bona-fide jargon. These guys are working with complexity theorists, such as Scott Aaronson and Michael Bremner, to nail down specific tasks that are going to be completely impossible for classical computing. Part of the motivation is to move beyond the whole D-wave saga, and build something that for-sure kicks the shit out of current classical computing technology. Calling it "quantum supremacy" is a nice bit of marketing, imho.

Earlier this year I sat down with the guy that runs the google quantum computing lab (Hartmut Neven) and chatted a bit about what they are building. One of the ideas is to be able to re-purpose silicon fab technology; whoever gets this to work (if it's possible) is going to win big.

There are many other approaches being explored, by many research labs around the world. One particularly interesting idea is to use photonics (think fiber-optics), which has the great advantage of not needing to be ridiculously cold [1].

Source: me doing a PhD in the area.

[1] https://arxiv.org/abs/1607.08535

Could you elaborate on "repurpose silicon fab technology"? I know pretty much nothing about quantum computing. How does the manufacturing process differ?

Most superconducting qubits, such as the "Transmons" used by Google and various other labs are made of Aluminium, as this is a material which has good properties and can be easily deposited using e.g. evaporation-based methods (i.e. heat Aluminium in a vacuum so it evaporates and gets deposited on the surface of a chip that you put nearby). The microwave circuits on the chips are typically made out of Niobium (another superconductor), which is not very suited for making qubits but which provides very good properties for building large-scale structures such as coplanar waveguides.

To fabricate a qubit, normally one first deposits Niobium on a Saphire/Silicon wafer and then etches structures into this using photo lithography. Afterwards, the Aluminium qubits get deposited using so-called shadow evaporation technique, where you create two layers of Aluminium separated by a thin insulating barrier that is created by first depositing a single layer of Aluminium, oxidizing it and then depositing another layer of Aluminium at a different angle. This process is quite finicky and the yield of working qubits you get out of it is (or at least was at the time I worked on this) rather low. Adapting techniques from conventional semiconductor manufacturing (were error rates are < 1^-9 per transistor) seems therefore very promising. So far it seems that there are still many hurdles to overcome for this, though the fabrication methods get more reliable every year.

The attempts to build "qubits" involve very complicated custom gadgets, with strange crystals or superconductors, or other weird things I don't know much about. And then refrigerated down to almost absolute zero. And it's not just about mass producing qubits, they need to talk to each other, ie. be entangled. And that is the really hard part about all of this.

Martinis has been working on a quantum supremacy demonstration chip for a few years already, and it's not unrealistic that they will make it work in the next 12-18 months. The term "supremacy" is a bit misleading though as it suggests that general purpose quantum computing might be around the corner, which it is not.

The "supremacy" algorithm is designed to demonstrate -for the first time- a quantum processor that runs an algorithm which cannot (or just barely can) be simulated on a classical computer (due to the number of qubits in the system). The circuit they're using for this contains around 45 qubits (which is enough to make a classical simulation very hard) and the result of the algorithm that runs on it can be verified using a conventional supercomputer (as otherwise it would be hard to prove that it actually did what it's meant to do). So all in all this is a very clever way of demonstrating the possible supremacy of a quantum computer with the currently available means of technology, but it's no general purpose quantum computer.

Quantum Supremacy sounds amazing, like a blockbuster movie! One imagines a disjunctive change, with humans either assuming the mantles of wise, beautiful gods or being dragged, screaming, from their place at the head of nature by blank eyed quantum computing powered robot deamons.

But actually my understanding is that the people who are driving the Quantum Computing research agenda mean something fundamental which is very distantly related to application. Quantum supremacy will be declared when :

"But what quantum supremacy means to me, is demonstrating a quantum speedup for some task as confidently as possible. Notice that I didn’t say a useful task! I like to say that for me, the #1 application of quantum computing—more than codebreaking, machine learning, or even quantum simulation—is just disproving the people who say quantum computing is impossible! So, quantum supremacy targets that application." [http://www.scottaaronson.com/blog/?p=2943]

Also, the article doesn't point out that there are many people working on this internationally, there are groups in the UK and Australia which are developing differentiated approaches which have a strong chance of underpinning the technology for QS - preserving QBITs for long periods and creating large numbers of them which can be usefully manipulated.

cmon now. This is a good attempt at pandering to the hackernews audience by being skeptical and cynical, but Quantum Supremacy is a big deal. From the same article:

If I have any policy advice, it’s this: recognize that a clear demonstration of quantum supremacy is at least as big a deal as (say) the discovery of the Higgs boson. After this scientific milestone is achieved, I predict that the whole discussion of commercial applications of quantum computing will shift to a new plane, much like the Manhattan Project shifted to a new plane after Fermi built his pile under the Chicago stadium in 1942.

I didn't want to be skeptical, much less cynical. I agree with the second bit of the article too but the operative word for me is if!

If a * clear* demonstration happens things will change, alot.

But this has not yet happened, it may never happen, that's science! Personally I bet into 5 or more years but would be delighted to be wrong.

I can't find any other source for this claim. It's written by Dr. Peter Diamandis. I don't know who he is, but his biography seems impressive. Could this be true?

Statement is about 2/3 down the article.

It's very unlikely. Quantum computing is just a bunch of random generators (very fast generators indeed) with bunch of filters on top of them. They good for searching a solution in broad space, but that is all. It's reinvention of analog computing with electric models, but using new tech.

"Searching for a solution in a broad space" describes virtually every hard problem in computer science.

For example optimization problems are everywhere. In planning or in machine learning, just to name a few.

Having efficient ways to solve such problems would very definitely make a difference.

I agree that it's unlikely to happen anytime soon, though.

The problem with analog computing was that they tend to oscillate or pickup noise which greatly limits their usefulness for precise calculations. With quantum computing, you will need to use series of runs and average them. For simple cases, it will be e.g. 10 runs, for complex cases you may need thousands or millions of runs. IMHO, it still will be faster (because quantum produces random values at mach faster rate than macrosystems), but (IMHO) efficient quantum algorithms can be reimplemented with electronic circuits, which will consume much less resources than general purpose quantum computer.

> efficient quantum algorithms can be reimplemented with electronic circuits,

You can emulate a quantum computer with a classical one, but it takes exponential time.

If you could emulate them efficiently, quantum computers would just be another Turing machine, with all the limitations we already know. I think this is the whole point of quantum computers, they are different from what we already have.

I talking about analog computing, not about digital computing.

https://en.wikipedia.org/wiki/Analog_computer

Quote: "Here’s the list of the underlying technologies:

(1) Electromechanical;

(2) Vacuum Tube;

(3) Relay;

(4) Transistors; and

(5) Integrated Circuits."

Surely not in that order? Because Relay = Electromechanical. The "Relay" item is certainly out of place. In the history of computing devices, relays (and purely mechanical computing devices like the Difference Engine) predated vacuum tubes, which predated transistors, which predated integrated circuits.

Perhaps electromechanical is in reference to something like the enigma machine. Characters light up on a board, but the character that lights up is chosen by a purely mechanical process that selects the current path. I'm just speculating.

> Perhaps electromechanical is in reference to something like the enigma machine.

Yes, I agree, good example, but that also predates vacuum tubes as computing/logical elements. So the temporal order is still wrong.

Maybe the author didn't really intend that the list be temporally ordered -- but it seems as though that was the intent.

I think by relay they are referring to reed switches: https://en.m.wikipedia.org/wiki/Reed_relay

Which were used for memory and other computing tasks after tubes were already obsolete.

This, VR and smart 3D television.

Vacuous comment pieces are great for generating and maintaining interest and pushing understanding and ideas forward. However, reality checks are essential to avoid this being little more than a sci-if short story, with realistic considerations.

Delving into why quantum computing may not be useful for implementing PHP 8 would be an interesting article. Encryption is the quantum computing poster child. What else will it give us?

Only being negative to help us find the positives and promote practical and realistic understanding of the subject.

For context, check out this 'Science International' video:

https://www.youtube.com/watch?v=fF8vLUEp29E

If you're of a certain age, you'll remember this series from when you are a kid, they talked about 'all the future things'. Except almost none of it came true.

The future is hard to predict.

Maybe they will make some progress in 12-18 months in Quantum, we we are a long way off from seeing anything working, I believe.

The Crytpo implications alone are pretty crazy.

By the time you get to 300 qubits, you’ve got a computer that can do more “calculations” than there are atoms in the universe.

I don't understand what this means. Last time I checked, the Universe was infinite.. So what does this claim exactly mean? Because it's not the first time I've seen it.

The author likely meant "observable universe". 2^300 ~= 2*10^90, and it's estimated that there are around 10^80 atoms in the observable universe[0].

[0] https://en.wikipedia.org/wiki/Observable_universe#Matter_con...

The universe is likely finite, with a countable number of matter constituents.

Where did you check that?

Judging by the sensationalist clickbait articles in the list on the left, I wouldn't hold my breath for anything that's said here.

Stop Quantum Supremacy Now!

In Trump's hands...!

Next article: "Making silicon computers great again!"

Nope.

Reply on news.ycombinator.com