If you've been dying to run some math on a dinky toy quantum computer, IBM may have something for you

IBM today claimed it will shortly sell the world's first commercial quantum computer – or, more accurately, calculation time on it. The IBM Q machine has been years in the making, and is an extremely specialized bit of equipment, requiring the environment around it to be very tightly controlled to work properly. Pretty much …

[Anonymous Coward]

How did they predict the development thing?

Is this like the early 1960s, when fusion power was just 20 years away*

Or like the bold prediction by [begins with g] that by 2015 as many digital as analog cameras would be sold?

I suspect the former.

*Seriously we were being bombarded as kids with propaganda about going into nuclear science and engineering - guaranteed big bucks. But our physics teacher, who had actually been in R&D and got disillusioned, was having nothing of it.

[Anonymous Coward]

Re: How did they predict the development thing?

I agree with your assessment. FTA:

"The qubits that make it so powerful lose their properties within 100 microseconds – and that time falls off significantly with any introduction of vibrations or temperature change."

The more important fact is that each additional qubit increases the instability exponentially, and there seems no way around this due to the quantum nature of reality. This is why IBM's quantum computer has only 20 qubits; They've approached a hard, built-in limit, not unlike the speed of light.

I see it as a good thing. It preserves the utility of public key encryption, and who knows what else? I for one would not relish going back to the days of physical couriers transporting crypto keys all over the planet, with briefcases handcuffed to their wrists. Online retailing is just too convenient!

[Noah.Zark]

Re: How did they predict the development thing?

Not so sure about that hard limit. Have a gander at these advances on error correcting quantum codes. Interesting intersection between computer science and the nature of quantum reality.

https://www.quantamagazine.org/how-space-and-time-could-be-a-quantum-error-correcting-code-20190103/

[Anonymous Coward]

Re: How did they predict the development thing?

"Or like the bold prediction by [begins with g] that by 2015 as many digital as analog cameras would be sold?"

If you count smartphones' digital cameras, I'm sure the digitals drastically outnumber the analogs at this point. And most folks don't use a standalone camera of any kind these days - just their phone.

[comradequinn]

Re: How did they predict the development thing?

Face palm

[Francis Boyle]

In the 60s

fusion power was 50 years away. OK, we missed that deadline but by nowhere near as much as common wisdom has it.

[jelabarre59]

Riiiiiiiiight....

...has just 20 qubits

"Riiiiiiiiiiiighj.... What's a Qbit?"

(oh, sorry, probably shouldn't be paraphrasing Bill Cosby now).

[Justin Case]

That video explanation...

...reminds me of the Monty Python "How to" sketch.

Illuminated and enlightened I am not.

[Andy Denton]

It's maths not math.

[Anonymous Coward]

Let's see your proofs.

[Milton]

It's not a 900lb gorilla ...

It's not a 900lb gorilla ... but I think it may be a Schrödinger feline. And it deniably exists at Ft Meade. NSA has always had beyond-cutting-edge stuff in its basement, both in terms of knowledge (still routinely hoovering up some of the smartest math talent on Earth) and hardware. Extremely motivated and gushingly funded, NSA can buy and, more importantly, build whatever it likes. It can buy, or by nature of its activities, steal whatever IP it wants. Furthermore, as a kind of icing on the cake, it can keep its best toys secret, on pain of dark, deep federal sanctions.

My point being that if you really want to know the most advanced status and capability of quantum computing right now, you'd have to ask NSA. And they sure as shinola will not tell you.

It is frustrating, I guess, to speculate about advances in stuff like QC, or even computing, comms and crypto generally, and know that out here we are probably at least five, more likely 10 years behind the best that has been achieved. But this has been true with respect to NSA since at least the 1960s and I don't expect it to change any time soon.

Does anyone who knows the field care to make an educated guess about What Lies Beneath ...?

[Rich 11]

Re: It's not a 900lb gorilla ...

What Lies Beneath ...

...was a naff film. I watched it once and gave the DVD to a charity shop.

[MJI]

Re: It's not a 900lb gorilla ...

Better than GCHQ?

[John Smith 19]

Money is not the issue.

No one knows how to build a quantum computer that's actually a computer in the sense people hear understand the term, IE whose functionality can be changed by changing its programming.

I'm starting to think finite automata may be the way forward. MIT were leading this in the early 00's with their CAM 8 machine, then someone trashed all the data after 1998 (which seems rather strange to put it mildly).

FSA can make very small tiles on an ASIC and have very high density. Their function is governed by a look up table (LUT) much like a single cell in an FPGA but with the proviso it's an SIMD. Like a RISC, it's all down to how good a compiler can compile the task into the LUT, and how often it has to trigger a LUT reload.

[#define INFINITY -1]

Re: Money is not the issue.

FA is a subset of UTM. Don't see how it could speed things up.

[Anonymous Coward]

Re: Money is not the issue.

I love the wording of that last paragraph... Tech poetry