Quantum computing as a field is obvious BS

Post Reply New Topic

I'm not telling anyone anything. Read the tone of my posts, I clearly advertise myself as a layman. Magz has graciously responded and I am learning through the framework of debate and discussion. You and Fnord have been... less gracious.

Nowhere it implies a need for physical elements for each quantum amplitude. Actually, tracking these amplitudes would ruin the quantum properties of the system. The very concept of a quantum computer is to have these 2^1,000 amplitudes evolving in a system of 1000 channels.

Yes, it is - in a sense - an analog computer (with continuous set of states). Digital computers "won" over analog computers because of easy data correction - say, electric potential of 0V means "0" and 5V means "1". So, you read a bit and your voltometer shows 4.3V - you can reasonably guess it should be 5V and refresh it as 5V.
You couldn't do a similar procedure in a computer where your voltage is exactly the number you're interested in - so, in a analog computer, cumulation of error is a much bigger challenge.
However, the amplifier in your HiFi is a kind of very specialized analog computer performing operations on a continuous signal - and it works very well in its very specialized field.

It's similar with quantum computers. They perform operations on quantum amplitudes that are - in a way - analog signals. The operations are performed on all the amplitudes at once, with one physical element for an operation on all the amplitudes (or many if you don't want all), so you don't need to expand the system exponentially to have exponential gain in computation capacity. That's the attractive property.

The real challenge is to isolate these amplitudes from environment noise so you don't lose entanglement - if they lose it, they start behaving like classical signals and you don't have this exponential gain, just one signal for one channel, like in a classical computer. It's a real engineering challenge to keep your signal in the desired quantum state for long enough.

Additionally, error correction is even more tricky than in classical analog computers because measurements fundamentally destroy quantum states - but there are some algorithms aimed at error correction, I remember being shown one on a lecture.

Most engineers - including really brilliant ones like my husband - are not really familiar with these regions of quantum mechanics, so it looks like they imagine quantum amplitudes as classical signals, requiring a separate channel for each. The very idea of quantum computer is that you don't - you can have 2^n signals in n channels but you need to keep them in fragile quantum states.

Well... as Randall Munroe illustrated in this chart:

quantum mechanics is particularily prone to being discussed by people with little understanding of the topic. It's really hard to answer some questions without at least a semester or two of specialized Math, aimed specifically at these problems.

TLDR: Quantum computer:

Pros:
Using quantum superposition and entanglement, you can have up to 2^n signals in n channels and perform operations on them simultanously. Classically, you have n signals in n channels.

Cons:
Quantum states are fragile, interactions with environment - including thermal fluctuations - degrade these special properties;
Error correction is really tricky.

there are still VIPs who think it is worth doing, and putting serious money into it trying to do it.

Yes I know. I'm still looking for where that idea came from. I've not found anything yet, but I did have a harebrained idea, which is probably wrong. In order to see the speed benefits of a quantum computer, would not the measurement & control circuitry need to be quantum in nature itself to keep up? If the control circuits are classical, you might end up with a bottleneck. I'm wondering if the exponential components idea comes from something like that. More complex quantum computation circuits themselves requiring quantum circuitry... requiring even more quantum circuitry and on and on...

I think the exponential components idea comes from an engineer expert in classical signal theory but knowing little about quantum information.

I spent a good few years studying crypto & graph theory as they relate to basically unlimited entropy. Turns out people hate the whole idea even though it is demonstrating considerable validity.

Magz is helpful here & OP needs a reality check. Discussion of a controversy does not invalidate the subject matter just because you aren't looking at the implementation of the tech.

This may be the origin of the "exponential hardware" idea, perhaps wrongly generalised from Shor's algorithm to quantum computing as a whole in the original article:

https://www.emergentchaos.com/archives/ ... gress.html

I was surprised to learn that Shor’s algorithm requires 72k³ quantum gates to be able to factor a number k bits long. Cubed is a somewhat high power. So I decided to look at a 4096-bit RSA key, which is the largest that most current software supports — the crypto experts all say that if you want something stronger, you should shift to elliptic curve, and the US government is pushing this, too, with their “Suite B” algorithms.

To factor a 4096-bit number, you need 72*4096³ or 4,947,802,324,992 quantum gates. Lets just round that up to an even 5 trillion. Five trillion is a big number. We’re only now getting to the point that we can put about that many normal bits on a disk drive. The first thing this tells me is that we aren’t going to wake up one day and find out that someone’s put that many q-gates on something you can buy from Fry’s from a white-box Taiwanese special.

A complication in my calculations is the relationship between quantum gates and quantum bits. For small numbers of qubits, you get about 200 qugates per qubit. But qubits are rum beasts. There are several major technologies that people are trying to tease qubits out of. There’s the adiabatic techlogies that D-Wave is trying. There are photon dots, and who knows how many semiconductor-based methods.

It isn’t clear that any of these have any legs. Read Scott Aaronson’s harumphing at D-Wave, more pointed yet sympathetic faint praise and these educated doubts on photonics. Interestingly, Aaronson says that adiabatic quantum computers like D-Wave need k¹¹ gates rather than k³ gates, which pretty much knocks them out of viability at all, if that’s so.

The arguments of the engineering being impossibly difficult are only part of it, indeed he states he cannot prove it. His main argument is that quantum computing has many, if not all the hallmarks of other now-forgotten hyped vaporware.

Well, there won't be any "something you can buy from Fry’s from a white-box Taiwanese special" cracking 4096-bit RSA keys any soon if that's the question asked

It's possible Shor's algorithm, while flagship for early development of the branch, won't be the one most commonly used, giving it to e.g. quantum annealing or other algorithms.

Multiple approaches to make the theoretical concepts a real machine - adiabatic technologies, photon dots, etc. - are normal in research. After all, research is all about exploring

Now, philosophically:
I think the "hyped vaporware" criticism misses exactly this exploring part of what research is. I mean, engineers building solid things using well-established technologies are one group of people we need - another group chases Higgs boson and oscillating neutrina, measures pulsars, etc., because, as R. Feynmann put it, "quarks are sexy". No need to complain the other group is different from yours, both contribute to growth of the humanity.

I've been reading quite a bit of his writing in the last day, he's particularly dismissive of high-energy physics. There and elsewhere he talks of the waste, not just money but the human capital too. Because "quarks are sexy" and not just to real scientists, we have too many people focussed on sexy science rather than, dare I use the term - useful science - the sort that can plausibly be applied by engineers and improve human lives.

Food for thought anyway.

There is a lot of "useful science" going on, too, it just doesn't need that much advertisement to get fundings - concerns pay for it and advertise only the end product, not the R&D.

The point of fundamental research is, you explore reality without any idea what would become useful 60 years from now. There is rocket technology in my bike. General relativity in my GPS. Pursuing knowledge for the sake of itself can bear fruit in unexpected places - and not bear fruit where expected. But you can't have real innovation without exploration.

This criticism is a bit like telling Roald Amundsen not to go to South Pole and measure cabbage fields west of London instead because this is less costly and more useful.

Fine for Roald, but do we really need hundreds of thousands of people going to the South Pole every year and as many more writing theoretical papers about building high tech spaceports there because cabbage fields are boring? I won't say he is definitely right but I think he does have a germ of a point somewhere.

We don't lack workforce in applied sciences People are developing medicines, do machine learning stuff, improve farming efficiency, improve weather forecasts, model earthquakes... there are money and people there.

We don't lack a workface...and, as Magz points out, we don't lack EFFICACY.

The first time I read about a quantum computer I really thought it was a joke - like one of those long deadpan jokes from the Journal of Unreproducable Results.

I later realized that it was not a joke.

People are perfectly free to ignore things like quantum or decide that they are BS.
People are also free to decide that global warming is BS.
Or that the earth is flat and the edges are guarded by NASA.

https://plato.stanford.edu/entries/truth

https://plato.stanford.edu/entries/postmodernism

Calling something "obvious BS" might not be the best way to invoke serious civil discussion of a complex topic.

You can also read about "Post Truth" - google it.

Dick Feynman (Nobel Prize winning physicist) was working on some lectures on computing.
The were published under the title "Feynman Lectures On Computation"
It is a bit of a hard read - it is alto hard to get a copy - I purchased 4 copies which were all canceled because the seller had already sold the same book on a different e-commerce website.
He crammed much of my 4 year degree in Computer Science into the first chapter - the rest was about the "physical limits of computing". I still haven't finished it.
Feynman died in 1988. The fellow who found his notes was also a very bright guy but he kept having to go to others in several different departments to make any sense of the notes - but the notes DID make sense. It took him a few years to get the work together in a form that could be published.

Some things are hard because people use jargon - but once you understand the jargon you can understand the topic. Some things are hard because they are hard.

Quantum is hard.
Because it is hard.

In a nutshell - below a particular level of "small" things stop being like billiard balls and start being like n-dimensional waves. So what is "obvious" simply stops being "obvious". Wave-particle duality is hard. Best I can figure it - everything really is n-dimensional waves - but sometimes that can look a lot like billiard balls. Sometimes it looks NOTHING like billiard balls. Turns out the universe doesn't much care about what is obvious.

When I was in grade school I learned about something called "action at a distance forces" like gravity and electro-static force.
Quantum says "no such thing - everything is messenger particles". And wave-particle duality says that the particles don't have to act like particles.

If you can figure out the Higgs-boson - and you are a better thinker than I if you can - then perhaps quantum computers might start to make sense.

Even Einstein didn't quite get quantum entanglement.