Fusion and Quantum computing

Factual Questions
41

Really? Because if you just google “potential of nuclear fusion” among the first hits are an article in scientific american summarizing 19 experts of the national academies of science urging the US to invest more in fusion because compared to fission, nuclear fusion creates almost no nuclear waste.
This interview with John Holdren, former head of the Federal Office of Science and Technology Policy, and a Ph.D in plasma physics describing it as “essentially inexhaustible energy”. (Interestingly he also seems positive about the potential for space travel).
And the IAEA suggesting that the amount and degree of harmful byproducts are vastly smaller than fission.

What I said is very clearly a fine one or two line summary of what scientists actually say.
And if this were a reasonable discussion, you’d grant that.
However, you came into this thread guns blazing, and I’ve no doubt you’ll have some snappy retort, probably involving linking someone being pessimistic about whether we can build nuclear fusion reactors (which would be irrelevant to whether my summary of what scientists say about the potential of nuclear fusion in the abstract is accurate or not).

I’m just replying because I wanted to emphasize the following point again, as I think this is a point that even people arguing in good faith sometimes miss:

Yes, obviously there are engineering limits that will constrain nuclear fusion beyond just the constraints of physical laws. We may never achieve the full potential of nuclear power, and I alluded to this in my first post.

However, we don’t know what the engineering limits are right now. We don’t. In the hypothetical case of “What if we had commercial nuclear fusion power plants?”, some or all of the known issues with reactors as we know them have necessarily been solved. Otherwise it would not be usable for this purpose.

You can argue we’ll never achieve that state, because it’s impossible in your opinion. That’s fine.
But if we’re imagining that state of affairs then necessarily many if not all of the current engineering roadblocks have been solved. The commercial reactor cannot be the same as experimental reactors today.

42

Typical computer architecture nowadays isn’t designed to handle thousand-bit integers, so you have to pull some programming tricks to deal with number that big. But those tricks exist and are well-known, and the physical computer itself has billions of bits of memory available, so it’s no problem. You might need to trick the computer into thinking that it’s something other than a number, but whatever you call it, the number fits into the computer, by a very easy margin.

When your entire computer only contains 54 bits of memory, though (or at least, only 54 bits of the right kind of memory), that’s a much more difficult constraint to work around. Maybe there’s some way to do the calculation on a hybrid computer, with oodles of classical bits and 54 qbits, that only needs 54 qbits at once, and which can solve problems slower than a fully quantum computer but faster than a fully classical computer. Maybe. But if so, nobody’s found it yet.

43

Yes, there’s a bunch of fusion hype. The OP asked for NO HYPE so that’s what I’m giving him. Googling for overoptimistic and vague statements that talk about theoretical possibilities that may never materialize in glowing terms is just going to find more hype, not solid answers.

I’ve already granted that it’s a summary of hype that scientists often put out, if you were engaged in reasonable discussion you’d stop pretending I didn’t. But the OP asked for a realistic, no hype discussion, not a bunch of ‘limitless free energy’ claims, and repeating hype that ignores the many real-world difficulties involved in a working fusion reactor doesn’t do that.

I’m not sure why you think that saying over and over again that your position is that ‘if we don’t know exactly how this will work, we should assume that all of the issues we see with attempts now will completely resolve themselves’ will make me think it’s reasonable to do so. Ignoring the real-world problems and relying on hype is exactly what the OP asked to avoid.

I could argue nearly any position, but I didn’t argue that position in this thread, and it’s not my opinion. The ‘never’ and ‘impossible’ are entirely your additions, and for all of your talk about a discussion in ‘good faith’ adding those kind of qualifiers to what someone said is not discussing things in good faith. It’s really absurd that you keep pretending that I’m not being reasonable when I’m referencing the issues encountered by people attempting real-world solutions to the problem, and you’re just repeating hype while shouting over and over that I’m being unreasonable - and this bit is just the icing on your ‘unreasonable, but claiming to be reasonable’ cake.

44

Quantum Annealers are definitely analog computers. A quantum annealing system is essentially a matrix of artificial spin systems, entangled. The inputs are the spins and the weighting of their coupling in order to create a higher energy ground state of the spin system. Once the ground state is set up, the whole system is allowed to “anneal”, relaxing to the lowest energy ground state. If you set the “problem” up right in the higher energy state, the lowest energy state will be the solution (great for optimization problems).

However, most of the “quantum supremacy” systems you read about are “gate model” computers. There the inputs to the system are to set up the superpositions within the qubits and the entanglement between qubits within each “gate” and from gate to gate and then the “gates” operate and influence one another to run whatever algorithm you have set up. This computer is much more like the CMOS processor, with digital values set as input and digital values out (though it is important to realize that all computers operate using analog electronics if you start to look at the individual transistors or qubits and there interactions).

45

Thanks, that’s what I was looking for.

So, does the output exist in a conventional tabular format or is it closer to a data cloud that is mined statistically?

46

It seems that the Quantum Computer is an analog computer.

In the sense that any computer used for simulation is an analog computer. I didn’t want to say that a quantum computer “is” an analog computer – just that solving the simultaneous conditions works in a fundamentally different way.

It appears to me that quantum computing is currently at the stage equivalent to when they were toggling the bits into a digital computer from the front panel. Setting up the initial conditions for a digital computer used to be pretty difficult too.

47

I’m very familiar with that part. The first digital computer I worked on was the IBM 704 at JPL (IBM Customers Engineer 1958).

48

Readout of both annealers and gate-model computers is accomplished by making a measurement that puts all the qubits into definite states. Generally, the states are the artificial spins version of “up” or “down”, which can be interpreted as a “1” or a “0”, so in that sense it is like reading a digital register. Until the measurement is made, the system can be viewed as a sort of “cloud” with all qubits in unknown, mixed states.

However, because all quantum is inherently probabilistic, both the annealers and the gates need to be set up, run, and read out multiple times in order to assure you have the right answer (which will be the peak of the distribution of answers).

Currently, all quantum computers are NISQ (Noisy Intermediate-Scale Quantum) computers, so even with error correction the algorithms are run many time and the resulting distribution’s peak is provided as “the answer”. But even if you could run a completely noise free quantum computer, there would be a (very small) chance that any single run would contain “wrong” bits, so there would still be the need to run it multiple times.

49

When it’s plasma physicists, atomic energy agencies etc saying this, it’s not hype any more. It’s humanity’s best understanding of this technology. And the right thing to summarize to someone asking about this technology.

A lot of your arguments seem to rely upon paraphrasing me into a straw man.
This is now the 4th time: It’s not about assuming the problems assume themselves. It’s about that in the hypothetical case where we have a commercial nuclear fusion reactor, it’s necessarily not the same as experimental reactors today. It would have to have solved essentially all of the containment, stability and other engineering issues.

There are two positions available: we can suggest that such a reactor is impossible. Or we can say it’s possible and it will be different from reactors today.

The one position we cannot take is suggesting that even if it’s possible such a reactor will function the same as experimental reactors today.

I never suggested you made such a statement, so calm down and try to read what I am saying before flying off like this.
I’m saying that the position that a commercial scale nuclear fusion reactor is impossible to make is at least a self-consistent position.

OTOH, suggesting that even if we make a reactor that can do many things that experimental reactors today cannot, it will still have all the functional limitations of reactors today, is a self-inconsistent position.

50

Sounds like something that would be better suited to weather modeling than large primes.

51

It doesn’t matter who’s spreading hype, hype from scientists is still just hype. “Humanity’s best understanding of this technology” is things like peer reviewed papers and experimental results, not misleading summaries made to get headlines and/or funding.

Fallacy of the excluded middle. It’s pretty clear at this point that you’re just repeating hype, baselessly denying the flaw in your argument over and over (four times by your count), engaging in logical fallacies, and accusing me of not arguing in good faith for refusing to accept incorrect assertions from you. I think I’ve answered the OP’s question, and that it’s clear who’s operating in good faith here, so have fun.

52

Nuclear Fusion already has, in that weapon yields on nuclear weapons can go beyond high thousands of Kilotons. But fusion being around the corner has been a persistent problem, and I suppose that solving it leads to questions of how that works.

My suspicion is that Nuclear Fusion isn’t going to work on small or even modern scale of power plants; Fusion power is going to require gigantic means to profitably convert hydrogen to helium. As our understanding of how to manage extreme temperatures increases, the output will rise, but building such a power plant is going to be a monumental effort even when it does become possible.

The glory of this all is that humanity can gain access to vast amounts of power. It’ll take longer, probably another twenty years, for second generation fusion plants to win on price, but that transition means a general conversion of power. Fusion Gen II is going to be going up against the grandchildren of modern renewable and the battery backups that give them more flexibility, and the dying tail of fossil fuels [At this point, largely natural gas]. The politics of this is going to be analogous to today’s scenario - Oil producing countries facing the end of high demand. There is a future in plastics, but the clock is ticking on fuel.

The key players, in this case, are nations exporting Lithium, critical for advanced battery design. Like oil, this is not Lithium becoming useless, but a collapse in demand. Several South American nations face a difficult transition out of resource boom economies. The world marches on, and there are winners and losers.

Fusion keeps on getting cheaper, and its around this point where R&D questions of ‘what can we do with more power’ start to get asked. There are many implications: Faster travel, gemstones and other materials becoming common, decreases in cost of living for the people of the world. If it hasn’t already happened, this pushes mankind to get into space, as it is probably possible for a third generation fusion design to survive going into space, and burning Hydrogen in space as fuel. There are formidable engineering problems here, but conceptually this design could go all the way to Alpha Centauri.

Quantum Computing doesn’t take 50 or 100 years to hit its stride. It’s advancing at a far faster pace, and we’re going to see implications from it in perhaps 5 years.

Conceptually, Quantum Computers have a marvelous answer to brute force style problems: ‘All of the Above’ and see what works at all, then focus on what works. This defeats our modern style of password security and may even defeat things like Chess. There are some embarassing outcomes, but people simply switch their security apparati. Fortunately, things like a Skynet scenario are not possible, as the US Nuclear Arsenal remains safely locked with 1960s era punch cards.

Computational power will rapidly expand, far beyond modern computers. There are questions about how programs would work with Quantum computing, including that the major problem with the design is a need for redundancy. Any new technology has teething problems, although there’s no reason that this provides any kind of hard limit. My suspicion is that quantum computers, with much fanfare, outpower the human brain in perhaps 20 years.

This can go badly in a lot of different ways, although I think the most likely answer is probably transhumanist: Given this power, it seems prudent that enhancing our ability to think would be the answer. Rejecting it with a Luddite style answer would only make its rise uncontrolled; a doomsday War of the Machines style answer ignores the part where the creators of these machines are likely businesses and research tanks, not sociopaths. There is probably some kind of instructive lesson that speaks to the cultural importance to responding in this way.

Plugging something able to process information on the order of sextillions of times faster than ourselves trivializes many routine tasks. This is probably the end of the human condition as we know it–there will be holdouts, but it doesn’t take a lot of imagination that a third step, going far beyond human intelligence, is coming. Futurists have talked about the singularity, a point in which technology simply explodes because a feedback loop that allows technology to advance triggers seismic shifts in technology. If this sort of thing has any merit, suddenly gaining vast increases to processing power and being able to think is going to trigger it. Differential equations and random trivia would be as intrinsic as seeing a red object and understanding that it’s red. In such a setting, what would we think about, what would we do?

And maybe we run into hard limits, of which we are yet to find. Moore’s law is drawing to a close; we are designing chips that are a few atoms thick and we’ve responded to the limitation by adding more cores to our processors. From Babbages difference engine or a room sized ENIAC, computers have heavily advanced even if we are now facing some of their hard limitations. That the computations made by Quantum computers would scale into the sextillions [remember, there are 50 sextillion atoms in 1 gram of carbon. The Andromeda Galaxy is about 350 sextillion feet from you right now], and so the math really does suggest some fairly insane consequences, not least of which because a Qubit doesn’t have a 1 or 0, it can have many, MANY values in between.

But I’m not suggesting that someone makes a quantum computer that spins that whole gram of carbon and makes it do a gigantic calculation of (500^10^21–which is such a big number that actually describing it using words leads to spamming the word Milli over and over again). It takes just 8 or 9 to compute as deeply as the human brain. And getting to 8 or 9 isn’t going to prove much harder than getting to 10, or 11, and so on. The potential really does seem insane, and plausible.

Tl;dr
Fusion Power takes a long time to start moving the needle, only gets online after renewable have dominated fossil fuels, but make major changes.
Quantum Computing moves faster and probably changes the human condition.

53

You might not need to re-run your quantum calculation if the problem is of a sort that has easily-verified solutions. Given a very large number, it’s very difficult (using a classical computer) to find a factor, but it’s very easy, given a potential factor, to double-check whether it works. So if the problem you’re solving on your quantum computer is factoring, then you run your quantum program once, get a number out, and test it. If it fails the test, then you have to re-run it, but if it passes, you’re done.

And I know that quantum computing is often described in the popular press as “try everything at once and keep only what works”, but that’s not actually how it works. There is a certain class of problems for which quantum computing works (or rather, for which it works better than a classical computer), and chess (for example) is known to be outside of that class.

54

As @Chronos notes and I pointed out above, this really doesn’t have anything to do with how a quantum computer works. In fact, there’s a lot of interesting number theory that goes into solving the factoring problem on a quantum computer, but the ‘quantum’ part of the algorithm works by encoding a certain sequence into a quantum state, and then using a quantum algorithm to figure out a global property of that quantum state—corresponding to the period with which the sequence repeats. This can be done efficiently, because one of the problems quantum computers can solve efficiently is Fourier transformation. This makes use of interference effects: components of the state corresponding to ‘false’ periods for the sequence cancel out, while those corresponding to the ‘right’ period reinforce.

The quantum Fourier transform is at the heart of many known quantum algorithms. But it’s not something that enables you to solve any and every problem faster; indeed, there are problems for which it’s known that they stump quantum computers just as much as classical computers (assuming widely-held beliefs about the complexity of these problems, and how they relate).

So the advent of the quantum computer will not mean magical computing power for each and every opportunity. There’s a narrow class of problems that quantum computers can perform more efficiently, some of which—most notably, factoring, but also searching through unstructured data—are highly important, but not necessarily for everyday applications (in principle, the factoring problem used to safeguard cryptography in present-day applications, like RSA, can be easily replaced by ‘quantum-safe’ encryption).

One possible exception here is the application of quantum mechanics to machine learning. Certain machine learning algorithms can be improved upon by quantum computers, which in turn leads to speedup for very widely-applicable problems, such as data classification. But that field is still very much in its infancy.

55

But who’s to say it’s hype?
Frankly, I am going to go with the judgement of scientists working in the field over some guy on the internet.

Speculating about what is possible for a technology going forward necessarily involves things that haven’t been unequivocally demonstrated yet. If it had been demonstrated, then a commercial fusion plant would already be running, or would be in the process of being built.
In terms of theoretical physics, no doubt many such papers have been written, I am not well-qualified to fully understand them, and have to go with the consensus judgement of experts in the field.

Also, you seem cool with speculation in other fields e.g. economically extracting useful amounts of uranium from seawater. The problem only seems to be with anything that contradicts the meme “Fusion power is overhyped”.

I’ll stop you there. What excluded middle? What’s the position I’m missing?

Because I would like to be clear on what you’re saying. In your first couple of posts you seemed to suggest that fusion reactors will always have the limitations of current day experimental reactors. When I’ve pointed out the issue with that logic, you’re saying I’ve excluded the middle. What middle?

As for you leaving the thread, OK, that’s your choice.
But better would be to concede the point and let the discussion continue. I know that my writing style on the dope can sometimes be a little acerbic and put people on the defensive. But you can also find examples of me conceding points, or apologizing to people. I am here to learn and grow, I’m not here to “win”.

56

It’s worse, because even among truly qualified experts, there are strong differences of opinion.

I’m old enough to remember some of the same debates over fission power. Some “real atomic scientists” said it would be great, and other “real atomic scientists” said it would be awful. Almost none of 'em said it would be “a little good and a little bad,” as it actually turned out.

57

Well it depends to what we’re referring.

In my opening post, I just talked about the hypothetical potential for fusion power. What’s so special about it? Why do people get so excited about potential breakthroughs?
I don’t think there’s much disagreement about this.

If we’re speaking more practically; Will we achieve such results in the next 50 years? Is it sensible to massively increase funding for fusion research? Now that’s much more open to debate.

The big back and forth that you saw in this thread between me and Pantastic I think really just started from a misunderstanding, that (s)he thought I was talking about the latter, and not the former.

58

Beyond fusion’s hyperbolic enthusiasm lies the practical questions… How do you turn energetic neutrons into heat, and how do you do this without creating radioactive waste. I’m not up on my nuclear reaction details, but the most practical fusion creates energetic neutrons. (Lack of neutrons was one of the clues that “cold fusion” wasn’t). You stop neutrons by having them collide with assorted materials - some stop neutrons sooner and better. But those collisions cause the shield materials to become radioactive (transmutation) plus also will cause structural degradation.

The key engineering question is how to deal with this problem, assuming that fusion can be successfully sustained.

59

Better would be for you to stop pushing for me to concede your point when your point is incorrect, but it’s clear that you won’t do that. And I’m not ‘leaving the thread’, I’m simply done trying to engage in discussion with someone who is not only failing to engage in a discussion in good faith but keeps accusing me of engaging in a bad faith discussion for not accepting blatant hype as facts.

60

Even worse, there’s the major political issue that if you use materials like ordinary uranium to stop the neutrons, you end up with bomb-grade plutonium. None of the hype about fusion even touches on how to deal with the fact that if you make a nice, economic fusion reactor, you’ve also made a plant for churning out fissile material. The risk of nuclear proliferation is the main reason fission breeder reactors fell out of favor, and none of the glowing praise of how clean and safe fusion is seem to even touch on this.