Quantum computing as a field is obvious bullshit | Hacker News Reader

Quantum computing as a field is obvious bullshit

scottlocklin.wordpress.com

33 pointssmsm4238 comments

Disclaimer: I'm near the end of a PhD in experimental quantum computing.

I'm not sure what he means very little progress... we're entangling over long distances, we're doing practical error correction, we're reaching really long coherence times. More importantly, we've seen exponential progress in coherence times over time. Moore's law-like progress. Quantum computing has also had the side benefit of generating obviously practical results like ultra wide bandwidth, high dynamic range quantum limited amplifiers, and debatably practical but stunning fundamental results like reversing a quantum jump in real time by feedback. This is all recent. Quantum computing research has exploded in the last 10 years. The blog post just sounds like some dude that doesn't really understand physics, didn't do a thorough literature search, didn't really even look at recent news articles, who pats himself on the back for being a real practical business guy instead of a Quantum Physics PhD, because some Quantum Physics PhD served him food or something and thus they must all be floundering. He then goes on to poorly explain why quantum computing is really really hard and therefore we shouldn't be spending money on it. Poorly I say because he can't even seem to explain properly one of MANY proposed implementations of a quantum computer. One that isn't even particularly popular anymore. I really think he just woke up angry and wrote this.

Nearly having finished a PhD in quantum computing, I wouldn't say the skills I'm most proud of I picked up are my understanding of quantum mechanics... I'm now an expert in low noise microwave electronics, high speed feedback control, cryogenics. Some of which are practically useful skills, some of which are highly applicable towards better understanding of the natural world in ways that I think the author would even agree are worthwhile. So no, my life hasn't been wasted and my career hasn't been ruined.

"my life hasn't been wasted and my career hasn't been ruined"

Talk to me in 10 years when you're not bringing me a steak for a living.

FWIIW I went to a Gordon conference on this in the 90s, and literally, the capabilities of quantum computers have not changed since then. Zero. None. There has been no "Moore's law like progress." There hasn't been any! That's the point! There's been more progress in interstellar space flight than quantum computing (I dunno, White-Juday is at least a good try), but at least warp drive creators aren't trying to convince people like you to ruin your lives.

https://imgur.com/a/HCJ7hLq

Again, exponential progress in coherence time.

Well, for one thing, consultants who can see through baloney like that are extremely valuable to people in business.

You ever look at Haroche's 1996 argument against QC? You should go have a look.

Read it. Pretty stale take, which it would be for being 23 years old.

Have any of his concerns been addressed? Do you, for example, have a pithy rejoinder to the idea that manipulating an exponential number of states to any degree of precision is kind of impossible?

>Talk to me in 10 years when you're not bringing me a steak for a living.

This comes off as an incredibly hostile sentence, though I suspect you may have not meant it as just an attack.

Dude, some guy with a Ph.D. in quantum computing from MIT brought me a steak for a $10 tip; does that not concern you?

Dude, some guy with a Ph.D. in quantum computing from MIT brought me a steak for a $10 tip; does that not concern you?

Dude, your n=1 anecdote isn’t as amusing or compelling as you seem to think. Your entire argument seems more like a thinly veiled excuse to rant than it resembles an intellectual exercise. That’s hardly an inspiring argument from a self-styled empiricist.

https://news.ycombinator.com/item?id=18961731

I thought it was hilarious, personally, as we were discussing at the dinner table why quantum computing is probably nonsense. My interlocutor (who was more or less going along with the conventional opinion that QC is about to "change everything") thought I bribed the waiter to say this.

If you only sleep 4 hours you can have two jobs and still work on quantum computing. Then when you make a break-though someone steals your idea. Science is only for the privileged.

He then goes on to poorly explain why quantum computing is really really hard and therefore we shouldn't be spending money on it

I think it read more like

We're still ages away from the ancilliary technologies that will allow us to tackle quantum computing

The whole XVIII century vs SR-71 example was pretty clear

But what technologies are needed to make quantum computing work with say superconducting qubits? High speed room temperature processors? FPGAs and fast DACs/ADCs will do it. Low loss cables? Got that. Low noise amplifiers? Got it. Idk what people want. It’s a good time to start throwing money at it to iron out the problems that nobody wants to do a thesis on. It’s not riskfree but it’s a reasonable bet.

Indeed, you gotta risk to win. But at which threshold you stop, and reconsider, given the lack of results? 1B? 10B? 1T?

I'm not an expert on the matter, but I understand "cut your losses".

What “lack of results?”

I very clearly pointed out that no QC researcher has created, for example:

1) An actually error corrected qubit

2) A quantum factorization of the number 15

You can set yourself artificial goalposts of "hey my entanglement lasted a little longer on architecture X this year, therefore Moore's law means it is inevitable" -but you still can't factor the number 15. Please factor the number 15. No "pre compilation" where you leave out the number of gates needed to get the answer -just factor the number. Show me this, and then I will proceed to make fun of the idea that you can factor, say, a 3 digit decimal number. This is the only demonstrated interesting use case for QC, and the only reason anyone is funding the topic.

Yes, yes, "modeling quantum systems" -OK, go model an interesting quantum system for me.

1) https://www.nature.com/articles/nature18949

https://arxiv.org/abs/1411.7403

2. http://science.sciencemag.org/content/351/6277/1068 seems like it might satisfy your requirements. It doesn't seem to use the answer in its calculation, but I'm not sure.

If the abstracts were actually true and someone implemented an actually error corrected qubit with lifetime 320ms, I guess it should be trivial to apply a few thousand operations to a mere 4 qubits and factor the number 15.

You're at Yale and presumably going to Berkeley later: maybe you should try to do this. Sounds to me like you could build a whole career on meeting the challenges involved in factoring the number 15.

Edit add: since you continue to add material, I will continue to add comments; half the text of the Blatt paper is verbiage making excuses for the fact they dropped the operations they knew were not needed to factor 15, just like the NMR quantum computers in the 90s did. In fact, this one is worse than 90s results.

The amount of insults per line compared to the number of actual points per line convinced me that this probably didn't make any noteworthy points.

"Quantum computing hasn't been practically useful yet!" Yes, we know that; thank you for your valuable insight.

You'd think that it would actually comment on the claims made of late, about BosonSampling and whatnot, but as far as I could tell, nope!

Agreed. Seems like the author had enough education on the subject to be able to form an opinion that sounds credible enough, but lacked the capability to hammer the point home.

Sure, many of the points made are valid - work is still ongoing. Progress however, has been staggering. Ignore the whole quantum computer aspect, and you still have condenser matter, nanomanufacturing, optics, material science, modelling, dilution refrigeration, metrology and many other fields that directly benefit from the funding that's going into quantum computing at the moment.

So like, if I claim I can build a perpetual motion machine in a dilution refrigerator, everything is great because "dilution refrigerator."

FWIIW I sat next to a dilution refrigerator in a lab in 1991 -can you name any significant improvements in them which came from QC funding?

I don't think the combative stance helps your argument at all. You know I didn't mean that.

In answer to your question: certainly.

First of all: the sample volume, total number of microwave lines and resulting cooling power have all been massively increased in standard dilution fridges so that the modern quantum computing experiments can fit in. Even if you don't care about phase qubits, but are interested in something else that needs to run at mK temperatures, you're going to have a better time now rather than 30 years back.

More importantly though I'd say isolation has drastically improved. Even at 4mK you can get infra red photons breaking apart Cooper-pairs: forming additional quasiparticles. That noise interferes with any qubit in the fridge. These infra red photons were originating from the fridge itself before this issue was fixed. We would never have identified this issue without studying the high sensitivity of qubits to quasiparticle creation.

Finally, if you don't think that these advances are significant enough, then I don't think you're just having a dig at quantum computing, you've got an issue with the progress of science as a whole.

"Combative stance?" Do you think I should be more collaborative in pointing out that quantum computing is a bullshit field? This sort of mealy mouthed attitude is how such nonsense has grown and thrived in the first place. This subject should have been laughed to scorn in 96.

You're right: I have an issue with the progress of science as a whole, though I particularly take issue with "quantum computing" being considered science or technology or much of anything beyond play acting with liquid nitrogen. On the upside, nothing I learned in school is likely to be out of date ... for the rest of my life!

For the sake of lolz, please characterize this massive increase in cooling power using ratios. I could have missed something.

In the spirit of charity I wish there were a version of this post that didn’t seem like it was written after injecting some strange cocktail of anabolic steroids and cocaine.

Maybe there are some good or at least interestingly wrong arguments but I find myself incapable of looking past the rhetoric.

Right, much like discussing political maters, when discussing important things; compassion and easily digestible chunks of information go farther. It's OK to be passionate and certainly OK to call things out, but for the most part I think the only people who are able to look past the rhetoric are people who already agree with its position.

For people who think this post is too much of a rant (not me; I like it), perhaps this article might provide a more interesting read: https://spectrum.ieee.org/tech-talk/computing/hardware/the-u... which summary of the finding of the national academy of sciences report which can be downloaded from here https://www.nap.edu/download/25196

Personally, I'm all for more progress on quantum information science topics, but I think it will be better for the field if we rein in some of the claims about the potential for the technology, and the timeframe when it will become practical.

One glaring fact in support of the original post is that even the companies who make quantum computers don't claim that they are useful for solving any real problems. The problems that they can "solve" have to be constructed to fit within the limits of the machine.

In other words, quantum computers are less powerful than Turing machines which can solve any problem that any computer can ever solve -- including one's we haven't seen yet.

That certainly doesn't seem to be true to the average layman:

http://www.dbta.com/Editorial/News-Flashes/IBM-Unveils-Quant...

https://www.forbes.com/sites/tiriasresearch/2019/01/17/ibm-l...

https://www.openpr.com/news/1516739/Quantum-Computing-Market...

https://www.forbes.com/sites/gemmamilne/2019/01/14/the-inves...

http://techgenix.com/quantum-computing-cloud/

https://techcrunch.com/2019/01/08/daily-crunch-the-age-of-qu...

http://discovermagazine.com/2019/jan/quantum-supremacy

https://jaxenter.com/quantum-computing-cryptography-security...

https://gizmodo.com/us-passes-bill-to-inject-1-2-billion-int...

Yes, they're very good at getting PR... but read what they put on their actual sites. These computers are expensive toys for research purposes.

The goal is to find out if quantum computers can ever be useful for real world problems. I'm with the OP; they can't.

Where do you think the PR comes from? Most of them put these links on their sites; for example, Rigetti: https://www.rigetti.com/news

Doesn't matter: everyone, literally everyone, is acting like this is just around the corner and RSA and friends are near doom. There needs to be more skepticism.

PR comes from PR companies. A great book on this topic is "Toxic Sludge Is Good For You," which I highly recommend. PR is designed for mass consumption.

If you poke around in the documentation on their websites which is designed for developers, you'll see that these computers cannot be used for general problem solving, which any computer (a Raspberry Pi, for example) can.

If you go further and actually sign up for a trial account on these services (which I have done), you'll find that the "problems" they can solve are quite contrived and, as the OP points out, only consist of issues where the programmer already knows the solution.

Quantum computers are not as useful as a Raspberry Pi. That's it in a nutshell.

Assuming everything he says is 100% true. Why can’t some people just work on things that interest them because they interest them? Why does everything need to be “useful” (as in useful right here right now)? Okay resource is finite and has to be allocated. But so what? What’s the point of science and technology if not to enable others to have a more meaningful life? For some, that meaning is in exploring “useless” problems. Let the people who want to make better classical computers do so. Let people who want to study quantum computers do so. I can assure you that the number of people who study/work on quantum computers is so low that convincing all of them to drop their work and do something more “practical” will not make the world a better place.

Everything I said is 100% true. Citation: nobody has found an error yet.

I think it would be great if people took this up as a hobby. I don't even mind wasting VC on stuff like this, and said so. I object to researchers selling this to young scientists as a potential career. It's not! It is a glass bead game, and acting like it isn't is the moral equivalent of fraud.

I normally go through HN comments before I decide to read the article. And so I decided to read it indeed to see how it could possibly be this bad. Wow! Just imagine how much different the article would have been if the author had just posted his misunderstandings as "hey I tried googling but did not understand these points, can an expert in the field help me understand the following...?". A world to live in indeed!

In everything I've read about quantum computing, I've always had nagging doubts about the actual engineering details required to make a QC a viable means of solving computing limitations. The original article sums all of these doubts up, and expounds on them. Can anyone who has actually worked on QC's offer an experiential rebuttle to the author's claims?

Addendum: I think most reasonable people reading these comments can agree that the author is hostile to the topic, bordering on being an asshole, and he probably knows this. However, this has NO impact on his assertions. They are good, especially the references to older arguments against the potential of QC's, which still seem quite valid.

"He also noticed that instead of making progress down fruitful lanes or improving detailed knowledge of important areas, most develop enthusiasms for the latest non-experimental wank fest; complexity theory, network theory, noodle theory."

What exactly is the issue with complexity and network theory?

A finance guy calling out physicists for wasting resources and doing useless work?

Whether quantum computing research will deliver anything useful soon or not (it won't), I think the picture at the end sums up the author's attitude nicely.

Reply on news.ycombinator.com