Funny how nobody talks about the Yeti being the first to summit everest and even more incredibly how it was the first to reach the summit WITHOUT the use of oxygen. I for one would like to pay homage to this creature by not shaving my beard or hair for a period of 10 years.
Furthermore I would like to announce my dismay that not only is the Yeti not discussed in climbing records, but also the lack of research of Yeti or Yetifolk collecting summit rocks. More funding needs to be devoted to this topic.
After doing some brief research I've found that some Yetis can be as large as 10 feet, meaning as many sherpas have speculated the second step wouldn't have been an empassable obstacle, some academics however remained unconvinced.
@@101noz101 And how much of that funding do you intend go to Yetifolk themselves! We cannot exploit Yetifolk for science, the way colonial science has been done in the past! 🤣
First of all when Richard Feynman says that "nobody knows how quantum physics works" he doesn't mean we can't calculate it. We can calculate it exactly. Everything down to 10 decimal places, aside from stuff we don't know yet like dark matter. So we can exactly calculate how quantum computers work. What Feynman meant is that nobody knows why the universe works the way that it works. It just does. Secondly due to superposition some mathematical problems can be solved extremely easily with quantum computing instead of classical computing and this does not violate entropy or information theory. The greatest drawback to quantum computing is translating quantum results into the classical world. This is where you typically add a bunch of entropy and heat to the process. It's extracting a classical result from quantum that takes time and energy. There is a problem in chemistry called the "protein folding" problem which requires so much classical resources that it is virtually insolvable except for certain simple cases such as 2 dimensional lattices. However, the proteins to in fact fold. How do they do it? Using quantum effects. If you model protein folding using a quantum computer it can be theoretically solved much easier and orders of magnitude faster. I mean if you wanted you could just grab some actual proteins and then measure them folding and that's your quantum computer right there. You could also use q-bits and an actual quantum computer to model the same. Quantum computing is real and it works. As for Shor's algorithm it is supposed to change the NP problem of factoring into polynomial time which would indeed break RSA and all our current encryption. The current quantum computers operate on the order of numbers between 1 and 100 maybe, and finding results there requires measuring the phase of that order. Finding the phase of a quantum computer, turning it from quantum into classical for a number that is 1024 bits which seems possible experimentally, but if you increase the size of the RSA key by a factor of 10 that puts it beyond any measurement machine that we can possibly conceive. The big question is does Shor's algorithm break RSA and the answer is maybe. I wish you would just drop the entire veratasium response and explain the entire thing from start to finish. You are requiring me to go watch that veritasium video to understand everything you are talking about and I'm not in the mood to watch a video about quantum computing where the youtuber doesn't know the subject. Of course I could be completely wrong about some of my response, but I'm still interested in you taking a deeper dive into quantum computing. I'm not sure I like your pie analogy as the key to quantum factoring is using Euler's discover that there's a pattern to taking the modulus of a number when it exponentiated, I forget what the theorem is but that theorem is key.
" Secondly due to superposition some mathematical problems can be solved extremely easily with quantum computing instead of classical computing and this does not violate entropy or information theory. " Cite one. One single problem that is proven to be faster on a quantum computer than any possible classical implementation. State the problem mathematically, not "Just use Grover's algorithm." State the exact mathematical problem. You can't use integer factorization because there is no proof that this cannot be performed in polynomial time. So, please state the mathematical problems you claim exist. They don't. "If you model protein folding using a quantum computer it can be theoretically solved much easier and orders of magnitude faster." As they say in Sparta, "If." "I wish you would just drop the entire veratasium response and explain the entire thing from start to finish." This is a channel about Mount Everest and misinformation. The point of this video is to develop the ability to spot misinformation, even if you do not know the underlying information. Certainly, if you were all knowing, you would not be fooled by misinformation. However, this is not a realistic goal. Instead, you have to accept that you will not always know everything, but you can still spot false information without knowing what the true information is. "I'm not sure I like your pie analogy as the key to quantum factoring is using Euler's discover that there's a pattern to taking the modulus of a number when it exponentiated, I forget what the theorem is but that theorem is key." It's Fermat's little theorem and Euler's generalization of that, and as I state in the video, Shor's quantum piece has nothing to do with factoring numbers, so that could not possibly be the "key." If you actually read Shor's paper, he gives two algorithms -- integer factorization and discrete algorithms. The same quantum technique is used for both. As discrete algorithms do not rely on Fermat's little theorem, it could not possibly be the "key" for an algorithm that doesn't use it -- nor does the quantum algorithm for integer factorization. As I clearly state in the video, it is a period finding function --- Wikipedia also labels it a period finding function. One such application of period finding is integer factorization and this relationship uses Fermat's little theorem. But it has nothing to do with period finding. I have a link in the description that has a reference implementation of Shor's that explains the whole thing from start to finish. Putting it in a video does not help to explain it any better.
@@michaeltracy2356 thanks for your in depth response. I read it once, and will have to read it again. I did watch your video twice but probably need to go over it again. If I understand you correctly you think all quantum computing is at best equal to classical computing? That's a bold claim.
@@michaeltracy2356 i thought the paper in nature Resource-efficient quantum algorithm for protein folding Anton Robert, Panagiotis Kl. Barkoutsos, Stefan Woerner & Ivano Tavernelli had promise
When you look at the Variational Quantum Eigensolver (VQE) -- which is essentially what they are using for that protein folding problem, and least for me, there is a tendency to be seduced by the math -- I was. You look at it and say, "Hey, I can just type numpy.linalg.eig and get all the Eigen values and those are exactly what I am looking for. I'll have this NP-hard problem solved in no time." Well, the problem with NP-hard problems is you can "shut up and calculate," but the calculation will take 100 billion years to finish. You can make a nice mathematic statement and prove it will solve the problem and go get it published. Shor's is the same. The value you are looking for is in the Eigen values. If there really were a nature.linalg.eig then quantum programming would be pretty easy. But once you realize that is not they way it work, you start to think.... No other "hard" problems have been solved by a quantum computer. And this integer factorization/hidden subgroup problem isn't proven to be "hard." Then you look at it and things like Euler's theorem, and Hilbert's 10th (not the actual result, but using Diophantines for primes), and all the patterns in primes and there is just a really good possibility a polynomial or near polynomial solution to the problem exists. Now, in the United States, there are 1.5 million people who have top-secret clearances, and the NSA is the largest employer of mathematicians in the world. So, it is not like there aren't people working on this problem. Those people just can't publish any of their work -- at least that we can read. When you really look at Shor's it is not all that "quantummy." The matrixes it uses are all permutation matrixes. It is nothing that complex. So, we are left with a couple of possibilities: Quantum computers really can solve NP-hard problems like protein folding and once we crack that every problem in the universe will be solved. Or if integers can be factored in polynomial time, then there really isn't anything left that Quantum Computers do any better and the universe will go on just as it always has. It would be nice if the universe would throw us a bone. But so far, she hasn't seen fit to do so.
So. The heavy lifting is done by a conventional computer, the quantum computer does the incredibly fast number crunching but has a 75% error rate. It passes the results back to conventional, which tosses the errors, and those don’t matter because the errors are discernible by the conventional. And the whole thing is so screaming fast that getting only 25% product is pure gravy?
Mostly. The quantum computer can find large periods at about the same speed it finds small one. For simplicity I'll say it takes 1 day to find the period for any sized number currently used in public key cryptography. However, it would also take the quantum computer about the same amount of time to factor 77 -- certainly it would take far longer for a quantum computer to factor 77 than a classical one. In addition, we have the ability to solve problems that take "1 billion years" to solve. That is, we have 1 billion computers, and we just run the problem on them for a year. The advantage of the classical computers is that this particular algorithm can be split up on a classical computer, but on a quantum computer you can only have multiple computers running the same problem to deal with the errors -- so roughly 4 quantum computers would give you a speed up, but 100 quantum computers would not factor that single problem any faster. But if you instead had 100 billion computers, then the classical computer would be 100 times faster than 1 billion (for this algorithm). This is because for factoring, you can split up which portions will be checked and don't need to communicate your negatives. And on classical computers, a "core" will count as a computer. So, if you have a 32 core computer, you have 32 "computers". Just write a malware script that takes over people's unused cores to factor your numbers and you will easily have 1 billion "computers" working on the problem.
@@michaeltracy2356 Well, there is definitely a HUGE unused core capacity out there. SETI tried to do that same sort of thing, but voluntarily. BTW, right now the brute force decryption of password hash files are done with a “farm” of graphic processor cards/ chips. People try to make complex passwords, but it is really the semi-complex password that are 28+ characters long are the ones that end up on the cutting room floor of there hash attacks.
as someone who enjoys the pbs spacetime vids on quantum mechanics, was extremely excited to see veritasiums vid on quantum computers. I dont have much actual math knowledge, so I was hoping his video would be a decent laymans explanation of how they work, but it seems the video might have some problems. appreciate your vid here!
He has made more videos about quantum computers in the past, but they are much more technical. He has leaned into more of a pop-sci type of format in his videos in recent years.
Lovely video Michael. There's a lot of hype surrounding quantum computers - mainly pushed by government funded scientists and their friends in the scientific press. Whilst a workable quantum computer would be a great achievement it would only be useful for a handful of computations. Classical computers will remain suitable for most tasks. One quantum bugbear I have is the misunderstandings I often come across with quantum entanglement - that it allows for faster-than-light communication.
I guess it depends on how you define "communication" and how you define "speed of light." See Veratasium: ruclips.net/video/pTn6Ewhb27k/видео.html. I suppose you could set this up to answer the question asked by Veratasium in that video -- the one he says is unknowable. That is, if the "speed of light" is indeed c/2 and then instantaneous, your "communication" would not be "faster than light" because of that phenomenon. Alternately, it could show the one-way speed of light.
This is like being in high school history class listening to your mild mannered favorite teacher talking about Mt. Everest when suddenly a government helicopter lands to sweep him away, all because it turns out your teacher is the only person in the world that can solve some interdimensional fracture in our space time.
This is my field of work, quantum hardware (PhD). Thanks for covering this topic. I participate in both the technical and non-technical coverage of quantum hardware. On that note, I'll be delivering some invited quantum hardware lecture series on the machines/chips we've been building at my research lab and more. It will be open to the public next week on April 12th with my IBM friends and QuantumGrad. You are all welcomed to join us to learn about what tools and equipment we use to build real quantum devices in the field. Ignore the naysayers, let's keep building and learning from what we build. -Onri the Diné Quantum Engineer
@@michaeltracy2356 thank you!!! you have me hooked. hope they find irvine soon. any news on Nat Geo's drone footage? any interest in doing a video about mallory in the US in 1923? was he struggling to raise funds in UK? all the best Michael.
What? i didnt know there was going to be math. Michael can you break down Mallory rope, found around his body and the ipmlications of where Irvine might be.
Hi Michael, have you ever made a video about the first attempt on Everest in 1922? Here is someone who had a go at it: ruclips.net/video/E5oJi2guPxM/видео.html
I'm not entirely sure I agree with you. Much of the way veritassium explains things is intended for the layemn audience. I'm happy to sit here and listen/understand the guts of the quantum forier transform because I'm highly interested in the subject, but for most people if you try to explain these actual ramifications you're going to lose them. He disclaims his own video at the end to say he's simplifying things. Take for example what you consider his disregard of the pigeonhole problem. I think it's unreasonable of you to expect the layperson (who doesn't watch 3 blue 1 brown) to even be aware of the pigeonhole problem! Sure, veritassium could have spent 5 minutes in his video explaining it. He could also have spent another 5 minutes explaining reversible formulas for quantum computing. He also could have spent another 5 minutes on the shore algorithm itself. But do you see where that's going? Yes, it's hand wavey to just say "oh it computes everything at the same time". Yes, it's not correct to say a quantum computer is faster for calculating smaller factors (he only implies this, and his explination goes on to show that it really is better for bigger factors). But I think you'd lose most people if you walked that road in a Veritassium video. That said, I am very grateful you put this together because it certianly answers some questions I had about quantum computing!
There is a difference between explaining things to a layman audience and being just plain wrong. If I explained that the "world is flat" because a layman can look out and it looks flat, that is not an explanation for a layman -- it is just plain wrong. It would make no difference that a layman would likely never meaningfully use the knowledge that the world is not flat. This also applies to simplifying things. He did not simplify them. He made them more complex because they were just wrong. The concepts are fairly easy to to understand, and while I did not criticize his video for disregarding the pigeonhole principle, as you state, the "pigeonhole principle" is so basic and easy to understand that I did not even describe it in the video -- I just assumed that people would know that two pigeons cannot fit into a hole that can only fit one pigeon. What you are missing from your analysis is that Veritasum has no idea how the algorithm works. That is why he explained it incorrectly. It is fine to put people on pedestals and think they are these great geniuses that deign to explain things to us "laymen." But your emperor has no clothes. And if you look at some of the other videos in this channel, you will see that is a recurring theme. As for explaining things to "laymen," from that point of view his entire video is wrong. RSA keys have been broken using numerous different techniques. Even implying to a layman that someone needs a quantum computer is misleading. The main problem with RSA is not attacks from "quantum computers." It is the numerous attacks that have successfully broken keys numerous times in the past and will continue to do so. So easy, even a layman could understand it. But there is a big difference between being able to understand and wanting to believe. And, again, if you look at the videos in this channel, you will see that the willingness to be deceived is frequently discussed. And using RSA is simply people being willing to be deceived. You don't need some quantum computer to break it, but better to believe that you do. If you are not familiar with some of the basic techniques to break RSA, you might start with Wikipedia which describes them.
@@michaeltracy2356 Let me address your points individually: 1. Your analogy about the world being round is misplaced. Veritassium is not calling quantum computers flat. He's calling them round. YOU are calling the world oblong and tilted on its axis. For the purpose of navigating the globe, creating GPS systems, and fully understanding our universe, you are more accurate. That doesn't make calling the world round-like-a-basketball entirely wrong, just less accurate. We need to start with 'earth is round' before we can get to 'earth is more than just round'. That isn't only true for scientific progress, it's true for learning any subject. You may find this easy to understand, but that doesn't mean that's the case for everyone. 2. He may very well not understand the full complexities of the algorithm. That's exactly why he isn't broaching the subject fully. but, he's using a simplification that is in common nomenclature. If that simplification were inaccurate, wouldn't more people be up in arms about it? I won't be convinced that the heavily policied Wikipedia is leaving false information up because nobody except YOU noticed. (And then, if that were the case, why haven't you corrected it)? Also, I don't appreciate you shitting on Veritassium. You yourself acknowledge that the majority of his videos are fully accurate. You're out here trying to dunk on him for clout, that's painfully obvious, just stop. 3. RSA Keys have been broken by virtue of their physical simplicity. Veritassium explains full well why a 2048 bit key hasn't been broken yet, but quantum mechanics could readily allow us to break them. The fact that you're not acknowledging the difference that the quantum Fourier transform makes for decrypting large keys shows to me that you either don't understand the implications yourself (which I highly doubt based on the content of the video), or you're genuinely just out here trying to dunk on Veritassium for views. So, in summary, you're intentionally misinterpreting a common simplification of a complex problem because you're 'so smart you get it', and you're dunking on Veritassium for being an idiot because he's a 'sham blinded by celebrity', and why would anybody watch his videos when you exist? In fact, your entire attitude right now is so patronizing. You have no idea who I am or what I know, and yet you talk down to me? You should really consider your method for handling criticism if you want to make it far in this world.
You are correct that I have no idea who you are. That is why I respond to your statements, not to your personality or character. I was not "shitting" on Veritatisum. I stated he didn't know how the algorithm worked -- a statement you seem to agree is correct or at least as to its "full complexities." 2048 bit keys have been broken. I directed you to the Wikipedia page. You ignored my advice, and then just spout off with "information" that fits your narrative without addressing anything I said. That is just talking past people, and while it may get you very far in the world, it is not the way deal with people. You also just define things to be whatever will make you "right." For instance, you say the 2048 hasn't been "broken" but this simply means it has not been "broken" by an attack by factoring a number that is not known to have a weakness. While "true," this is also meaningless, as 2048 keys have been "broken" by other means and continue to be broken by other means. Had you bothered to read the Wikipedia page, you would know this. Nor is there a need to simplify things to me when it makes your statement less accurate. You state, "RSA Keys have been broken by virtue of their physical simplicity." I do not know what that means. They are not physical objects. They have been broken using a number of techniques, the most inventive being tricking people into using easy to factor numbers -- a technique used by the NSA. Another is to check for duplicate keys which also broke a large number. Another is to exploit the "randomness" of the generated keys. I do not see how any of that related to "physical simplicity," but that is just a contrived phrase that could mean anything. If you meant that it had been broken by tricking the generator to use an algorithm to generate keys with small periods, you should just say so. You also assume that there is a quantum speed up in factoring numbers -- a belief, but nothing more. Understanding Shor's algorithm does not require you know whether a classical algorithm in polynomial time exists or not. However, understanding how some aspect of quantum computing provides a definitive advantage over classical algorithms for factoring does require that you know no such polynomial time algorithm exists. While various aspects of "quantum supremacy" have been demonstrated, none of them show that factoring cannot be done classically in polynomial time. Should a polynomial time algorithm for factoring exist, then obviously the QFT does not provide any "key" to the speed up, as there would be no speed up. I any case, Beaurerad shows how to use a single qubit to measure out the result as a type of "single qubit QFT." As a quantum algorithm exists that performs the algorithm without performing a full QFT on the register "all at once," this is clearly not the big secret to the algorithm. Beuargard's algorithm replaces much of the "QFT" with classical computations of the rotations, so clearly the "QFT" can be reduced to single qubit measurements coupled with classical computations. While I do not expect Veritasium to get into such detail, that is not the basis for the criticism. There have been a dearth of algorithms that provide exponential speed ups over classical methods. You seem more interested is stating, without proof, that these classical problems are indeed "hard" rather than looking into why these exponential speed ups have failed to materialize. This fundamentally gets to the "compute everything at once" problem.
As to: I won't be convinced that the heavily policied Wikipedia is leaving false information up because nobody except YOU noticed. (And then, if that were the case, why haven't you corrected it)?" I have a whole video about this: ruclips.net/video/GJBJYtJfspM/видео.html It is very common for people to believe that about Wikipedia -- that there is a bunch of people out there policing it and keeping it accurate. Simply is not the case, but people like to believe it is. Note again, the willingness to be deceived there because it is much easier to go along with the mainstream than to step into the debate yourself. Wikipedia is just a tool. If it helps you learn, then great. It has some useful info on the RSA hacks. But thinking it is accurate because of the reasons you state is not very realistic.
Just after saying that a quantum computer can do a lot of calculations at the same time, he says that in reality with one measurement we cannot access everything A bit of bad faith on this argument, but the rest is maybe correct
How does the quantum computer store the factions? It doesn't. How does the measurement contain multiple of the same remainder? It doesn't. As I point out, little to nothing in the video is correct. The "all the calculations at the same time" is just the tip of the pseudo-science nonsense iceberg that makes up the video. The rest is not "maybe correct." It is definitely incorrect, but I am interested in why you think it is correct. What about the presentation causes you to believe a quantum computer works in the fashion he describes? Why do you believe that when you take a measurement, it gives you multiple copies of the same remainder? Given that Shor's paper clearly shows this is not the way the algorithm works, why would you believe some RUclips video that doesn't even mention the paper nor cite a single reference? Did you try running the reference implementation I provided? Or anyone else's implementation?
@@michaeltracy2356 the maybe correct was referring to your arguments I was just saying the all calculations at the same time + you need them to interfer to make the correct answer rise is not so bad Like if you take Grover algorithm, you start from a superpostion of all n bits strings, then you apply the oracle f to all the bitstrings and get the superpostion sum over |i> |f(i)> and them make them interfer to get the i where f equals 1
That is how religions get started. It is a great explanation that "god makes the rain." And it doesn't rain when she is angry for whatever reason. So, I can follow your "logic." It can never be proven incorrect because you can never say exactly what makes god angry -- for instance, perhaps airplanes seeding the clouds makes the god happy and that is what causes it to rain. The problem with both the "god makes it rain" argument and the "everything at once" argument is that it gives you no ability to use nature for our benefit -- while actual science does. They are both trite explanations that "explain" things in a simple fashion. Your explanation of Grover's is also incorrect. Apply the "oracle" to a uniform superposition will not do anything. You have to alter one of the encoded states first. Also not sure what your notation is. Click the link and see my implementation of Grovers, and it will show exactly what it does. A big part of this channel is how we can know something without having absolute certainty. Quantum computing is probabilistic computing and Verisasium's video ignores this. Your own "example" misses the difference between amplitude encoding and amplitude amplification -- which is what the algorithm really does. While it is common to do simple demonstrations with a deterministic result, the only time you really get that is with an extremely simple problem. Grovers, like Shors, is a probabilistic algorithm for all but the most trivial of problems. The advantage of Shor's is that we have an easy way to test whether the probabilistic answer is correct and thus we can know with certainty that it provided the correct answer. This is not the case for Grovers for anything other than the academic problem -- that is, it is returning the exact same result that you just "marked." So, you know it is correct because you are the one that put the correct answer into the system and you just compare the answer you put into the system with the one the computer output. I hope you are starting to see what the problem is with using Grover's as an example. When you have the correct answer, you input it into the computer, it "computes everything at the same time", and then gives you an incorrect answer some percentage of the time -- your whole explanation of how quantum computers work doesn't work. At least if you went with the quantum god being angry, there would be an explanation as to why the incorrect answer came out.
Funny how nobody talks about the Yeti being the first to summit everest and even more incredibly how it was the first to reach the summit WITHOUT the use of oxygen. I for one would like to pay homage to this creature by not shaving my beard or hair for a period of 10 years.
Furthermore I would like to announce my dismay that not only is the Yeti not discussed in climbing records, but also the lack of research of Yeti or Yetifolk collecting summit rocks. More funding needs to be devoted to this topic.
After doing some brief research I've found that some Yetis can be as large as 10 feet, meaning as many sherpas have speculated the second step wouldn't have been an empassable obstacle, some academics however remained unconvinced.
starting....NOW!
@@101noz101 And how much of that funding do you intend go to Yetifolk themselves! We cannot exploit Yetifolk for science, the way colonial science has been done in the past! 🤣
I basically understood nothing in this video but entertaining! lmao
Huh? You lost me after the 2nd step!
First of all when Richard Feynman says that "nobody knows how quantum physics works" he doesn't mean we can't calculate it. We can calculate it exactly. Everything down to 10 decimal places, aside from stuff we don't know yet like dark matter. So we can exactly calculate how quantum computers work. What Feynman meant is that nobody knows why the universe works the way that it works. It just does. Secondly due to superposition some mathematical problems can be solved extremely easily with quantum computing instead of classical computing and this does not violate entropy or information theory. The greatest drawback to quantum computing is translating quantum results into the classical world. This is where you typically add a bunch of entropy and heat to the process. It's extracting a classical result from quantum that takes time and energy. There is a problem in chemistry called the "protein folding" problem which requires so much classical resources that it is virtually insolvable except for certain simple cases such as 2 dimensional lattices. However, the proteins to in fact fold. How do they do it? Using quantum effects. If you model protein folding using a quantum computer it can be theoretically solved much easier and orders of magnitude faster. I mean if you wanted you could just grab some actual proteins and then measure them folding and that's your quantum computer right there. You could also use q-bits and an actual quantum computer to model the same. Quantum computing is real and it works. As for Shor's algorithm it is supposed to change the NP problem of factoring into polynomial time which would indeed break RSA and all our current encryption. The current quantum computers operate on the order of numbers between 1 and 100 maybe, and finding results there requires measuring the phase of that order. Finding the phase of a quantum computer, turning it from quantum into classical for a number that is 1024 bits which seems possible experimentally, but if you increase the size of the RSA key by a factor of 10 that puts it beyond any measurement machine that we can possibly conceive. The big question is does Shor's algorithm break RSA and the answer is maybe. I wish you would just drop the entire veratasium response and explain the entire thing from start to finish. You are requiring me to go watch that veritasium video to understand everything you are talking about and I'm not in the mood to watch a video about quantum computing where the youtuber doesn't know the subject. Of course I could be completely wrong about some of my response, but I'm still interested in you taking a deeper dive into quantum computing. I'm not sure I like your pie analogy as the key to quantum factoring is using Euler's discover that there's a pattern to taking the modulus of a number when it exponentiated, I forget what the theorem is but that theorem is key.
" Secondly due to superposition some mathematical problems can be solved extremely easily with quantum computing instead of classical computing and this does not violate entropy or information theory. " Cite one. One single problem that is proven to be faster on a quantum computer than any possible classical implementation. State the problem mathematically, not "Just use Grover's algorithm." State the exact mathematical problem. You can't use integer factorization because there is no proof that this cannot be performed in polynomial time. So, please state the mathematical problems you claim exist. They don't.
"If you model protein folding using a quantum computer it can be theoretically solved much easier and orders of magnitude faster."
As they say in Sparta, "If."
"I wish you would just drop the entire veratasium response and explain the entire thing from start to finish." This is a channel about Mount Everest and misinformation. The point of this video is to develop the ability to spot misinformation, even if you do not know the underlying information. Certainly, if you were all knowing, you would not be fooled by misinformation. However, this is not a realistic goal. Instead, you have to accept that you will not always know everything, but you can still spot false information without knowing what the true information is.
"I'm not sure I like your pie analogy as the key to quantum factoring is using Euler's discover that there's a pattern to taking the modulus of a number when it exponentiated, I forget what the theorem is but that theorem is key." It's Fermat's little theorem and Euler's generalization of that, and as I state in the video, Shor's quantum piece has nothing to do with factoring numbers, so that could not possibly be the "key." If you actually read Shor's paper, he gives two algorithms -- integer factorization and discrete algorithms. The same quantum technique is used for both. As discrete algorithms do not rely on Fermat's little theorem, it could not possibly be the "key" for an algorithm that doesn't use it -- nor does the quantum algorithm for integer factorization. As I clearly state in the video, it is a period finding function --- Wikipedia also labels it a period finding function. One such application of period finding is integer factorization and this relationship uses Fermat's little theorem. But it has nothing to do with period finding.
I have a link in the description that has a reference implementation of Shor's that explains the whole thing from start to finish. Putting it in a video does not help to explain it any better.
@@michaeltracy2356 thanks for your in depth response. I read it once, and will have to read it again. I did watch your video twice but probably need to go over it again. If I understand you correctly you think all quantum computing is at best equal to classical computing? That's a bold claim.
@@michaeltracy2356 by the way I do think quantum computing is somewhat over hyped but not a total loss
@@michaeltracy2356 i thought the paper in nature Resource-efficient quantum algorithm for protein folding
Anton Robert, Panagiotis Kl. Barkoutsos, Stefan Woerner & Ivano Tavernelli had promise
When you look at the Variational Quantum Eigensolver (VQE) -- which is essentially what they are using for that protein folding problem, and least for me, there is a tendency to be seduced by the math -- I was. You look at it and say, "Hey, I can just type numpy.linalg.eig and get all the Eigen values and those are exactly what I am looking for. I'll have this NP-hard problem solved in no time." Well, the problem with NP-hard problems is you can "shut up and calculate," but the calculation will take 100 billion years to finish. You can make a nice mathematic statement and prove it will solve the problem and go get it published.
Shor's is the same. The value you are looking for is in the Eigen values. If there really were a nature.linalg.eig then quantum programming would be pretty easy. But once you realize that is not they way it work, you start to think....
No other "hard" problems have been solved by a quantum computer. And this integer factorization/hidden subgroup problem isn't proven to be "hard." Then you look at it and things like Euler's theorem, and Hilbert's 10th (not the actual result, but using Diophantines for primes), and all the patterns in primes and there is just a really good possibility a polynomial or near polynomial solution to the problem exists.
Now, in the United States, there are 1.5 million people who have top-secret clearances, and the NSA is the largest employer of mathematicians in the world. So, it is not like there aren't people working on this problem. Those people just can't publish any of their work -- at least that we can read.
When you really look at Shor's it is not all that "quantummy." The matrixes it uses are all permutation matrixes. It is nothing that complex.
So, we are left with a couple of possibilities: Quantum computers really can solve NP-hard problems like protein folding and once we crack that every problem in the universe will be solved. Or if integers can be factored in polynomial time, then there really isn't anything left that Quantum Computers do any better and the universe will go on just as it always has.
It would be nice if the universe would throw us a bone. But so far, she hasn't seen fit to do so.
A break from the norm, I'm in 😊
Update: Really enjoyed that, and it was put very politely too.
So. The heavy lifting is done by a conventional computer, the quantum computer does the incredibly fast number crunching but has a 75% error rate. It passes the results back to conventional, which tosses the errors, and those don’t matter because the errors are discernible by the conventional. And the whole thing is so screaming fast that getting only 25% product is pure gravy?
Mostly. The quantum computer can find large periods at about the same speed it finds small one. For simplicity I'll say it takes 1 day to find the period for any sized number currently used in public key cryptography. However, it would also take the quantum computer about the same amount of time to factor 77 -- certainly it would take far longer for a quantum computer to factor 77 than a classical one. In addition, we have the ability to solve problems that take "1 billion years" to solve. That is, we have 1 billion computers, and we just run the problem on them for a year. The advantage of the classical computers is that this particular algorithm can be split up on a classical computer, but on a quantum computer you can only have multiple computers running the same problem to deal with the errors -- so roughly 4 quantum computers would give you a speed up, but 100 quantum computers would not factor that single problem any faster. But if you instead had 100 billion computers, then the classical computer would be 100 times faster than 1 billion (for this algorithm). This is because for factoring, you can split up which portions will be checked and don't need to communicate your negatives.
And on classical computers, a "core" will count as a computer. So, if you have a 32 core computer, you have 32 "computers". Just write a malware script that takes over people's unused cores to factor your numbers and you will easily have 1 billion "computers" working on the problem.
@@michaeltracy2356 Well, there is definitely a HUGE unused core capacity out there. SETI tried to do that same sort of thing, but voluntarily. BTW, right now the brute force decryption of password hash files are done with a “farm” of graphic processor cards/ chips. People try to make complex passwords, but it is really the semi-complex password that are 28+ characters long are the ones that end up on the cutting room floor of there hash attacks.
I got lost getting on the ladder at the second step 😂
minus one? 🤔😏
Michael Tracy is the most intelligent mountaineer, no doubt about it.
as someone who enjoys the pbs spacetime vids on quantum mechanics, was extremely excited to see veritasiums vid on quantum computers. I dont have much actual math knowledge, so I was hoping his video would be a decent laymans explanation of how they work, but it seems the video might have some problems. appreciate your vid here!
He has made more videos about quantum computers in the past, but they are much more technical. He has leaned into more of a pop-sci type of format in his videos in recent years.
That redemption suggestion is perfect.
One can hope.
Thank you. I tried to do a Python experiment to understand what Veritassium was saying and it wouldn't make any sense.
Lovely video Michael. There's a lot of hype surrounding quantum computers - mainly pushed by government funded scientists and their friends in the scientific press. Whilst a workable quantum computer would be a great achievement it would only be useful for a handful of computations. Classical computers will remain suitable for most tasks. One quantum bugbear I have is the misunderstandings I often come across with quantum entanglement - that it allows for faster-than-light communication.
I guess it depends on how you define "communication" and how you define "speed of light." See Veratasium: ruclips.net/video/pTn6Ewhb27k/видео.html. I suppose you could set this up to answer the question asked by Veratasium in that video -- the one he says is unknowable. That is, if the "speed of light" is indeed c/2 and then instantaneous, your "communication" would not be "faster than light" because of that phenomenon. Alternately, it could show the one-way speed of light.
So what do you make of the recent quantum teleportation findings?
"Forget about computers" -- H. W. Tilman
A most interesting video. Thanks and thumbs up.
This is like being in high school history class listening to your mild mannered favorite teacher talking about Mt. Everest when suddenly a government helicopter lands to sweep him away, all because it turns out your teacher is the only person in the world that can solve some interdimensional fracture in our space time.
Andy Weir's "Project Hail Mary" opens pretty much like this. Except it's just outer space, not an interdimensional fracture
Well, this is interesting!
This is my field of work, quantum hardware (PhD). Thanks for covering this topic. I participate in both the technical and non-technical coverage of quantum hardware. On that note, I'll be delivering some invited quantum hardware lecture series on the machines/chips we've been building at my research lab and more. It will be open to the public next week on April 12th with my IBM friends and QuantumGrad. You are all welcomed to join us to learn about what tools and equipment we use to build real quantum devices in the field. Ignore the naysayers, let's keep building and learning from what we build. -Onri the Diné Quantum Engineer
Maybe gpt4 can tell us if M&I made it to the summit....
Wow, I had just watched the Veritasium's video and never imagined so much might be wrong.
Hello Michael, I was wondering if you could provide a link to your Discord server?
discord.gg/JAMb2VUp
more mallory and irvine please
Going to be looking more at the Chinese 1960 "summit." I'll have more Mallory and Irvine content after the visit to the archives.
@@michaeltracy2356 thank you!!! you have me hooked. hope they find irvine soon. any news on Nat Geo's drone footage? any interest in doing a video about mallory in the US in 1923? was he struggling to raise funds in UK? all the best Michael.
What? i didnt know there was going to be math. Michael can you break down Mallory rope, found around his body and the ipmlications of where Irvine might be.
Hi Michael, have you ever made a video about the first attempt on Everest in 1922? Here is someone who had a go at it: ruclips.net/video/E5oJi2guPxM/видео.html
That's a pretty good video. Hope he makes one on the 1924.
I'm not entirely sure I agree with you.
Much of the way veritassium explains things is intended for the layemn audience. I'm happy to sit here and listen/understand the guts of the quantum forier transform because I'm highly interested in the subject, but for most people if you try to explain these actual ramifications you're going to lose them.
He disclaims his own video at the end to say he's simplifying things.
Take for example what you consider his disregard of the pigeonhole problem. I think it's unreasonable of you to expect the layperson (who doesn't watch 3 blue 1 brown) to even be aware of the pigeonhole problem!
Sure, veritassium could have spent 5 minutes in his video explaining it. He could also have spent another 5 minutes explaining reversible formulas for quantum computing. He also could have spent another 5 minutes on the shore algorithm itself.
But do you see where that's going? Yes, it's hand wavey to just say "oh it computes everything at the same time". Yes, it's not correct to say a quantum computer is faster for calculating smaller factors (he only implies this, and his explination goes on to show that it really is better for bigger factors).
But I think you'd lose most people if you walked that road in a Veritassium video.
That said, I am very grateful you put this together because it certianly answers some questions I had about quantum computing!
There is a difference between explaining things to a layman audience and being just plain wrong. If I explained that the "world is flat" because a layman can look out and it looks flat, that is not an explanation for a layman -- it is just plain wrong. It would make no difference that a layman would likely never meaningfully use the knowledge that the world is not flat.
This also applies to simplifying things. He did not simplify them. He made them more complex because they were just wrong. The concepts are fairly easy to to understand, and while I did not criticize his video for disregarding the pigeonhole principle, as you state, the "pigeonhole principle" is so basic and easy to understand that I did not even describe it in the video -- I just assumed that people would know that two pigeons cannot fit into a hole that can only fit one pigeon.
What you are missing from your analysis is that Veritasum has no idea how the algorithm works. That is why he explained it incorrectly. It is fine to put people on pedestals and think they are these great geniuses that deign to explain things to us "laymen." But your emperor has no clothes. And if you look at some of the other videos in this channel, you will see that is a recurring theme.
As for explaining things to "laymen," from that point of view his entire video is wrong. RSA keys have been broken using numerous different techniques. Even implying to a layman that someone needs a quantum computer is misleading. The main problem with RSA is not attacks from "quantum computers." It is the numerous attacks that have successfully broken keys numerous times in the past and will continue to do so. So easy, even a layman could understand it. But there is a big difference between being able to understand and wanting to believe. And, again, if you look at the videos in this channel, you will see that the willingness to be deceived is frequently discussed. And using RSA is simply people being willing to be deceived. You don't need some quantum computer to break it, but better to believe that you do. If you are not familiar with some of the basic techniques to break RSA, you might start with Wikipedia which describes them.
@@michaeltracy2356 Let me address your points individually:
1. Your analogy about the world being round is misplaced. Veritassium is not calling quantum computers flat. He's calling them round. YOU are calling the world oblong and tilted on its axis. For the purpose of navigating the globe, creating GPS systems, and fully understanding our universe, you are more accurate. That doesn't make calling the world round-like-a-basketball entirely wrong, just less accurate. We need to start with 'earth is round' before we can get to 'earth is more than just round'. That isn't only true for scientific progress, it's true for learning any subject. You may find this easy to understand, but that doesn't mean that's the case for everyone.
2. He may very well not understand the full complexities of the algorithm. That's exactly why he isn't broaching the subject fully. but, he's using a simplification that is in common nomenclature. If that simplification were inaccurate, wouldn't more people be up in arms about it? I won't be convinced that the heavily policied Wikipedia is leaving false information up because nobody except YOU noticed. (And then, if that were the case, why haven't you corrected it)?
Also, I don't appreciate you shitting on Veritassium. You yourself acknowledge that the majority of his videos are fully accurate. You're out here trying to dunk on him for clout, that's painfully obvious, just stop.
3. RSA Keys have been broken by virtue of their physical simplicity. Veritassium explains full well why a 2048 bit key hasn't been broken yet, but quantum mechanics could readily allow us to break them. The fact that you're not acknowledging the difference that the quantum Fourier transform makes for decrypting large keys shows to me that you either don't understand the implications yourself (which I highly doubt based on the content of the video), or you're genuinely just out here trying to dunk on Veritassium for views.
So, in summary, you're intentionally misinterpreting a common simplification of a complex problem because you're 'so smart you get it', and you're dunking on Veritassium for being an idiot because he's a 'sham blinded by celebrity', and why would anybody watch his videos when you exist?
In fact, your entire attitude right now is so patronizing. You have no idea who I am or what I know, and yet you talk down to me?
You should really consider your method for handling criticism if you want to make it far in this world.
You are correct that I have no idea who you are. That is why I respond to your statements, not to your personality or character.
I was not "shitting" on Veritatisum. I stated he didn't know how the algorithm worked -- a statement you seem to agree is correct or at least as to its "full complexities."
2048 bit keys have been broken. I directed you to the Wikipedia page. You ignored my advice, and then just spout off with "information" that fits your narrative without addressing anything I said. That is just talking past people, and while it may get you very far in the world, it is not the way deal with people.
You also just define things to be whatever will make you "right." For instance, you say the 2048 hasn't been "broken" but this simply means it has not been "broken" by an attack by factoring a number that is not known to have a weakness. While "true," this is also meaningless, as 2048 keys have been "broken" by other means and continue to be broken by other means. Had you bothered to read the Wikipedia page, you would know this.
Nor is there a need to simplify things to me when it makes your statement less accurate. You state, "RSA Keys have been broken by virtue of their physical simplicity." I do not know what that means. They are not physical objects. They have been broken using a number of techniques, the most inventive being tricking people into using easy to factor numbers -- a technique used by the NSA. Another is to check for duplicate keys which also broke a large number. Another is to exploit the "randomness" of the generated keys. I do not see how any of that related to "physical simplicity," but that is just a contrived phrase that could mean anything. If you meant that it had been broken by tricking the generator to use an algorithm to generate keys with small periods, you should just say so.
You also assume that there is a quantum speed up in factoring numbers -- a belief, but nothing more. Understanding Shor's algorithm does not require you know whether a classical algorithm in polynomial time exists or not. However, understanding how some aspect of quantum computing provides a definitive advantage over classical algorithms for factoring does require that you know no such polynomial time algorithm exists. While various aspects of "quantum supremacy" have been demonstrated, none of them show that factoring cannot be done classically in polynomial time. Should a polynomial time algorithm for factoring exist, then obviously the QFT does not provide any "key" to the speed up, as there would be no speed up. I any case, Beaurerad shows how to use a single qubit to measure out the result as a type of "single qubit QFT." As a quantum algorithm exists that performs the algorithm without performing a full QFT on the register "all at once," this is clearly not the big secret to the algorithm. Beuargard's algorithm replaces much of the "QFT" with classical computations of the rotations, so clearly the "QFT" can be reduced to single qubit measurements coupled with classical computations. While I do not expect Veritasium to get into such detail, that is not the basis for the criticism.
There have been a dearth of algorithms that provide exponential speed ups over classical methods. You seem more interested is stating, without proof, that these classical problems are indeed "hard" rather than looking into why these exponential speed ups have failed to materialize. This fundamentally gets to the "compute everything at once" problem.
As to: I won't be convinced that the heavily policied Wikipedia is leaving false information up because nobody except YOU noticed. (And then, if that were the case, why haven't you corrected it)?"
I have a whole video about this: ruclips.net/video/GJBJYtJfspM/видео.html
It is very common for people to believe that about Wikipedia -- that there is a bunch of people out there policing it and keeping it accurate. Simply is not the case, but people like to believe it is. Note again, the willingness to be deceived there because it is much easier to go along with the mainstream than to step into the debate yourself. Wikipedia is just a tool. If it helps you learn, then great. It has some useful info on the RSA hacks. But thinking it is accurate because of the reasons you state is not very realistic.
Just after saying that a quantum computer can do a lot of calculations at the same time, he says that in reality with one measurement we cannot access everything
A bit of bad faith on this argument, but the rest is maybe correct
How does the quantum computer store the factions? It doesn't. How does the measurement contain multiple of the same remainder? It doesn't.
As I point out, little to nothing in the video is correct. The "all the calculations at the same time" is just the tip of the pseudo-science nonsense iceberg that makes up the video. The rest is not "maybe correct." It is definitely incorrect, but I am interested in why you think it is correct.
What about the presentation causes you to believe a quantum computer works in the fashion he describes? Why do you believe that when you take a measurement, it gives you multiple copies of the same remainder? Given that Shor's paper clearly shows this is not the way the algorithm works, why would you believe some RUclips video that doesn't even mention the paper nor cite a single reference? Did you try running the reference implementation I provided? Or anyone else's implementation?
@@michaeltracy2356 the maybe correct was referring to your arguments
I was just saying the all calculations at the same time + you need them to interfer to make the correct answer rise is not so bad
Like if you take Grover algorithm, you start from a superpostion of all n bits strings, then you apply the oracle f to all the bitstrings and get the superpostion sum over |i> |f(i)> and them make them interfer to get the i where f equals 1
That is how religions get started. It is a great explanation that "god makes the rain." And it doesn't rain when she is angry for whatever reason. So, I can follow your "logic." It can never be proven incorrect because you can never say exactly what makes god angry -- for instance, perhaps airplanes seeding the clouds makes the god happy and that is what causes it to rain.
The problem with both the "god makes it rain" argument and the "everything at once" argument is that it gives you no ability to use nature for our benefit -- while actual science does. They are both trite explanations that "explain" things in a simple fashion.
Your explanation of Grover's is also incorrect. Apply the "oracle" to a uniform superposition will not do anything. You have to alter one of the encoded states first. Also not sure what your notation is. Click the link and see my implementation of Grovers, and it will show exactly what it does.
A big part of this channel is how we can know something without having absolute certainty. Quantum computing is probabilistic computing and Verisasium's video ignores this. Your own "example" misses the difference between amplitude encoding and amplitude amplification -- which is what the algorithm really does. While it is common to do simple demonstrations with a deterministic result, the only time you really get that is with an extremely simple problem. Grovers, like Shors, is a probabilistic algorithm for all but the most trivial of problems. The advantage of Shor's is that we have an easy way to test whether the probabilistic answer is correct and thus we can know with certainty that it provided the correct answer. This is not the case for Grovers for anything other than the academic problem -- that is, it is returning the exact same result that you just "marked." So, you know it is correct because you are the one that put the correct answer into the system and you just compare the answer you put into the system with the one the computer output.
I hope you are starting to see what the problem is with using Grover's as an example. When you have the correct answer, you input it into the computer, it "computes everything at the same time", and then gives you an incorrect answer some percentage of the time -- your whole explanation of how quantum computers work doesn't work. At least if you went with the quantum god being angry, there would be an explanation as to why the incorrect answer came out.