100 Comments

  • Roman Klimenko

    Thanks a lot!
    All those "popular science" explanations that try to oversimplify are in fact just confusing even more.

  • Igor Jovcevski

    doesn;t hadamard gate break the quantum coherence, and quantum computer ability to perform all operations in superposition state? I am mainly thinking of using it in the chain of perrforming multiple operations.

  • Chris Davison

    quantum will never beat classical computing on all calculations, there's all kinds of physical systems we could leverage to calculating niche problems. I just dont get the hype. If you had a stack of i9's the size of a quantum computer cooled with liquid nitrogen, it would kick quantum arse. No?

  • CaliPark

    My Millennial colleagues really need to stop their poor habit of asking/saying "Right?" There's some use for this word when you're teaching something, but if you're in the midst of explaining one part of a larger puzzle, asking "Right" doesn't exactly help considering you can't always understand one part of the puzzle until you learn more parts of the puzzle — if that makes sense. Which is ironic, because of all things he's trying to teach Quantum Theory and its applications in Computer Science. It shouldn't be taught classically, using traditional logic at every step to qualify what's true (1) or false (0), when you DON'T KNOW THE ANSWER YET. That's literally the "black box." If you're asking "right" after explaining something that is not readily apparent using traditional logic — you should expect that the answer in the individual's mind will be "yes? no? I don't know." Which is ironic, because you shouldn't be teaching them to always apply logic to a science that has a lot of illogic in it — yet as a system of illogical subsystems, the outcome with those taken as a whole becomes logical… Assbackwards approach to teaching this, my friend. Teach them the outcome and work backwards if you expect them to 1) put together the puzzle pieces and 2) expect to understand how that puzzle was solved and how those puzzle pieces being such as they are, solved the larger puzzle… It's called reverse engineering. Better to teach Quantum Theory using reverse engineering, because scientists are still trying to figure out how the hell it works. Because it's still a THEORY and is actively being explored, with many unsolved paradoxes that seemingly disprove the theory's premises entirely — yet the outcomes say otherwise. So don't expect a layman to comprehend until more concepts are introduced to them. It must be explained to students, constantly, that we still can't explain all of this perfectly because it's not a settled science (though no science is truly settled), and thus they need to roll with it sometimes. That should always be a given. If not, we will not get anywhere in terms of accelerating our scientific discovery..

  • Ken Snyder

    Wow!!! Microsoft has a Research Division? Who knew? I knew that they had a kick-ass marketing division that is without rival. And their legal department is absolutely stellar. (Or should I say quasi-legal dept?). But honestly, I thought that the only research that they did was how to steal intellectual property. This is news to me.

  • Sierox

    1:06:39 If I measure one of the entangled qubits, can someone else know when I measured it, causing their qubit to collapse? Or do they not know WHEN I measured it. If they can know, then FTL communication would be possible. You could measure qubits in short and long intervals which could be interpreted as 0's and 1's, which would enable the transfer of data.

  • Stanley Dodds

    I feel like this whole thing could have been condensed to "a quantum computer is a unitary matrix" and "the input/output is a unit vector"

  • Duck Blockschild

    Finally a real explanation of Quantum computing. Popular stupid analogies of flipping coins don't work. I am not a computer scientists and this video finally helped me understand what QC really are

  • Devon Parsons

    I don't understand the claim that the quantum computer outperformed the classical computer because it only took 'one query', because you've added another input and output. Isn't that 'equivalent' in some sense as the cost of a second query?

  • Sun

    It was hard to follow for me but my interest in this has grown due to this. I did run through this multiple times before it started making sense. I am still not 100% there but getting there. QC is really cool. I love it. Nature is so weird and wonderful. This was a great video. Thanks!

  • __iHexx__

    For anyone else stumbling upon this, here's the slides
    https://www.microsoft.com/en-us/research/uploads/prod/2018/05/40655.compressed.pdf

  • parityviolation

    nice talk, but he should stick to the computing part and leave the physics to the xperts (especially around 1hr in +- a couple of minutes) … What he calls "coordination" and claims to have no term yet due to language limitations is actually called "causality"… And Bell's theorem does not do away with locality but with local realism! Huge difference! I could go on here, but I dont want to bring in negativity to a topic and effort worth applauding.
    I suppose it's a good thing, physicists don't watch 90mins videos with "for computer scientists" in the title ^^ otherwise the comment section probably would look somewhat different^^

  • AFastidiousCuber

    2 Questions:
    1. How do we interpret the probabilities if we allow complex numbers? For example, if the spin of particle is in a superposition represented by (√2/2) + (√2/2)i, what will we expect to see when we make a measurement?
    2. How does this model deal with more than 2 q-bits? Would an n-qubit system be represented by a vector of length n for each q-bit, or of length 2^n for each possible binary string of length n?

  • Normen Yu

    In reference to the confusion in 43:14 regarding naming confusion. Just wondering, is this confusion typical because traditionally, as you mentioned earlier, in a classical computer, you change the variable. Hence the variable wire changes from "input" to "output". However, in the quantum world, the variable stays in place while you change the program. Hence, the wire, which represents a variable instead of a function, is always called the output wire or the input wire. I wonder if it is this shift in perspective that causes the confusion?

  • テェぺシュヴラッド

    I'd like to see things like human choice shown as a Qbit in a superposition (Do/Do Not) until the physical action occurs which is the collapsed stated. I think this kind of thinking is going to revolutionize our way of thinking let alone our way of computing.

  • Funny Geeks

    43:24 I think the output input thing could be more easily understood by C programmers, since we have the concept of returning through parameters.

  • Funny Geeks

    1:10:07 The guy is questioning whether or not superposition actually exists. Is the probabilities due to the fact that we have imprecise tools that cannot observe the particles, or are the particles actually in multiple states in the same time? Copenhagen interpretation says that particles truly are in multiple positions at the same time, and alternative theories like pilot wave theory says that there are things that exist, that we cannot precisely measure. (I don't know if the math behind pilot wave theory has been fully worked out yet)

  • Mea Hula Pelelina

    9:55 "…only use operations that are their own inverses"
    He is wrong there, the phase shift gate S being a counterexample.

  • Ben Ng

    Ok… if I have 2 qubits and I entangle the qubits they can be in a state of either entangled or not 1 or 0. If I take the qubits and I separate them I now have 2 qubits and they are apart. I use a robot to push a button if the qubit is not entangled any more. If I wanted the robot to push the button I would measure the one qubit and the other one would collapse making the button get pushed. How is not a transfer of information.

  • MrHatoi

    I think it would help in the Black Box problem to draw the CNOT gate as having 2 inputs and 2 outputs so the fact that it changes the input makes more sense.

  • Shivansh Beohar

    What if I use two input bits in a classical computer, one is always zero as in QC, other one as 1:
    Const 1 -> 11
    Const 0 -> 00
    Identity -> 10
    Negation -> 01
    Then use XOR gate on the output I will get 0 for constant functions and 1 for variable functions in one query. And since QC was allowed to use 2 qbits in the problem, I think it is fair to give CC that much if we are comparing them.Where is the speed up we are talking about ?

  • vaibhav singh

    1:06:45 So does this mean that I physically separate two entangled qubits, run some operation on one (so that its probability of returning 0 or 1 changes) and then if I collapse the one qubit the second qubits' probability will get updated "somehow"?

  • CameraKad

    Actually As I understand quantum computing is just a different concept of laying/extracting bits using schrödingers state function?? With a bit of advancement and time I think we can reflect and implement those Qbits principles onto the current classic computing?

  • Lordslothable

    Great video but don’t you think opening parallel gates to other worlds is a bit much just to play minesweeper 10,000x slower than my Commodore 64?

  • Mihael Rogelj

    This is supposed to be for scientists, but it's the video from which I learned the most of Q.C. out of them all. All other videos are just barely understandable analogies. And I'm not a scientis, nor a programmer.

  • paul snor

    The whole discussion on the deutsch oracle (36:00) would be much clearer if he finished the arrows and adapted his termology: how must we interpret a system where ‘output’ goes in and ‘input’ comes out?

  • Bernd Eckenfels

    The observation that all metaphors ring wrong for describing quantum algorithms bothered me for a while as well (3:40) however shutup and calculate is bothering in its own way 🙂

  • Evan O'Leary

    25:27 is world-shattering. So quantum physics is not just, naively, "all possibilities are realized", it's "all collectively reversible possibilities are realized"?

  • bmoneeboi

    So for clarification, the constraints of the Deutsch oracle problem allow there to be usage of multiple bits? It's not required that there is just a single bit used for a classical single bit operation as per the constraints of the problem?

  • sjsjsjj1

    Hi I have a question. For the 4 operations, the preprocessing was done, with the assumption that both starts with both values as 0. However when we were solving for variable, X, the preprocessing was applied from the starting point of 0. But X is not 0, right? Shouldn't the preprocessing have started from somewhere else?

  • sjsjsjj1

    Hope to find out what he means by neutralising the difference within the categories, and magnifying the difference between them. What are the categories he is referring to? How can I tell it is neutralised, how can I tell it is magnified?

  • bmoneeboi

    When you have n qbits that are being represented with a single vector through an n-fold tensor product of the product states, there are probabilities for the system to collapse with a 1 at one of the 2^n numbers in the vector. How does this n-qbit system collapse? Can all qbits be measured at once, are they required to be measured all at once? And what does the 1 in the product state represent? With a single qbit, there are probabilities for collapsing to 1 and collapsing to 0. What do the probabilities correspond to with a larger product state of multiple qbits?

  • Chad Bloxham

    Watching this 15 months after upload. Anyone know if Google has reached "quantum supremacy" i.e. do we know if the resulting noise makes the quantum computer unusable?

  • study related

    At 1:00:00, using entanglement for communication is possible in one case that I can think of. If I take a Q-bit, which is entangled to my friend's Q-bit. imagine me traveling to some far place. I tell my friend that if the Q-bit collapses to 1, then perform task A. And if it collapses to 0, then perform task B. And I would know which task he performed because of the Q-bit that I have. This is communication even though I don't send information implicitly…what do u guys think…?

  • Brian Decker

    Quantum Supremacy – still as far in the future as ever. What the last year has shown is that the scaling problem is far larger than anyone had anticipated, and by implication, perhaps the foundation expectations of quantum computing may be flawed – if not impossible.

  • Stuart Heinrich

    Great intro to quantum programming, but the tangent at the end gives false information. Bell's inequality proved that there can be no hidden variable theory only under the assumption that the universe is local. The later discovery of quantum entanglement showed that Bell's inequality actually was proof that all theories must be non-local. Hidden variable theories are NOT disproved by Bell's inequality, they merely fell out of popularity after the initial misinterpretation of Bell's inequality. Bohmian mechanics is a perfect example of a hidden variable theory that is consistent with all known quantum mechanical experiments thus far.

  • Gerardo Moscatelli

    Bell's inequality proves that there aren't hidden binary variables but it still doesn't prove there is any "teleportation" or "spooky action at a distance", you just collapse the wave and measure the same state both sides 100% of the time. Nothing was teleported, we just can't explain this action with binary variables which is quite normal because we are measuring quantum particles not tennis balls.

  • Gerardo Moscatelli

    There is no way to entangle q-bits if they weren't created from the same particle before. The fact that there aren't binary hidden variables doesn't mean they are being teleported faster than the speed of light. You need to entangle the particles before. No magic happening just quantum mechanics

  • Gerardo Moscatelli

    Honestly the usage of the word "teleportation" should be banned from QM. Let's keep this term for Star Trek. There is NO INFORMATION TELEPORTED, Alice and Bob have random results, they need to communicate via normal channel about the operation they performed on the q-bit to have the same sequence of bits, this is not clear in this presentation. No need of mystification of quantum physics with Teleportation, flying saucers, time travel, flying fairies and unicorns

  • Gerardo Moscatelli

    In the Teleportation slide where the output is psi after the operator X or Z is applied there is a very important piece of information missing. Bob needs to know what operator is valid to have the correct sequence. Thus there is NO TELEPORTATION

  • Linda Mull

    0 & 1, meaning the symbolism used representing female and male principle. How dare you! The outer "glow" surrounding the earth is been as if a lockdown surrounding the earth. Fools. 1619-2019 🙂 Yeah. Live with the knowledge that your mathematical equations have always been about minimizing what is universally available to us all. Good luck in your endeavors. Your time is short! 🙂

  • Micheal Bee

    If I write a SQL query that does the filter before the join, I can get a tremendous speedup in processing. Isn't the pre-processing you do for Deutsch-Jozsa the same sort of thing? How is this considered a fundemental accelleration of processing?

  • Dan Deeteeyem

    Holy sh*t.. The audience is not buying his weasel words – they are obviously not afraid to ask the most obvious question. If I have an "entangled coin" and split it, and separate the two halves by great distances.. Why is it so amazing that measuring one half reveals "tails"? Wow! faster than the speed of light the other half is "heads"? Seriously? How much money are they scamming for this snake oil? Lol

  • Oke Uwechue

    Thanks for the vid. A lot of good info here.

    If anyone wants more detail on exactly how the teleportation process works, check out this other vid:
    https://www.youtube.com/watch?v=3HiYm9SB-xU

  • Alain Proviste

    1:05:54 If it's not collapse, measurement will make it collapse. If it's already collapsed, measurement will make known the already collapsed state.

    Communication ~= transport the qbit
    Coordination ~= collapse at both end

  • Diogenes TheDog

    21:10 Raises hand yeah i'll be back after i take like a million pre-req courses….

    i barely understand heaviside step functions and am a psych major lololol

  • Al Ry

    Hi. Could you explain?
    45:58
    We have BB, which realises identity function. But also we have "No-cloning theorem". Is it posible to construct that BB in physical world?

  • Roger Froud

    At 40:23, the real confusion is labelling an Input as Output which makes no sense. I can see that it's related to the Output' in the same way that Input is related to Input', but someone really needs to think of a better way to name these things so it's not so confusing.

  • Data Intensive

    Great Video! I have looked at various articles and videos online which described pieces (generally confusing) of the quantum computing puzzle. This brought it all together and I now feel I have a solid basis to go further.

  • AnteConfig

    1:11:56 if Alice has to measure A and T then send the results of those measurements to B so that bob can collapse the system forcing teleportation, why couldn't Alice just collapse the system from her end?

    As far as I know the decoherence can be initiated from either alice's or bob's particle. Why Should bob be the one to collapse the wave function if Alice has everything needed to measure and collapse the system to enable teleportation?

  • DD

    Any videos or links about more stuff like this, please? The video is extremely simple made and clear. Very well done, thanks for that!

  • Merijn Vogel

    This is the best explantation / introduction of QC I've seen, and indeed very well targeted at computer scientists. Thanks for that.

Leave a Reply

Your email address will not be published. Required fields are marked *