Five Quantum Computing Misconceptions
Articles,  Blog

Five Quantum Computing Misconceptions


Quantum computing is getting quite a lot of
coverage these days in the media, which I think is fantastic, but when I read some of
the things, there are a few things that are a bit wrong. And I’m not really criticising
anyone, quantum computing is very hard. But I think there is some room for some more nuance,
so here are my top five clarifications about quantum computing. A standard description of quantum computing
normally goes something like this. Quantum computers are made of qubits which can be
in a state of 0 and 1 at the same time. All of the qubits are then entangled so they are
all treated as one object that is in many different states at the same time. And so
a quantum computer is like an infinitely parallel computer. Now that’s not quite right. So it’s true
that quantum computers are in multiple states at the same time, before you measure them,
but as soon as you measure them, you only get one state back. Parallel computing deals
with breaking down large problems in small chunks which can each be solved on a processor
before the whole thing is recombined. But the architecture is very different to quantum
computers in a parallel computer you’ve got many different independent processes running
at the same time rather than an entangled state. And also you can read out any of the
states of any of the processors at any time you like. Number two. Imagine this situation, you’ve
got someone who has developed a quantum algorithm that needs a hundred qubits in order to run,
and you’ve got another person who says I’ve built this quantum computer and it’s got
a hundred qubits in it. So any normal person would say, okay great you can just take that
algorithm and put it on that machine and get an answer. But in most cases that’s really
not true. And it’s because they are talking about two different kinds of qubits. In the algorithms sense what they are talking
about is a theoretical ideal qubit which is commonly referred to as a logical qubit, whereas
in a quantum computer sense they are talking about a real physical device, which could
be like a loop of metal, an atom, an electron, a photon, something like that. And these physical
qubits have noise in them because they are real devices living in the real world, and
noise is a killer for quantum algorithms, it will destroy your quantum state. So people are all trying to make qubits with
as low a noise as possible but the quality of your qubit matters a huge amount, to how
good your quantum computer is. Now, theoretically there’s a way of getting around noise in
qubits using a technique called quantum error correction which is where you use multiple
physical qubits to simulate one logical qubit. And how many physical qubits you actually
need to simulate a logical qubit depends on how good quality those physical qubits are.
The higher quality they are the fewer you need, but estimates range from say ten thousand
physical qubits to a hundred physical qubits per logical qubit. That brings me on to point three, quantum
computers and internet encryption. So a lot of the internet encryption is based on the
fact that it’s really hard to factor large numbers on a classical computer. And there
is this quantum algorithm called Shor’s algorithm that can theoretically factor these
large numbers exponentially faster than the best classical algorithm. The trouble is for a standard 128 bit encryption
you’d need about a thousand qubits, for Shor’s algorithm to run. And that means
you’d need a million or more physical qubits in order to run it. And that’s a really
massive number of qubits. Currently we have, the best is 72 qubits that google has in the
universal quantum computing scheme and so it is going to be a very long time until hit
a million. So for the time being, your internet secrets are safe. Point number four is the fact that, nobody
knows for sure that quantum computing will ever work at scale. Some people argue that
noise is such a significant issue that it will be impossible to get say a million qubits
all working together, without noise coming along and ruining everything. And it’s a
fair point. Personally I’m optimistic. I’ve seen that
there are. Human ingenuity can go a very long way. Some people said that we would never
detect gravitational waves on Earth because noise is a too significant an issue for that,
and yet the people at LIGO did and absolute incredible job over many years and now we
can detect gravitational waves on Earth, which is an amazing achievement. It’s not exactly
equivalent to quantum computing but, you know, the only way we can find out if we can do
it, is by trying to do it. And like I say, I’m optimistic, but it is always worth bearing
mind that it’s not guaranteed. And the final point I want to make is about
quantum supremacy. Now quantum supremacy is a bad name because it poorly describes the
thing that it’s trying to describe. Now, I’ve done a whole video on quantum supremacy
so check that out if you want to find more details about it. But basically it is the
moment in time where are quantum computer can do one thing better than the best classical
supercomputers can do. Now currently, classical computers can do everything a quantum computer
can do, and more. But if you think about it, quantum supremacy makes it sound like it’s
the time when a quantum computer can do everything a classical computer can do, and more. But
that’s really not what it is. It’s a time when a quantum computer can do one tiny little
thing better than a classical computer so it shouldn’t really be called quantum supremacy,
it should be called something more like, a quantum glimmer of hope, or a lot of people
are calling it a quantum advantage, which I think is probably a better name. Well I hope that cleared up a few things,
gave you a bit more subtlety into what quantum computers can and can’t do Personally I’m really excited about where
quantum computers are going to go in the future. I think the potential applications of it are
massive especially in the realm of quantum simulation because that’s something that
we really struggle to do on classical computers because, quantum, quantum systems are so hard
to simulate. A quantum computer can do it a lot better, theoretically, and that could
be revolutionary in say like simulating materials, to find out what properties different materials
have with, strength, durability. Or, exotic things like high temperature superconductors.
Also in chemistry if you could simulate how molecules interact, say, a molecule in a drug,
how that interacts with many different biomolecules in your body, that would be an amazing thing
to do, and something we can’t do right now. So, I’m very excited to see what happens
in the future and, and I’ll keep you updated here. Thanks again to the sponsor of this video
brilliant.org. What they’ve done is taken proper maths and science and put it in this
framework where it pretty much feels like you are solving puzzles but you are solving
real problems from maths and science and learning real STEM skills, and, when I’m looking
for a brainteaser, I like doing their problems of the week, and the problems range from being
very approachable through to some stuff that is legitimately difficult. They also have
courses on physics, mathematics, computer science and they are adding more content all
the time. And the thing I love is when you are doing a hard problem and you finally see
how to solve it, you crack the problem, and you get it right and you have a moment of
feeling like, yes! I’m not a complete idiot. And that’s what I enjoy. So if that sounds
interesting go to brilliant.org/dos and I’ve also put a like in the description below.
Thanks for watching my video and I’ll see on the next one.

100 Comments

  • Akmon Ra

    This really helped clear things up. I should really get information from scientists rather than news sources, since I had been confused about what the real challenges were with quantum computing and also believed at any random time we could achieve the ability to break every encryption on the planet.

  • Heather Rainbowlady

    My legacy programmer's brain can't seem to get past binary. Every time I try to consider quantum physics and Qubits my brain gets stuck in an 'IF' loop and crashes out into a default routine involving the kind of cubits that Noah used. 🦄🙃

  • xeni9000

    Maybe I missed this part in the video or didn’t get it but why exactly is the noise the metal makes an issue?

  • Chris Castaneda

    can you explain how simulating a chemical reaction using a quantum computer would work? I cant seem to wrap my head around it? Thank you

  • hurktang

    Quantum computers are not "noisy" they are inherently random. From what i heard so far about how it really work, i'm affraid it won't have much application. Unless we can do Qbit cheaper then actual bits of course.

  • KappaW

    quantum computers are just a bunch of scientist trying desperately to find a use for their Ph.D. in quantum physics. Was never destined or intended to be practical, it's more of an experiment than a product. An experiment that has proven to fall significantly short of the promises made, they will never and can never compete with traditional computers. Its absurd to even compare the two.

  • Cargill Seiveright

    I believe it can get better… Transistors shrunk from the size we were able to see to the size smaller than a virus…

  • Your friendly neighborhood prog snob

    This is one of those things where I don’t understand a thing he’s saying but I just enjoy sitting here, blissfully staring blankly at my screen.

  • Michele Pellegrino

    May I point out something?
    1) The feature enabling "infinite parallelism" is SUPERPOSITION, not ENTANGLEMENT; the latter is still crucial to q-computing, but not in that stage;
    2) Logical vs physical qbits: nice! Someone are finally pointing this out!!!
    3) Shor algorithm isn't really for FACTORING, but for PERIOD-FINDING… the latter entails the former, in some way…
    4) The problem is that too many people are optimistic about q-computing and there's to few criticism, even if both the premises and the potentiality of this new technology are… dubious?
    5) In any case, q-computers won't be able to work without classical computers; so, is there any reason to use a q-computer to do those tasks that can be efficiently performed by a classical one?

  • mongke

    Can you do a video on quantum physics misconceptions?

    Specifically the observer effect and it's non-relationship to consciousness. There's a cult called Quanto Life surrounding this misunderstanding.

    Thanks!

  • call me zagh

    Hey, thank you for the video but can you please tell me: Is it better to have a masters in computational physics or computer science to get into quantum computing research?

  • poop catapult

    Not sure if you know, but last month a paper was published in which the authors showed that BQP is not a subset of PH. So there is a computational class of problems that only a quantum computer is able to solve efficiently. I think this is a huge step.
    [edit: messed up, the class names]

  • Joshua

    For point number 2: isn't the logical qubit all that matters? Why would ppl state something else then the real, usable performance?

  • Craig Gidney

    Minor mistake in the video: you say and show "128 bit RSA key", but RSA key sizes are actually more like 4096 bits. 128 bits is not nearly enough for security in the context of RSA. For example, in the RSA factoring challenge, the 129 digit (not even bit, digit!) challenge number was factored back in 1994! ( https://en.wikipedia.org/wiki/RSA_Factoring_Challenge )

  • Igor Henrique

    Your explanation actually helped me understand better this subject, you put it so simple with a very enthusiastic tone, thanks for the video, a suggestion to another "map video" is biotechnology 🙂

  • Korakys

    That was a very clarifying video, thanks.
    I can see why they call it supremacy though, the feeling that must come from finally making a QC that can do something better than a classical computer would be akin to supreme satisfaction.

  • Mahou Maou

    Great now explain to me what are "grover operator" ?

    6:31 – 7:03
    Quantum Computing theoretically revolutionary simulating materials, to find out what properties different materials have with strength, durability. or, exotic things like high temperature superconductors
    also in chemistry if you could simulate how molecules interact, say, a molecule in a drug, how that interacts with many different biomolecules in your body
    ????
    ^^^^^^^
    Please Research and read this here
    ^^^^^^^^
    scans technology

    https://en.wikipedia.org/wiki/3D_scanner#Technology

    https://en.wikipedia.org/wiki/Image_scanner

    https://en.wikipedia.org/wiki/Intelligent_lighting

    https://en.wikipedia.org/wiki/Laser_scanning

    detect and scanning light ball

    sensor scanner wiki

    https://en.wikipedia.org/wiki/Image_sensor

    https://en.wikipedia.org/wiki/Sensor

    https://en.wikipedia.org/wiki/Structured-light_3D_scanner

    https://en.wikipedia.org/wiki/X-ray

    ______________________________

    yes this technology already exist you just need to combine it with computer and WowLa

  • Augusto Napuri

    Sería un mejor video si dijeran lo bueno de la computación cuántica y no las cosas que se creen "imposibles" a llegar a pasar.

  • Auxilium

    so you can't listen to music on a quantum computer… there is just too much noise
    okay, that was just… ehm .. an example of a joke in a quantum state.. it is both funny and not.
    thanks for the video!

  • Zenytram Searom

    So you need 125kquByte to break a 16byte encryption, so unless it reaches the range of Tera quBytes, our credit card system and internet dicpic are "safe."

  • Xizheng Ma

    Please google "quantum error correction" before giving numbers about how many physical qubits you need to achieve a logical qubit… Otherwise, nice video!

  • Thibaut Hanson

    Looking forward to quantum cards/coprocessors, like those floating-point units coprocessors in the early days. Not convinced about full-on quantum for the whole computer. At least in the foreseeable future.

  • Iago Silva

    6:04 Call it a quantum edge – sounds good enough to be a video game title, at the least (on the matter of bad names, what are your thoughts on "quantum teleportation"? I played for some time with "quantum imprinting" as an alternative, but IDK)

  • Semyon Galtsev

    Great video, Dominic! The "movie night" reel popped up in my suggestions, and after it and this one, you got a new enthusiastic sub in me.

  • Viv B

    A very rudimentary processor like a current console (pro one x) run extremely hot! 75 Celsius on the surface. How will quantum computers handle heat?

  • Empower

    Please do a video, which is about laboratories, observatories, for example:

    Very Large Telescope
    Los Alamos National Laboratory
    CERN (mention types of accelerators)
    LIGO
    ITER
    VIRGO
    National Ignition Laboratory
    Jet Propulsion Laboratory
    Fermilab
    etc.

  • Robert *

    Quantum coding is even harder 😀 (its a joke.)
    About quantum Supremacy… in the sense you explain it its true… but in another sense it is not true… that one thing it is really good at… x^2 is pretty much a whole new language we need to learn to understand and code upon. if we could write programs under those rules and they would work in a state in which we could understand then quantum computing would be the exact opposite of what your are describing. yes we are still millions of years away from that but i think you know what i mean.
    and the noise… that is just a matter of technology evolving… at some point in time someone will find a solution for it. and then a 100 qubits will become a true 100 qubits.
    now its just error correcting itself which is something we should see a temporary thing.

    IF we have a "True" quantum working system that science could use to run simulations on it would be used 99% of the time for Bio-engineering.

  • Josh Duriden

    It will take about 15 years to reach a million physical qubits if we double the amount of physical qubits we can make each year.

  • Ivy-ron Quinto

    I am a lawyer and i don't know why i am here, considering i detest math consequently computing. But i like listening to things i know nothing of and learning about it. P.s. your beard and accent helped me to watch the entire video 😁

  • SkyCharger001

    #2: there is also the question of algorithm translation: sure the algorithm would need only 100 logical qubits if it were run on an ideal quantum computer, but it's more likely that the quantum computer will need more to deal with the translation overhead.

  • Fresh Green Jello

    Oh wow, you actually explained it such a way that an average Joe (like me) can understand it. Thanks for the insight!

  • Bluemilk92

    You should check out "PBS Space Time." They are one of the few channels I've seen that doesn't patronize their audience. I'm a college dropout, and only barely graduated high school, but their series on String Theory was certainly comprehensible.

  • crazieeez

    3:50 is wrong. Topological quantum computer can take 1M physical qubits down to 4,000 physical qubits to create 1,000 logical qubits.

  • MrMastadox

    Good video, the media always reports on quantum computing in a flawed manner. But i gues they have to be overly positive about it. Because quantum computing might face an investment drought before they ever become functional. I'm not as positive about the quantum computer's future. But time will tell.

  • Guillaume Muller

    Hi,

    I'm starting to study QC. Your clarifications are very good. Nice work. Thanks a lot.
    Would be great to find the drawing presented in this video in your Flikr stream!
    Have a nice day

Leave a Reply

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