That's the first time I've 'got' how quantum computers work. -'it's not the speed of the individual operations, it's the number of operations it takes to reach the result.' No magic. Andrea is a superb scientist. I love his explanations. Thank you!
Well, even though some explanation is done, I still don't understand what an operation is and how it affects a qubit's state. And how operations mutualy use those states. And how one programs a set of operations anyways. If I'm not mistaken, a qubit only holds the intermediate states between the beginning and ending of a sequence of operations. And after you read it the state is gone. I'd interpret this as operations can't read qubits either but just influence it. Sounds a bit like analog values getting amplified and dampened until the outcome is reached. Anyways, if you understand it better then please let my know
@@mmehdi3437 you want him to explain whole quantum physics in a youtube video? He talks very understandably and he's probably a great teacher. he knows what he's talking about.
If you're talking about the long hair dude on the whiteboard then you smoking rock... He didn't make a lick of sense to me. I've watched it twice now and have no clue how it works nor how it could be better than a standard computer.
It's just really fast computing, the quantum state stuff is just on and off just faster"slower" (think of space time slowing down) like a dimming switch on a light. An easier computing way would be to have a computer register 5 and 0 as on and off 1-4 and 6-9 as quantum numbers (dimensional movement) no such thing as quantum it's just really fast. The photon pairing example can be just seen as gluons and bosons(slower)
Where he was describing N qubits? All N is is a variable that is representing how many of something there are for instance if you might own 2 cars and someone else owns 3 cars you can describe both situations with both people owning Ncars where in your case N = 2 and for the other person N = 3. A coefficient is just a variable number placed before something that multiplies it. So what he was saying is if you have 1 qubit it can be in both positions at the same time called a superposition. where as classic computers can only be in one position at a time. As you add more qubits the positions can be multiplied exponentually. So he was saying if they have 300 qubits which means N = 300 and that 300 qubits can produce 2^N or 2^300 possible positions in comparison to classical computations.
Yes, but how does the computer *actually* work? How does it store a qubit? How does it entangle the qubits? How does it read the basis state? How is he algorithm strucured?
almost 8 years since the video was made and i still think this is one of the best explanations i can find for my tiny brain to comprehend the basic definition of quantum computing.
I have what scientists call QI or Quantum Intelligence, which exists in two states, before you measure it, I am both Intelligent and not, but once you measure it, what you find is that I am not intelligent 99% of the time. By the way, I don't even understand this enough to make an appropriate metaphor, and this is the 1% trying it's hardest here.
Finally someone actually explained how they work! I’ve know for years about how, “it’s nothing like a classical computer, it can have bits in superposition” but no one told me how to read the qbits or how they interact.
For those wandering, why when you add up the coefficients in front of the possible states of the electron spin at 1:35 you don’t find one (which would mean he made an error in the probabilities) , it’s because these numbers are just coefficients, if you want to get the actual probability of getting an electron spin up/down, you have to square the coefficients: (0,80)^2 = 0,64 (0,60)^2 = 0,36 And when you sum up these numbers you get one.
@@johncarson5436 From memory it is because of how the probabilities are calculated. Determining the probability is a function of the electron charge and its momentum. I realise that probably means nothing. Its a super abstract concept and is only expressed by the underlying maths.
@@johncarson5436 if both coefficients are 0,50, than you must have made an error: 0,5 squared is 0,25. There are only 2 possible states for that electron: spin up of spin down, which means that your probabilities don’t add up to 1 but to 0,5 (which means there are other possible states) If the probability coefficients are both 0,5 (and not just the simple coefficients) you have a 50% chance of observing that electron spin up, and 50% chance of observing it spin down. PS: if the probability coefficients are both 0.5, that means the coefficients are equal to square root of 0.5, so that if you do the math backwards: sqrt(0.5)^2 = 0.5 :) I emphasize on not confusing the simple coefficients and the probability coefficients
I remember my physics professor at Georgia Tech was building a quantum computer back in the early 2000’s, but it was probably a 2^1 or 2*2 Qbits at that point, 😂. Smarter than I’ll ever be!
I guess the future is hybrid CPU's. Just like we use specialized CPUs called GPUs to render images, we will use QPUs (?) to calculate something which benifits from it, while still using CPUs for all the other cases.
to better wet expand on your concept i belive the "domestic" application that can benefit the most from these kind of operations are exactly those made by the gpu,that has to process lots and lots of shadows and physics calculations at the same time,or to render a large video,as per say if you encode 2 bits of it at a time or 2000 it doest make a diference,so i belive that we are going to see 2 things from this 1st is quantum based gpu,s and second something faster or even a improvement on ssd,s so that they can keep up with the large amount of data transfers needed to acomodate such power,otherwise your gigantic quantum computer will be as fast as you can transfer data arround...
***** Or perhaps at some point we will hit a bottleneck,either because of the limitations or the possible price tag as an outcome.You can see it happening right now,instead of going bigger in sheer amount of processing power,developers create techniques to do whatever they want/can with the existing tech. Think of it this way: You can stack cards only so high,before they collapse.
Kronguard Price is not a problem. Simple economics say that so long as the demand and supply are high enough, the price will drop. Developpers are creating techniques to harness as much power as possible with existing tech only because of the existence of consoles. Being limited by generations, instead of the continuous evolution of the PC, they have to look for more sophisticated techniques to get as much power out of those machines as possible. This is both a good thing and a bad thing. Creating these techniques frees up memory, but being limited and having to spend money on the technical aspect of the game instead of the gameplay and to some degree graphics is not that good.
You're not considering the nature of GPU's and the way they process information. Cpu's work in serial; each line of code is taken in sequence, one by one. Really fast, but still, one by one, and has a few cores dedicated to that. Gpu's work in "parallel"; they work with floating point numbers, and can take several lines of code and process them at the same time, it has thousands upon thousands of specialized cores. The better an application is optimized for GPU-acceleration, the more "room" is given to a CPU for other processes, and thus the bottleneck someone mentioned can be taken care of. Eventually yes, a CPU bottleneck will occur because transistors will eventually reach the limit in manufacturing process, and so will GPU's, but we still have a few years for that, and there's a big chance that by then there will be a different material that will allow for a smaller manufacturing process of the transistors, which will account for that.
+-=[Kuledude Gaming]=- You have to understand the properties of 2^n. Think of it this way. Put a grain of rice on the first square of the chest board. Double it it for the next square, adding 2 grains of rice, 4 for the second, 8 for the third and 16 for the forth and so on. When you reach the last square you will have (2^64)-1 or 18,446,744,073,709,551,615 grains of rice, a pile of rice the size of Mount Everest and 1000 years of the entire world's rice production at 2010 levels. So, if you manage to build a quantum computer with n bits, and manage to keep those n bits entangled, it can represent 2^n states. So 4 for n=2, 256 for n=8 and 9 trillion at n=64. So in other words you can do much more for less. But only for certain (important) types of problems, because you need to be able to collapse all these states out to one you can actually read.
The minute I saw him in the video, I felt like this dude has some really good energy. Just someone you would immediately trust! A genuine but also extremely smart person! People like that are very rare.
don't pretend you understood it... kidding, i just don't get it. i understand the computation power reference, but i don't get what superposition means and how to use it
Guys, this is really a great explanation of Qubits and the best one I've seen so far. I suggest that you guys look at some videos on the basic properties/phenomenons of Quantum particles/physics before diving into Quantum Computers.
Just loved that he did explain in the technical but understandable terms, and as a highschool physical chemist student, I understand all this very well. Also love the nutsell telling that it just won't be improving our classical computing like browsing web or stuff
That long haired guy is an example of a great communicator; I think it's very easy for similarly intelligent guys to spew tech talk that passes over many heads
That's raci...!! Kidding. By the accent; sounds very romance (as in romance languages), the way he speaks, the hard pronunciation of consonants, the intonation, the way he constructs his sentences and some words he uses. Also the name is a big hint, but you as an Indian probably have some of these features in your speech as well. There are Europeans who look noticeably different (i.e. Spaniards-Norwegians) but in some places they are really different to tell apart, even from Americans, and that's because it's a society built mainly by European immigrants that mixed maybe as much as in Europe itself.
@@kangkanlahkar9045 as an italian, I knew he was italian before reading his name. He just speaks in the same way most italians do while speaking english at school. I don't know the specific features that make me say so, he just sounds italian. That's the way we speak I guess
He does a great job at explaining it. Picks his words with great care. Now I want to know which types of calculations would benefit from quantum computing.
+yrjosmiel73 As far as I know, no truly working quantum computer exists right now, so probably not. But then again, I won't know until I've searched for it. It is currently in a state of superposition.
+yrjosmiel73 nope . Crysis wasn't made for quantum computers and like the guy said it may be slower cuz it would use power just to convert to traditional code. Maybe when quantum PCs become more relevant can the cry engine use it's advantages.
What I'm still having trouble understanding is how can we get useful results out of a machine whose state is based in probability and chance. I've been looking up answers all day, but very few make sense to me.
Pretty much they are accessing the a wave function superposition meaning every possible answer you can think of and they are also with holding information on how powerful this machine really is .this machine is precognitive and it's been around and in use longer than they say it's accessing information from every source of possibilities that exists . Oh and it does so much more .
To summarize what they said in the video, the computer starts with all possible states, putting all possible states through your algorithm as the variables and simultaneously calculating all possible results, and then you read the result, but by reading it, only one result is created, influenced by probability, because you can only read 1 or 0 and not the probabilities of superpositions. For example, lets take a formula like a = (x && !y) || z. Lets say, for simplicity, we assign just one bit to each number on the right hand side, so x, y, and z equal either 0 or 1. If you calculated with a quantum computer it would calculate everything at once. x | y | z | =a 0 | 0 | 0 | =0 1 | 0 | 0 | =1 0 | 1 | 0 | =0 1 | 1 | 0 | =0 0 | 0 | 1 | =1 1 | 0 | 1 | =1 0 | 1 | 1 | =1 1 | 1 | 1 | =1 I'd assume that every bit starts at 50% chance of being 0 or 1, so a would have a 62.5% chance of being 1. I'm not sure, but from what they were saying I think they may have some kind of method of checking each line of the table one at a time. I'm still not sure myself, it's a confusing field and I haven't yet found any absolutely complete explanation. Basically, this could be used for powerful algorithms that consider every possibility at once. For example, a physics engine that calculates everything about a particle with every possible starting condition and then applies those calculations to each individual particle within the simulation at each tick of the clock; it would be like having parallel processors for each of the thousands or millions of particles in your simulation. This could be used for everything from gaming to theoretical physics simulations.
A basic example would be trying to obtain cryptographic secret key that matches known public key (this is mathematically possible but requires billions of years on classic computers). So you set the quantum computer so that all results will initially have the same probability and then you check it against the secret key with carefully designed algorithm. All possibilities will be evaluated at the same time and the correct result will increase its probability. Read the qubits and repeat the calculation a few hundered/thousand/million times. Then check the values on classic computer in order of how often they appear and you will find the result probably in a few seconds.
I have more simpler analogous example for everyone to understand here. Mr X can understand 1 question from one of the 3 people throwing questions at him, at a time. So he will take 3 turns to understands all 3 ppl 1 by 1. But here we have the genius Mr . Y. He has the special ability to absorb all 3 questions at the same time. Why wait 1 by 1 when you could take all at once ?? !! Awesome isnt it?? Now 2 Mr X (2 bits) will understand 2 questions at a time. But 2 Mr Y (2 qubits) will understand 2^2^6=256 questions at a time !! Because they can superimpose their input ability in exponential!! Keep adding Mr. Y (the qubit) and you get the capability to absorb billions of times larger amounts of questions in 1 go. Freaking awesome !! Done. ------ Now talk about why they so large and ultra expensive?? Quantum computer basically lifts the limits of hardware construction. Current cpu design is such that it can only be some mm thick. But quantum computer breaks this limit and goes all way up in hardware, thus you se those huge quantum CPUs. More simple: Like intel or Amd cannot make faster cpu by increasing cpu size, it's counter productive and that cpu will not work. Like your brain cannot be the size of an elephant to make it more powerful, it will err and die. But quantum cpu can be made as bigger as you want and it's processing power will keep multiplying exponentially!!! Thus it breaks the limits of current microarchitecture and utilizes full hardware possibilities in all directions. But the problem of cost and power consumption will remain there...it will all depend how much semiconductor technology advance, allowing some smaller quantum computers still millions of times faster reach to the public. Will take some decades though. Until then, we will only see them installed in big tech firms.
so basically you need (n * precision)^2 normal bits to determine a single qbit. For example with regular 32 bits floating point variables you'll need 128 bits to determine 2 qbits.
Its interesting that he says they can be both 0 and 1 at the same time. I'd venture to say that, they are not 0 and 1, at the same time. Instead, they are interchanging, just at such a rapid rate it is perceived as being at the same time.
@Mario A quantum bit can be used to count faster than a normal bit because a normal bit can only go on or off but a quantum bit can go on off or half off so it is just better for some things.
listen the sentences carefully and the moment you realize you're loosing the track start again from where you remember the last thing you fully understand!
So, basically, it's good for making huge combinations in a short time period, but not for transforming information, because the position of each electron has to be measured every time. So it's good for storing data and data analysis, as I understand. Think of a grid of electrons that can go up and down, instead of the classic bit, where the electrons have to go back and forth.
@@hamedkadkhodaie7715 So, what they're doing is basically trying to design more efficient heuristics by exploiting quantum properties. Unlike classical algorithm design, which is mainly concerned with reducing the asymptotic complexity of the problem itself.
Cool. I always thought Quantum Computers are the perfect thing to replace traditional silicon-based computers. Turns out I was wrong. Thanks a lot for the very helpful information!
So quantum computers aren't fast due to their individual operations being faster, but due to the fact that they require far fewer operations. Very insightful. Imagine a classical program that outputs you the quickest route to get from your house to the store. It must analyze every possible path there in order to compare and determine which path is shortest. It does this incredibly quick, however it requires enormous amounts of operations to give you the answer. A quantum computer, on the other hand, can process every possible path simultaneously, therefore requiring far fewer operations in the first place.
That's the right idea, but there's a catch. While the quantum computer can process each path simultaneously, only the information about one path can be retrieved. When the state of the qubits is measured, it collapses the superposition into a regular, non-superposition state. That is, all the information about the other states vanishes and it is essentially as if you had only analyzed one path. There are still ways to gain useful information out of the superposition (depending on the problem), they're just not always obvious and the information might not necessarily be useful.
@Trius IBM has a framework called Qiskit that you can use to write quantum computer software and you can run your code on their quantum computer connected to the cloud too. Here is the link and they also have video series on UA-cam explaining this. qiskit.org/
@@hatemel-kharashy8856 I think in the future computers would have 3 processor types: CPU for the main processing, GPU for intensely parallel processing, and a quantum processing unit for all this quantum stuff.
I just understand that there will be excessively combination per second. so this can be used for password cracking? is that right? because electrons move insanely per second which means that we can't know where an electron is but they can make 4 meaningful letters. The electron move randomly and insanely so this is the case. Per a moment a 1 letter will be observed and it will give signal. But per second, there will be nearly endless moment because electrons move so fast. There will come nearly endless and random signals per second because of the movement of the electron. The electrons move so fast that we can't even determine where it exactly is. So, it is not matter how long your password is, it will be detected very very very quickly. They can define the letters, numbers and characters to that system and easily solve your password. Because all of the possibilities will be just numbers near the infinity. Number divided by infinity equals zero. Actually maybe the number of letters (4 electron positions) are not the case. The number of the moments and randomness are the cases. It might be just 2 positions too. They are just making use of the randomness and speed of the electrons andthey achieve endless number of combinations. If I understand it correctly.
Anyone else notice that the numbers did not add up to 1 but the percentages added up to 100%. he messed up on his probability, the numbers were 0.80 and 0.60 which add up to 1.40, while the percentages were 64% and 36% which add up to 100%, he did the percentage math correctly, but did not do the probability part correctly.
0.8 * 0.8 is 0.64. 0.6 * 0.6 is .36. 0.64+0.36 = 1 those values were not the probabilities but, think of them as a weighting coefficient. when you multiply by the complex conjugate you get the probability. since these numbers do not contain i, the complex conjugate is itself.
TheBscit It's called "i" learn it in Adv. Algebra II. Imaginary numbers, they were a bitch. If that "i" refers to something from Calculus then I have no idea what he's talking about.
They're called Born probabilities. Oversimplifying: The probability of measuring a certain outcome is equal to the squared modulus of the given "weight" (or amplitude). Aaron Miller is correct.
I understand that quantum computers are using probability instead of solidly well defined positions which allows quantum computers to increase computing power exponentially instead of multiplicative. Since Moore's law is ending instead of creating quantum computers why don't we use variable voltages in transistors as as more states then just on and off 1 and zero. Like if you know anything about transistors there on state is within some range of voltage and the off state is also within some range of voltage called the threshold voltage. Well why don't we make computers slightly more precise so that there are maybe 4 voltage ranges instead of just two. Use a 4 number system instead of binary this would make computers far more powerful and wouldn't require a complete materials or computing revolution.
This is more along the lines of what I was thinking. But I really don't have much in depth experience with components so I didn't know if something like this would actually be possible.
You realy don't understand the architecture of a regular CPU....Hi works only beacuse trasistors understand two states(power ON and power OFF) there is no alternative just YES or NO. To do this, what you think, you need to develop some new kind of transistor and that wouldn't be a transistor anymore.
Actually transistors do act like a variable resistor between certain ranges, he's right. The issue is they do not act reliably in that regime. With HIGH versus LOW, there can be a range under which a transistor is closed or open. In that grey area where the metaphorical switch is not fully open or closed, things get very uncertain and complicated. It would be incredibly difficult to utilize that property of transistors for many, many reasons. Manufacturing uncertainty would be enough to throw off any individual transistor in a set so that it doesn't read the same value as its neighbor under the same applied voltage to the gate.
I believe you would need a larger voltage since there is always a small threshold above 2 different voltage to represent 1s or a 0s so if we are going to add more states there has to be more voltages to fill the gaps and then we will need transistors which are able know 3 or more states and that's where it all fails. What i really want is a light based mother board. But to make this any good we would need extremely fast conversion between electricity and light which i don't think will ever be possible.
There is no reason why we couldn't, it's just that the payoff isn't anywhere near as high. Both ideas have similar problems - stability and programming - that binary doesn't have because it's as simple as it gets. Using binary is like using triangles in computer graphics. Since triangles can only exist on one plane (in terms of filling in the shape when given the vertices), while other shapes could have ambiguity, they're ideal for precision (getting the same thing every time), but not for accuracy/efficiency (it takes a lot of triangles to approximate a circle). Similarly, binary is always right (in terms of mechanical calculations, floating point is irrelevant), but it's relatively limited. It would be possible to compute in base 4, and we could do so with essentially the same architecture that we currently use, but it would require far more accurate and expensive technology that would still be incredibly finicky (say, if you wanted to overclock it). Secondly, we'd need to create a new operating/logic system to account for the differing outputs, as any conversion would waste the potential. Based of what was said in the video, switching to base 4 would increase efficiency, as the gradience of each output would need its own measurement - more information - but that number would strictly double (currently when computing, each place value is recorded in a separate bus lane, so the effect only applies to each output). While these hurdles are even larger in terms of quantum computing, it's still just a one-off, and the benefits increase exponentially as opposed to multiplicatively (by that, I mean that the equation for the information density of a q-bit is 2^2^n, as opposed to just x^n. Even if we were to move into hexadecimal, it would only take 5 bits for the quantum computer to win out). However, (as far as I can tell) the benefit of voltage gradience is density as opposed to speed (it would only really improve memory bus width, but physical bottlenecks would still render that pointless, if not, worse than they already are), and one one the major problems/bottlenecks in quantum computing is the need to convert it back to binary, so if this were to be mastered, the quantum equation could be rewritten to 2^x^n. (Again, this part is outside of my knowledge) Of course, this is just an educated guess based off my understanding of how classical and quantum computers work in theory, as well as a brain dump to organize my thoughts on the matter. I could be completely wrong, and I'd love to be informed about that last part.
Lol! I have now! But no i do not think that is myself. I have had a natural interest in physics since i was 15 and was reading books such as goldie locks theory which actually introduced me to the multiverse theory, string theory and thus quantum mechanics which now incorporates string theory to make the multiverse theory possible.
Well, I never expected it to be like this. I got the idea of quantum computing (last year - 2020) even before I came across this vid. My basic concept was to make components smaller than what it is today. So small that it would use sub-atomic particles to process information, like some sort of hyper computer. Now I came to know that these computers actually exist and are "not" suitable for personal use.
OHH SO THATS WHAT MY PROFESSOR MEANT WHEN HE SAID MAGNETIC SPIN what theheck you explain it so simply my prof didn't bother telling us what magnetic spin was exactly
He probably didn't even fully understand it, so he couln't explain it. As Einstein said: If you can't explain it simply, you don't understand it well enough.
+s vashi Its not quite that simple, if you watch the explanation of spin with this same guy on the secondary channel you'll see what I mean. The 'spin' of an electron (and in fact every other particle) is a purely quantum property that has nothing to do with magnetism in the fundamental sense. So while the electron's spin has a magnetic interaction, the neutron's spin, for example, does not. So your term 'magnetic spin' is a misnomer. Hope this helps.
@@qiwi111 no, it's not this simple. This video makes it look like actual rotation to align with a magnetic field. Electron spin is a very unfortunate name. It is a purely quantum effect and does not have a classical equivalent. It's not a tiny solar system. For example I read a paper somewhere that said that a 360 degree rotation of a magnetic field was not enough to return the state of an electron to its original state. It took another 360 degrees. So 720 degrees rotation to return to its original state. But I bet I'm oversimplifying it too, I'm still learning about it and I have a degree in physics. Your professor probably didn't want to lose the whole class and have everyone throwing their desks and walking out. Read more about this topic it gets crazy.
This guy made me realize that my top-of-the-class awards in first to fourth grade didn't really reflect my intellectual capacity. The more he explained it the farther I got from grasping it.
the summary: the quantum computation - for now- can be used only for huge amount of data processing but can not be translated into our classical computational systems that means we need an inter-translation system to connect between both worlds I believe that what are we going to see in the near future
A quantum computer can do everything a classical computer can do, just much slower and using much more energy. And, somewhat ironically, to program a QC and read out its results, you need a classical computer anyway. (Well, not really, but it does make things easier.) The hope is that the qubits can hold superpositions of bit vectors, and resolve them into defined results. This would be done via quantum gates, which are logic gates for quantum states. And because a qubit can hold a superposition which is only stochastically defined until it is measured, you don't need to backtrack through all possible permutations to find a consistent configuration. At the end it can only hold a state that the computation allows. And that means it should be able to do some computations over large sets faster than a classical computer ever could. Quantum computers are basically their own class of computational complexity. But until quantum supremacy is confirmed, it might not actually be any faster. In practice you have to deal with decoherence, which happens whenever a quantum particle interacts with anything else. And the more qubits you have, the more it will happen when you don't want it to happen, which is why you also need quantum error correction. You have to do the calculation several times, compare the results, and test them. And it might just be that that erases quantum supremacy. In theory it shouldn't, and if it does, new physics is required. On the other hand, if quantum supremacy is confirmed, new maths will be required. And so far, a few pre-existing fields of mathematics have been sufficient to serve physicists well.
yo mang i've scoured youtube and not a single actual in-depth explanation of why it is able to solve problems faster. namely how is an algorithm designed such that measuring the states yields an actual answer and how do you even know how to interpret that answer. Everyone is just like "well herp derp I got an intro and now I know how they work"
Tiwaking Tiwaking They solve certain SPECIFIC problems like integer factorization faster. Problem is no one has ever come up with a concise yet thorough explanation as to how they achieve this.
***** that's exactly what i mean by bs hand wavy explanation. It doesn't really explain how it happens. And though you may have followed the gist of the explanation, you still really have no idea exactly why an exponential growth in possible superposition states actually translates to faster problem solving. it is obviously more involved than just increasing the amount of working memory.
+Max Loh Its BS and handwavy because its exceptionally complicated. If you want a thorough understanding of how quantum computers work, you aren't going to get it from a 6-7 minute youtube video. You probably need at least a masters degree. This video is VERY simplified because otherwise nobody outside of the field would have any hope of understanding what was going on.
Dan Albrecht Perhaps so. I was hoping for something that might help us actually understand the idea, without getting technical. Because really all they did was explain that the number of possible states stored grows exponentially, not explain how that's even leveraged to solve anything faster, and then some people are claiming that they "understood" the explanation just because they grasp the concept of exponential growth. For example, for "why does diffusion happen" the hand-wavy explanation would be "particles tend to state of disorder", but it's much more informative to explain that the reason this happens is there are so many more disorderly states over time so it's just a matter of probability -- then a layperson can completely understand how this works without delving into any equations. Similar explanations are available for many other physical phenomena and computer science problems. But maybe for quantum computing there is no such middle ground.
+Max Loh Totally agree! I've been casually looking for this explanation myself for years and have yet to come upon it. Namely how does one setup an optimization problem to leverage qbits and how do you then read qbits to get the answer. I don't think the reason is hasn't been explained is because "it's technical" and I really don't want a watered-down non-technical answer. I'm sure that equally complex things have been explained both in video and in literature. For some reason this one has been elusive. I mean no one has even tried to explain it. Weird.
I still don't understand how the randomness of superposition helps in calculation. Y'know what? This is magic. That's the only reasonable explanation. I'm gonna go look for witches now.
“The speed is not in the amount of operations that it can perform but just in the amount of operations it needs to get to the mathematical result”. He couldn’t have been explained better, this is awesome.
@@mikeg4972 There is a famous thought experiment called Schrödinger's cat. In this experiment, a cat is in a box with a radioactive source and a poison. If an atom from the radioactive source decays, a Geiger counter detects the decay, triggers a relay, and releases a poison that kills the cat. If an atom doesn't decay, the cat is alive. Because the situation is dependent on the superposition of the radioactive element, an observer can't really know whether the cat is alive or dead until they look in the box. Therefore, the cat is both alive and dead at the same time. Now, there are lots of interpretations that make this paradox a non-issue (and you can read about them on the wikipedia page) but the point is that you can't know the state of a subatomic particle until it's observed, but the act of observing it affects its state. This makes sense because if you want to measure the position of a particle, as an example, the methods needed to observe it (i.e. a visual observation depends on photons) will move the particle as they interact with it. It's not that a qubit is actually both one and zero at the same time, it's just that we don't know what the state is until it's observed, but observing it will give us the wrong answer because our observation modifies its state. So instead we think about it as a probability space. If we know the probability that a given qubit is 1, and that probability is, say, 85%, and we rerun the algorithm maybe a few dozen times and the mean probability that qubit is 1 is still around 85%, then we can be reasonably certain that the value really is 1, and we never need to look at it.
So, I still don't know how quantum computers work. How does the computer actually use the quantum states to do calculations? Or is it just a metaphor, and it still works with normal electric switches? Because theoretical information is useless, we need ACTUAL information! Really wish someone would explain it to me ....
For as far as I understand. When you have 2 qbits, they are not yet defined as up up, down down, up down, down up, until you read them. So on 2 qbits you can store 4 combinations. The only problem, and that is what is keeping us from having quantum computers, is that we can not yet be sure that we read the state that we want to read.
This might make the explanation a bit easier for some: The classical bit system allows for information to be stored in binary, or base 2. This means for each bit there can only be 2 possible entries, 0 or 1. The qbits allow for 4 possible entries making it a base 4. The decimal system we use every day is base 10 (0-9). This means that more information can be stored in a smaller space. As for how to calculate anything with the data, I still can't get my head around being in both possible states at once. Quantum mechanics is beyond me for now.
I can't tell you exactly HOW these calculations are performed. To do that, I would have to teach you quantum mechanics. This is not that surprising, even teaching you how a classical computer works would require teaching you about logic gates and stuff. But I'll try to answer your questions in a sort of "high-level" pop science way. I'll try to explain how a classical computer works in a language that will make it easier to "get" what a quantum computer does. "How does the computer actually use the quantum states to do calculations?" Every computer is essentially a physical system that realizes some model of a computer. The computer you have on your desk uses classical states to represent zeros and ones -- specifically, it uses electrical currents. If you can distinguish between there being a current and there not being a current, for instance, you have two states that you can label "zero" or "one" as you wish. For definiteness, let's say that "there is a current" is the "one" state. The second step is to apply some operation on these states: in the abstract they're called "logic gates", and they have simple rules like "and", "or", "not", etc: these are realized by a circuit. An "and" gate is a circuit that outputs a current if both of its inputs receive a current. A "not" gate outputs the opposite of its input, and so on. With these basic operations you can build any classical calculation on binary numbers. The third step is to measure the result: let's say your computer isn't very sophisticated and doesn't have anything like a screen. Then you'd take an ammeter to the "outputs" and you measure if there is or isn't a current on each of the output bits. Then you'd know the answer! A quantum computer follows the same basic paradigm: you start with a physical realization of a computer model, but you can't use anything quite as simple as electric currents. You have to use objects that obey the laws of quantum mechanics. The video talked about spins, I think spins are way too complicated to explain to the layman. So let's just say you have a quantum switch that can be "on" or "off", or any superposition of the two. Don't think too hard about what a "superposition" is, in practice it just means that there's some probability that when you measure, the switch will either be on or off. This is not a classical probability though, it's not like a coin toss where if you just knew a lot about the dynamics, you could in principle predict the answer. This randomness is fundamental to the system. That's all. Then, just like on a classical computer, you encode a problem by "preparing" a quantum state. You're just going to mess with these switches in a way such that they contain the input to a question you want answered. Then, just like on the classical computer, you apply operations that correspond to "logic gates". But they're not quite as simple as "and" or "not". Just think that these are a way of changing the quantum switches -- and the probabilities -- in a definite way. You have to be a bit cleverer in choosing what operations to apply because quantum algorithms by nature are a bit harder to design. In the end, the third step, if you mucked the probabilities correctly, the answer will pop out when you measure the "output". "The only problem, and that is what is keeping us from having quantum computers, is that we can not yet be sure that we read the state that we want to read." That's not true, because quantum algorithms are designed in such a way that you don't need to be lucky: you get the right answer with some good probability. Designing such an algorithm can be very hard, as I said, but people can be very smart. The reason we don't have quantum algorithms is that quantum correlations tend to be very delicate and the interaction with the environment spoils the states before the calculation's done. It's a bit like expecting your computer to function inside a thunderstorm (assuming it wouldn't be outright destroyed, of course). There are many ways to protect the states from interference or compensate for the noise, but so far we've only been able to make quantum computers with a handful of qubits. So we can only do silly sounding things like factor the number 15. "The qbits allow for 4 possible entries making it a base 4." That's not true either, although I see why you're confused. The researcher on the video actually misspoke, he didn't mean 4 classical bits. He meant 4 real numbers (well, really 3, since one of them doesn't matter, but okay). Technically to store one real number, let alone 3, you need an infinite number of classical bits. So qubits can carry a *lot* more information, in that sense. The trouble is though, you can't read them out. It's a theorem that the maximum number of classical bits that can be reliably read out of N qubits is... N. So no free lunch there. So, to tl;dr, it's not that quantum computers allow you to store more information, they just allow you to "shuffle" it around in ways that make certain calculations more efficient.
More clarity: According to the principles of quantum mechanics, an electron can exist in a state of "superposition," where it has both spin-up and spin-down states simultaneously. This means that until a measurement is made, the electron's spin is not in a definite state, but rather is described by a probabilistic distribution that includes both spin-up and spin-down states. The concept of superposition arises from the fact that in the quantum world, particles like electrons do not have a definite state until they are observed or measured. Before measurement, their properties exist as a range of possible outcomes that are described by a wave function. The wave function gives the probabilities of different states that the electron could have when it is measured. Therefore, an electron can be in a superposition of spin states, which means it has both spin-up and spin-down states simultaneously until a measurement is made, at which point the superposition collapses into one of the two possible states. This is a fundamental aspect of quantum mechanics and is responsible for many of the peculiar and counter-intuitive properties of the quantum world.
This is probably the best explanation from all the comments I dug through...... "ideally each q-bit should be in either state 0 or 1. That state can be reached by aligning it with or against the magnetic field. They align it by giving each q-bit a coefficient of something, which they didnt explain. Once it is in state 0 or 1, it's just like a classical computer from there onwards." My replay.... Yes, he did not explain the coefficient. But your mentioning of alignment is related to the coefficient does make sense. I just hope that he explains that further. He said, in classical bits, all you have to do is giving 2 number of information (i.e. 00, 01, 10 or 11), but in two qubit system, you need to give 4 coefficient numbers in order to define the state of the two qubit system. Ah, I think you are right....the coefficient is related to the magnetic field. And that's how it is measured or stored. By having a certain amount of magnetic coefficient, the N-qubit system will be forced to be a certain state and that is how it's measured. AHHHHH....
From what I understand; classical bits can have two positions and can carry the information with combination of these positions. But in quantum computation every qubit can have a unique position(unique mathematical numbers for its position) which can indicate a spesific result, and which can have incredible amount of different combinations. That means a singular qubit by itself can carry the meaning of big amounts of classical bit combinations. As qubits increase and the number of computation steps increases, those dozens or maybe hundreds of qubits can carry the weight of impossible amount of classical bits. But only on that spesific task. (please if you got it better correct me)
Hey, that is good. Much better explanation than the 4 coefficients he is talking about. I didn't get what he says about having those 4 coefficients for the two qubit, but you explaining it better by saying a single qubit can have many states other than UP or DOWN. For example, if a qubit can have 3 states, i.e., UP, DOWN, MIDDLE, then three number can be represented by these states with a single qubit..... (0, 1, 2). And if you put 2 qbits together then, and each can represent 3 states, then there are 3^2 = 9 possible combinations. If a qubit can have 4 states, i.e. UP, DOWN, MIDDLE UP, MIDDLE DOWN, then 4 numbers can be represented by these 4 different states .... (0, 1, 2, 3). And 2 qubits with 4 states each can have 4^2 = 16 combinations. Then the question is...I thought qubit once it's measured, it is in either UP or DOWN state and other states cannot be measured ?
I like how the professor explained part of quantum states, but it really doesn’t help explain WHY it’s better, etc. Thanks for your additional thoughts on the subject 😊
Bingo! I finally understand the basic difference between classical and quantum computers. The different states of a quantum bit allow fornanwhole lot more variables, etc. to be set for comparison, and much faster because you’re doing stuff at the molecular level informed now. Good stuff. So much more to learn, but after seeing this vid and a few others, I’m much more
Are we in a simulation? Yes. Did this simulation come from a super powered quantum computer that our ancestors used? Maybe. A different being? Who knows. Whys life in the ratio 1.618? Some entity computed that. What programmed that entity?
Correct me if I am wrong, but what I gathered here are two things. 1. Quantum computers are more like Quantum storage devices where they store 2^n bits in a space of n bits (thanks to super position) 1.a Also you don't want to read the values in those bits during a calculation. 2. You also need a specially tailored algorithm that can store the bits and perform certain sets of classical operations using a proxy quantum operation, like instead of adding a number 5 times, you are able to directly multiply with 5. 2.b Once the operation is done, you can measure the output and get the answer and the data is lost.
Wasn't expecting Loki to do an explanation about quantum computing
The tesseract must have brought him here
C'mon you stoll my thought.
Bro that guys creepylookin af
I low key wasn't expecting it either.
@@無名兄弟-i7m random (dumb) question: the hanzi/kanji in your username looks quite complicated, that's not your actually name is it?
Classical computer: true, false
Quantum computer: maybe
I prefer yesn't
lol this poppped up in your recommendations lol this way made 8 years ago-
PERHAPS
Bruh as a programmer imagine we have individually force the bool to the system
Bruh….
That's the first time I've 'got' how quantum computers work. -'it's not the speed of the individual operations, it's the number of operations it takes to reach the result.' No magic. Andrea is a superb scientist. I love his explanations. Thank you!
Agreed
Well, even though some explanation is done, I still don't understand what an operation is and how it affects a qubit's state. And how operations mutualy use those states. And how one programs a set of operations anyways.
If I'm not mistaken, a qubit only holds the intermediate states between the beginning and ending of a sequence of operations. And after you read it the state is gone.
I'd interpret this as operations can't read qubits either but just influence it. Sounds a bit like analog values getting amplified and dampened until the outcome is reached.
Anyways, if you understand it better then please let my know
@@what9418 I just came here for the free headache. Good night.
Yes, even now I realise why they are used in HPC - high performance computers for parallel computing..
Just WHAT do you mean "no magic" ????
This electronic engineering is BEYOND magic !!!
Absolutely Astronomical out of this universe ...
Really like how this professor teaches. He's very understandable.
I disagree, I didn't understand anything, although I don't think that is the Professor's fault...
To be fair he didnt get into any details, it was just general information about the topic
@sr1nu he's italian and I think he works in Australia, so it's a superposition of the italian and australian accent
@@mmehdi3437 you want him to explain whole quantum physics in a youtube video? He talks very understandably and he's probably a great teacher. he knows what he's talking about.
If you're talking about the long hair dude on the whiteboard then you smoking rock... He didn't make a lick of sense to me. I've watched it twice now and have no clue how it works nor how it could be better than a standard computer.
it's amazing I don't even understand a single bit.
Neither do quantum computers.
Write these two comments in history books.
It's just really fast computing, the quantum state stuff is just on and off just faster"slower" (think of space time slowing down) like a dimming switch on a light. An easier computing way would be to have a computer register 5 and 0 as on and off 1-4 and 6-9 as quantum numbers (dimensional movement) no such thing as quantum it's just really fast. The photon pairing example can be just seen as gluons and bosons(slower)
+Ganjanaut moores law continues
+Ganjanaut LSRSL
Epic accent: Check
Epic hair: check
Epic soul patch: check
This dude's got it down
Yes, where are all the likes..
He is prolly french
@@navindamansitha3684 italian
Or does he have it... Up?
He is Italian
The comments section:
100% : I don't understand
100% : I understand
.. And that's quantum computing
That is true until you actually read the comments.
Underated comment
Its called understandn't
That's true and false at the same time
Aashish Singh so you understand then? Lol. I measured your state and saw that you understood. Maybe next time I measure, you don’t understand lol.
I lost it at the when he started explaining about the coefficient part
I totally don't understand the diagram at 3:01 and his explanation onwards.
same
math sucks :(
Where he was describing N qubits? All N is is a variable that is representing how many of something there are for instance if you might own 2 cars and someone else owns 3 cars you can describe both situations with both people owning Ncars where in your case N = 2 and for the other person N = 3. A coefficient is just a variable number placed before something that multiplies it. So what he was saying is if you have 1 qubit it can be in both positions at the same time called a superposition. where as classic computers can only be in one position at a time. As you add more qubits the positions can be multiplied exponentually. So he was saying if they have 300 qubits which means N = 300 and that 300 qubits can produce 2^N or 2^300 possible positions in comparison to classical computations.
Agreed. I keep looking for simpler videos but they're either not in depth enough or too complicated for me to understand.
Yes, but how does the computer *actually* work? How does it store a qubit? How does it entangle the qubits? How does it read the basis state? How is he algorithm strucured?
These are my questions, too.
Einstein spooky theory
Top 10 questions that even science cannot answer.
@@jcf20010 Shut up im smart!
@@jcf20010 This was literally the next video suggestion for me
almost 8 years since the video was made and i still think this is one of the best explanations i can find for my tiny brain to comprehend the basic definition of quantum computing.
Its amazing that last week a 500 qubit quanum computer was developed
@@glendisshiko8182 So that's the same as 2^500 classical bits? Amazing.
I have what scientists call QI or Quantum Intelligence, which exists in two states, before you measure it, I am both Intelligent and not, but once you measure it, what you find is that I am not intelligent 99% of the time. By the way, I don't even understand this enough to make an appropriate metaphor, and this is the 1% trying it's hardest here.
Haha
Lol
At least you were able to make a proper analogy
Couldn't that also be called SI? Schroedinger's Intelligence? :)
brilliant!
This video should be titled “Italian Metalhead Explains About Qubits”
probably romanian tho
@@nicholas132edm no, he was born in Pinerolo, Italy.
LOL most likely prog metal
He looks like Fabio Lione (power metal singer in Angra and Rhapsody) and his voice is almost exactly like Fabio too
A 'scientist' trying to look cool just comes off as a pretentious douche. Credibility dubious at best.
Andrea Morello is my new hero.
Explains things so nicely!
ikr
Did veritasium go on a holiday to Australia to make the video? ;) Edit: Oh he's Australian with a Canadian accent...
Debajyoti Sengupta you should check out how nice his hand jobs are.
@@Dhirallin he is italian
Tha guy's accent, it's legendary.
And his Adams apple
What accent?
And the way he writes (Inverted lefty?)
He somehow remind me of Loki, from Thor movie
Now you mention it, indeed!
02:33 - the moment he realized the average viewer won't understand anything. And he was correct.
To him, the average view both understood and didn't understand at the same time
i feel like i kind of understand, but actually not at all
You know what this is 8 years ago and I wanna learn more about quantum computers but don't have any recent resources
@@Wraient check IBM’s public resources and documentation on quantum computing.
@@hem9483 Thanks for letting me know
Finally someone actually explained how they work! I’ve know for years about how, “it’s nothing like a classical computer, it can have bits in superposition” but no one told me how to read the qbits or how they interact.
0:40 Ping!
0:57 Ping!
That physicist thought he was explaining himself but he really wasn't..
i lost it at 0:00
LMAO
XD
ramadhan karim
Youcef Mahdadi Allah akbar
ikr
For those wandering, why when you add up the coefficients in front of the possible states of the electron spin at 1:35 you don’t find one (which would mean he made an error in the probabilities) , it’s because these numbers are just coefficients, if you want to get the actual probability of getting an electron spin up/down, you have to square the coefficients:
(0,80)^2 = 0,64
(0,60)^2 = 0,36
And when you sum up these numbers you get one.
What happens when both coefficients are 0.50?
@@johncarson5436 From memory it is because of how the probabilities are calculated. Determining the probability is a function of the electron charge and its momentum. I realise that probably means nothing. Its a super abstract concept and is only expressed by the underlying maths.
@@johncarson5436this is not a physical state
No clue what your saying. lol
@@johncarson5436 if both coefficients are 0,50, than you must have made an error: 0,5 squared is 0,25. There are only 2 possible states for that electron: spin up of spin down, which means that your probabilities don’t add up to 1 but to 0,5 (which means there are other possible states)
If the probability coefficients are both 0,5 (and not just the simple coefficients) you have a 50% chance of observing that electron spin up, and 50% chance of observing it spin down.
PS: if the probability coefficients are both 0.5, that means the coefficients are equal to square root of 0.5, so that if you do the math backwards: sqrt(0.5)^2 = 0.5 :)
I emphasize on not confusing the simple coefficients and the probability coefficients
I really like that dude, his eyes are so expressive. Bitchin soul patch too.
Milky Way Laniakea Superclusterite His eyes are amazing
Milky Way Laniakea Superclusterite if you’re gay that’s cool. You GAY
Im so proud to say he used to be my lecturer
Seriously... when this guy looks at the camera I'm sure he can read my mind.
I admire how accurately and gracefully Andrea dissipates all the myths built around quantum computers.
I remember my physics professor at Georgia Tech was building a quantum computer back in the early 2000’s, but it was probably a 2^1 or 2*2 Qbits at that point, 😂. Smarter than I’ll ever be!
0:40 *BING*
0:57
😂
1:08 the big brother
I guess the future is hybrid CPU's.
Just like we use specialized CPUs called GPUs to render images, we will use QPUs (?) to calculate something which benifits from it, while still using CPUs for all the other cases.
to better wet expand on your concept i belive the "domestic" application that can benefit the most from these kind of operations are exactly those made by the gpu,that has to process lots and lots of shadows and physics calculations at the same time,or to render a large video,as per say if you encode 2 bits of it at a time or 2000 it doest make a diference,so i belive that we are going to see 2 things from this 1st is quantum based gpu,s and second something faster or even a improvement on ssd,s so that they can keep up with the large amount of data transfers needed to acomodate such power,otherwise your gigantic quantum computer will be as fast as you can transfer data arround...
*****
Or perhaps at some point we will hit a bottleneck,either because of the limitations or the possible price tag as an outcome.You can see it happening right now,instead of going bigger in sheer amount of processing power,developers create techniques to do whatever they want/can with the existing tech.
Think of it this way: You can stack cards only so high,before they collapse.
imagine the cooling you would need...
Kronguard Price is not a problem. Simple economics say that so long as the demand and supply are high enough, the price will drop. Developpers are creating techniques to harness as much power as possible with existing tech only because of the existence of consoles. Being limited by generations, instead of the continuous evolution of the PC, they have to look for more sophisticated techniques to get as much power out of those machines as possible. This is both a good thing and a bad thing. Creating these techniques frees up memory, but being limited and having to spend money on the technical aspect of the game instead of the gameplay and to some degree graphics is not that good.
You're not considering the nature of GPU's and the way they process information.
Cpu's work in serial; each line of code is taken in sequence, one by one. Really fast, but still, one by one, and has a few cores dedicated to that.
Gpu's work in "parallel"; they work with floating point numbers, and can take several lines of code and process them at the same time, it has thousands upon thousands of specialized cores.
The better an application is optimized for GPU-acceleration, the more "room" is given to a CPU for other processes, and thus the bottleneck someone mentioned can be taken care of.
Eventually yes, a CPU bottleneck will occur because transistors will eventually reach the limit in manufacturing process, and so will GPU's, but we still have a few years for that, and there's a big chance that by then there will be a different material that will allow for a smaller manufacturing process of the transistors, which will account for that.
Well that explains a lot, still don't know a thing
+Justin Zh I came on here to find out why people were so hyped about it and what it could do but i guess that info just isnt here.
+-=[Kuledude Gaming]=- You have to understand the properties of 2^n. Think of it this way. Put a grain of rice on the first square of the chest board. Double it it for the next square, adding 2 grains of rice, 4 for the second, 8 for the third and 16 for the forth and so on. When you reach the last square you will have (2^64)-1 or 18,446,744,073,709,551,615 grains of rice, a pile of rice the size of Mount Everest and 1000 years of the entire world's rice production at 2010 levels.
So, if you manage to build a quantum computer with n bits, and manage to keep those n bits entangled, it can represent 2^n states. So 4 for n=2, 256 for n=8 and 9 trillion at n=64. So in other words you can do much more for less. But only for certain (important) types of problems, because you need to be able to collapse all these states out to one you can actually read.
Uh... very long, much words, very confusing. XD Anyways, i think i got what you mean, but not sure.
uh ok? xD
xD
professor Morello is amazing he can explain this complex concepts so easily... the rockstar of Physics
The minute I saw him in the video, I felt like this dude has some really good energy. Just someone you would immediately trust! A genuine but also extremely smart person! People like that are very rare.
Don't understand a thing but still watching.
Now i understand it but still watching.
Nethkrill Vesta You've been watching for two months? Dedication, my friend.
Victor Kyrg Oh yes, my friend. These science videos are like ganja to me. lol
this is the best explanation of qbits i've ever heard. thanks!
cubits in the bible.....get it??
Agreed, this was very well explained.
don't pretend you understood it...
kidding, i just don't get it. i understand the computation power reference, but i don't get what superposition means and how to use it
Guys, this is really a great explanation of Qubits and the best one I've seen so far.
I suggest that you guys look at some videos on the basic properties/phenomenons of Quantum particles/physics before diving into Quantum Computers.
Why do I like veritasium dressed as phosphorus atom so much?
What I like even more is referring to Derek as "Veritasium". Which I will do every day from now on.
ping
+Mikolaj Gackowski Dr. Derek Muller
Vacso Kagazzle Laloobay Hoophorn Wacago Seiliu bb
I think I heard someone laughing below their breath when he was the phosphorus the first time around.
Finally, with this I can run task manager at 60 fps
veritasium dressed as a phosphorus atom is my fetish
Omygod
phos-play
So are your gonna ask him to put his huge electron in your vacant orbital?
this does not deserve any of the ~500 like it has
@@vivekbarjod6815 who says hes gay?
When you accept to wear in a big red atom disguise, that's the moment when we know you're truely dedicated to you channel!
Its not an atom its a proton
@@auredio6838 He is talking about Phosphorus, so no its not a Proton
i dress up like this every day
@@eeevoo oh
i watched this three times, just enough to gain 3% of what he is explaining. Im taking what little i have learned and protecting my sanity.
Lol
well i watched it three times and i am already at a superposition
You just need to watch it 97 more times:)
@@royhsieh4307 🤣🤣🤣🤣
@@101perspective hmm
Just loved that he did explain in the technical but understandable terms, and as a highschool physical chemist student, I understand all this very well.
Also love the nutsell telling that it just won't be improving our classical computing like browsing web or stuff
That long haired guy is an example of a great communicator; I think it's very easy for similarly intelligent guys to spew tech talk that passes over many heads
And on the side, his accent is great...
Agreed.
They could of chosen a professor from Cambridge, most I've seen are great explainers.
I was thinking:
"Wait... is he italian?"
Me controlling that he has an italian name:
"Yes"
he was half italian until you saw this video
How do you differentiate a westerner? For an Indian guy, all westerners are the same
That's raci...!! Kidding. By the accent; sounds very romance (as in romance languages), the way he speaks, the hard pronunciation of consonants, the intonation, the way he constructs his sentences and some words he uses. Also the name is a big hint, but you as an Indian probably have some of these features in your speech as well. There are Europeans who look noticeably different (i.e. Spaniards-Norwegians) but in some places they are really different to tell apart, even from Americans, and that's because it's a society built mainly by European immigrants that mixed maybe as much as in Europe itself.
@@kangkanlahkar9045 I guess by their accent, am an Indian btw
@@kangkanlahkar9045 as an italian, I knew he was italian before reading his name. He just speaks in the same way most italians do while speaking english at school. I don't know the specific features that make me say so, he just sounds italian. That's the way we speak I guess
He talked me out of building one.
Me too. I was gonna run Windows 95 on mine. Damn.
hahahah
what do u mean ? :P
一排污3哦2U切E 865 . , 0.,0000009.00=0
w。
.,, ,. :-);):D 可哦E
Don’t bother. Can’t even run Crysis... har har har.
I would love to see another video from you on the advancements (if any) made in quantum computing !
0:40 - "bing" that made my day.
No computer is out of the realm of the blue screen of death!
+carlos carrion Any computer that does not run on Windows
+sweiland75 they still can crash though
+sweiland75 OS X has crashed on me a few times. Linux too.
wowitsbryce
They do not have the BSOD
sweiland75 BSOD?
He does a great job at explaining it. Picks his words with great care. Now I want to know which types of calculations would benefit from quantum computing.
This is the most amazing ~6-minute explanation of quantum computing principle that I had ever seen.
but can it run crysis 3 at ultra?
+yrjosmiel73 If they're talking about it being good for massively parallel operations, I can imagine it may actually be applicable to GPUs.
+yrjosmiel73 As far as I know, no truly working quantum computer exists right now, so probably not. But then again, I won't know until I've searched for it. It is currently in a state of superposition.
+cjdrey Google has recently made a quantum computer, just look it up.
+yrjosmiel73 nope . Crysis wasn't made for quantum computers and like the guy said it may be slower cuz it would use power just to convert to traditional code. Maybe when quantum PCs become more relevant can the cry engine use it's advantages.
+yrjosmiel73 asking the right questions
0:39 Ping
nice
ping
spin not ping
Amber Spirit says ping
M1 garand
What I'm still having trouble understanding is how can we get useful results out of a machine whose state is based in probability and chance. I've been looking up answers all day, but very few make sense to me.
My thoughts exactly
Pretty much they are accessing the a wave function superposition meaning every possible answer you can think of and they are also with holding information on how powerful this machine really is .this machine is precognitive and it's been around and in use longer than they say it's accessing information from every source of possibilities that exists . Oh and it does so much more .
Alex Enschede You mean to say we can solve TSP with O(n^a) complexity with quantum computing?
To summarize what they said in the video, the computer starts with all possible states, putting all possible states through your algorithm as the variables and simultaneously calculating all possible results, and then you read the result, but by reading it, only one result is created, influenced by probability, because you can only read 1 or 0 and not the probabilities of superpositions.
For example, lets take a formula like a = (x && !y) || z. Lets say, for simplicity, we assign just one bit to each number on the right hand side, so x, y, and z equal either 0 or 1. If you calculated with a quantum computer it would calculate everything at once.
x | y | z | =a
0 | 0 | 0 | =0
1 | 0 | 0 | =1
0 | 1 | 0 | =0
1 | 1 | 0 | =0
0 | 0 | 1 | =1
1 | 0 | 1 | =1
0 | 1 | 1 | =1
1 | 1 | 1 | =1
I'd assume that every bit starts at 50% chance of being 0 or 1, so a would have a 62.5% chance of being 1. I'm not sure, but from what they were saying I think they may have some kind of method of checking each line of the table one at a time. I'm still not sure myself, it's a confusing field and I haven't yet found any absolutely complete explanation.
Basically, this could be used for powerful algorithms that consider every possibility at once. For example, a physics engine that calculates everything about a particle with every possible starting condition and then applies those calculations to each individual particle within the simulation at each tick of the clock; it would be like having parallel processors for each of the thousands or millions of particles in your simulation. This could be used for everything from gaming to theoretical physics simulations.
A basic example would be trying to obtain cryptographic secret key that matches known public key (this is mathematically possible but requires billions of years on classic computers). So you set the quantum computer so that all results will initially have the same probability and then you check it against the secret key with carefully designed algorithm. All possibilities will be evaluated at the same time and the correct result will increase its probability. Read the qubits and repeat the calculation a few hundered/thousand/million times. Then check the values on classic computer in order of how often they appear and you will find the result probably in a few seconds.
I have more simpler analogous example for everyone to understand here.
Mr X can understand 1 question from one of the 3 people throwing questions at him, at a time. So he will take 3 turns to understands all 3 ppl 1 by 1.
But here we have the genius Mr . Y. He has the special ability to absorb all 3 questions at the same time. Why wait 1 by 1 when you could take all at once ?? !! Awesome isnt it??
Now 2 Mr X (2 bits) will understand 2 questions at a time.
But 2 Mr Y (2 qubits) will understand 2^2^6=256 questions at a time !! Because they can superimpose their input ability in exponential!!
Keep adding Mr. Y (the qubit) and you get the capability to absorb billions of times larger amounts of questions in 1 go. Freaking awesome !!
Done.
------
Now talk about why they so large and ultra expensive??
Quantum computer basically lifts the limits of hardware construction. Current cpu design is such that it can only be some mm thick.
But quantum computer breaks this limit and goes all way up in hardware, thus you se those huge quantum CPUs.
More simple: Like intel or Amd cannot make faster cpu by increasing cpu size, it's counter productive and that cpu will not work. Like your brain cannot be the size of an elephant to make it more powerful, it will err and die.
But quantum cpu can be made as bigger as you want and it's processing power will keep multiplying exponentially!!! Thus it breaks the limits of current microarchitecture and utilizes full hardware possibilities in all directions.
But the problem of cost and power consumption will remain there...it will all depend how much semiconductor technology advance, allowing some smaller quantum computers still millions of times faster reach to the public. Will take some decades though.
Until then, we will only see them installed in big tech firms.
i didnt understand anything, but was oddly compelled to keep watching...
that long heir guy looks like a head of Vampire club xD
You just made my day
+mheboob khan hahahaha
+mheboob khan I live in Transylvania!
hair*
so basically you need (n * precision)^2 normal bits to determine a single qbit.
For example with regular 32 bits floating point variables you'll need 128 bits to determine 2 qbits.
yoni0505 Yeah !
it means we can play GTA 99999 With high or ultra .. also we can watch 4K or 99K on any phone
or any small device
THE Future
xSniperU Did you even watch the video? ua-cam.com/video/g_IaVepNDT4/v-deo.html 6:23
UA-cam suddenly decided to recommend this cool video after 10 years 😅
Basically you won't be able to watch porn faster, but it'll solve your math homework in a blink of an eye.
+1992mikern porn makes you stupid and impotent.
It should take far less than 300 milliseconds though..
Faster is better. I accomplish the same task in 300 milliseconds that takes other men 7 minutes
1992mikern lmao u bust after 300ms? XD
The NSA would be able to crack your Gmail password in the blink of an eye.
Its interesting that he says they can be both 0 and 1 at the same time. I'd venture to say that, they are not 0 and 1, at the same time. Instead, they are interchanging, just at such a rapid rate it is perceived as being at the same time.
its taken me almost 4 years to understand anything said in this video
Enlight me pls, i dont have 4 years, i wanna know now :D
@@mario2872 haha. Nice . Each of our messages is a year apart . I'm gonna go jack off now . Be back in a year
Revolution NOW Ur gunna whack it for a year?
@@trilexi 1 hour
@Mario A quantum bit can be used to count faster than a normal bit because a normal bit can only go on or off but a quantum bit can go on off or half off so it is just better for some things.
1:09 the sound of turning needle other way
how in the hell did I manage to understand that?
You must know english language.
quantum physics and computer science too
becuz ur fucking pretty
***** wouldn't that imply the opposite
listen the sentences carefully and the moment you realize you're loosing the track start again from where you remember the last thing you fully understand!
So, basically, it's good for making huge combinations in a short time period, but not for transforming information, because the position of each electron has to be measured every time. So it's good for storing data and data analysis, as I understand. Think of a grid of electrons that can go up and down, instead of the classic bit, where the electrons have to go back and forth.
Summarised it all perfectly in 3 sentences.
In other words, good for brute force computations.
@@Diana_L. Yes in other words :)
@@hamedkadkhodaie7715 So, what they're doing is basically trying to design more efficient heuristics by exploiting quantum properties. Unlike classical algorithm design, which is mainly concerned with reducing the asymptotic complexity of the problem itself.
@@Diana_L. exactly you explained it better than me
Cool. I always thought Quantum Computers are the perfect thing to replace traditional silicon-based computers. Turns out I was wrong. Thanks a lot for the very helpful information!
you was stupid as hell.. that's what you was..
Cant wait to play quantumm Minecraft
With every incremental discover in technology we get these huge leaps forward, i can't imagine what people will eventually invent with it.
So quantum computers aren't fast due to their individual operations being faster, but due to the fact that they require far fewer operations. Very insightful. Imagine a classical program that outputs you the quickest route to get from your house to the store. It must analyze every possible path there in order to compare and determine which path is shortest. It does this incredibly quick, however it requires enormous amounts of operations to give you the answer. A quantum computer, on the other hand, can process every possible path simultaneously, therefore requiring far fewer operations in the first place.
That's the right idea, but there's a catch. While the quantum computer can process each path simultaneously, only the information about one path can be retrieved. When the state of the qubits is measured, it collapses the superposition into a regular, non-superposition state. That is, all the information about the other states vanishes and it is essentially as if you had only analyzed one path. There are still ways to gain useful information out of the superposition (depending on the problem), they're just not always obvious and the information might not necessarily be useful.
@Trius IBM has a framework called Qiskit that you can use to write quantum computer software and you can run your code on their quantum computer connected to the cloud too. Here is the link and they also have video series on UA-cam explaining this.
qiskit.org/
@@hatemel-kharashy8856 I think in the future computers would have 3 processor types: CPU for the main processing, GPU for intensely parallel processing, and a quantum processing unit for all this quantum stuff.
I just understand that there will be excessively combination per second. so this can be used for password cracking? is that right? because electrons move insanely per second which means that we can't know where an electron is but they can make 4 meaningful letters. The electron move randomly and insanely so this is the case. Per a moment a 1 letter will be observed and it will give signal. But per second, there will be nearly endless moment because electrons move so fast. There will come nearly endless and random signals per second because of the movement of the electron. The electrons move so fast that we can't even determine where it exactly is. So, it is not matter how long your password is, it will be detected very very very quickly. They can define the letters, numbers and characters to that system and easily solve your password. Because all of the possibilities will be just numbers near the infinity. Number divided by infinity equals zero. Actually maybe the number of letters (4 electron positions) are not the case. The number of the moments and randomness are the cases. It might be just 2 positions too. They are just making use of the randomness and speed of the electrons andthey achieve endless number of combinations. If I understand it correctly.
This is the first time I actually understand the real difference between a Quantum computer vs a legacy one! Thank you heaps!
My thesis is that every result is based on the same quantum principle. Any result has to be handled as: true, false, both, none.
***** interesting
This is the best explanation I have heard yet of how quantum computing actually works and what it is good for
Anyone else notice that the numbers did not add up to 1 but the percentages added up to 100%. he messed up on his probability, the numbers were 0.80 and 0.60 which add up to 1.40, while the percentages were 64% and 36% which add up to 100%, he did the percentage math correctly, but did not do the probability part correctly.
0.8 * 0.8 is 0.64. 0.6 * 0.6 is .36.
0.64+0.36 = 1
those values were not the probabilities but, think of them as a weighting coefficient. when you multiply by the complex conjugate you get the probability. since these numbers do not contain i, the complex conjugate is itself.
Oh man! We got a genious here!!!!!!!
TheBscit It's called "i" learn it in Adv. Algebra II. Imaginary numbers, they were a bitch. If that "i" refers to something from Calculus then I have no idea what he's talking about.
Evan Watters the number "i" is and always will mean the same thing. Sqrt(-1).
They're called Born probabilities. Oversimplifying: The probability of measuring a certain outcome is equal to the squared modulus of the given "weight" (or amplitude). Aaron Miller is correct.
That guy is amazing(ly brilliant). He explains it well enough, despite the language barrier.
yeah but...can it run crysis?
Only at 800x600 on Medium, shadows off
It's amazing how well the professor explained it! High level overview, while also mentioning all the relevant stuff.
One of the best explanation I ever heard!!!
I understand that quantum computers are using probability instead of solidly well defined positions which allows quantum computers to increase computing power exponentially instead of multiplicative. Since Moore's law is ending instead of creating quantum computers why don't we use variable voltages in transistors as as more states then just on and off 1 and zero. Like if you know anything about transistors there on state is within some range of voltage and the off state is also within some range of voltage called the threshold voltage. Well why don't we make computers slightly more precise so that there are maybe 4 voltage ranges instead of just two. Use a 4 number system instead of binary this would make computers far more powerful and wouldn't require a complete materials or computing revolution.
This is more along the lines of what I was thinking. But I really don't have much in depth experience with components so I didn't know if something like this would actually be possible.
You realy don't understand the architecture of a regular CPU....Hi works only beacuse trasistors understand two states(power ON and power OFF) there is no alternative just YES or NO. To do this, what you think, you need to develop some new kind of transistor and that wouldn't be a transistor anymore.
Actually transistors do act like a variable resistor between certain ranges, he's right. The issue is they do not act reliably in that regime. With HIGH versus LOW, there can be a range under which a transistor is closed or open. In that grey area where the metaphorical switch is not fully open or closed, things get very uncertain and complicated. It would be incredibly difficult to utilize that property of transistors for many, many reasons. Manufacturing uncertainty would be enough to throw off any individual transistor in a set so that it doesn't read the same value as its neighbor under the same applied voltage to the gate.
I believe you would need a larger voltage since there is always a small threshold above 2 different voltage to represent 1s or a 0s so if we are going to add more states there has to be more voltages to fill the gaps and then we will need transistors which are able know 3 or more states and that's where it all fails.
What i really want is a light based mother board. But to make this any good we would need extremely fast conversion between electricity and light which i don't think will ever be possible.
There is no reason why we couldn't, it's just that the payoff isn't anywhere near as high. Both ideas have similar problems - stability and programming - that binary doesn't have because it's as simple as it gets. Using binary is like using triangles in computer graphics. Since triangles can only exist on one plane (in terms of filling in the shape when given the vertices), while other shapes could have ambiguity, they're ideal for precision (getting the same thing every time), but not for accuracy/efficiency (it takes a lot of triangles to approximate a circle). Similarly, binary is always right (in terms of mechanical calculations, floating point is irrelevant), but it's relatively limited.
It would be possible to compute in base 4, and we could do so with essentially the same architecture that we currently use, but it would require far more accurate and expensive technology that would still be incredibly finicky (say, if you wanted to overclock it). Secondly, we'd need to create a new operating/logic system to account for the differing outputs, as any conversion would waste the potential. Based of what was said in the video, switching to base 4 would increase efficiency, as the gradience of each output would need its own measurement - more information - but that number would strictly double (currently when computing, each place value is recorded in a separate bus lane, so the effect only applies to each output). While these hurdles are even larger in terms of quantum computing, it's still just a one-off, and the benefits increase exponentially as opposed to multiplicatively (by that, I mean that the equation for the information density of a q-bit is 2^2^n, as opposed to just x^n. Even if we were to move into hexadecimal, it would only take 5 bits for the quantum computer to win out).
However, (as far as I can tell) the benefit of voltage gradience is density as opposed to speed (it would only really improve memory bus width, but physical bottlenecks would still render that pointless, if not, worse than they already are), and one one the major problems/bottlenecks in quantum computing is the need to convert it back to binary, so if this were to be mastered, the quantum equation could be rewritten to 2^x^n. (Again, this part is outside of my knowledge)
Of course, this is just an educated guess based off my understanding of how classical and quantum computers work in theory, as well as a brain dump to organize my thoughts on the matter. I could be completely wrong, and I'd love to be informed about that last part.
Beyond me, lol
Mordant Victor my brain just exploded lol
+BennyDACHO I have zero qualifications and i can grasp it.
No, I'm just saying that you dont need to do computer science to grasp this concept.
No need for being pretentious.
BennyDACHO
And I lied.
I do have qualifications just they are in the automotive industry, not anything worth while.
Lol! I have now! But no i do not think that is myself.
I have had a natural interest in physics since i was 15 and was reading books such as goldie locks theory which actually introduced me to the multiverse theory, string theory and thus quantum mechanics which now incorporates string theory to make the multiverse theory possible.
Love the clarity the prof has and the nice silliness of breaking it down to my level 😀
"Does anyone would like to take a break?" Me: Please, let me out of here..
The "one type of computing" they never seem to tell you is brute force decryption.
Didn't know that Weird Al became a physicist.
lol
LOL
lol
If you get an answer from a quantum computer how can you verify the info is even correct? It's basically an educated guess, isn't it ?
Well, I never expected it to be like this. I got the idea of quantum computing (last year - 2020) even before I came across this vid. My basic concept was to make components smaller than what it is today. So small that it would use sub-atomic particles to process information, like some sort of hyper computer. Now I came to know that these computers actually exist and are "not" suitable for personal use.
OHH SO THATS WHAT MY PROFESSOR MEANT WHEN HE SAID MAGNETIC SPIN what theheck you explain it so simply my prof didn't bother telling us what magnetic spin was exactly
He probably didn't even fully understand it, so he couln't explain it. As Einstein said: If you can't explain it simply, you don't understand it well enough.
+s vashi Many professors don't go into detail because they want you to research on your own.
+qiwi111 What about people who do not do well at explaining?
+s vashi Its not quite that simple, if you watch the explanation of spin with this same guy on the secondary channel you'll see what I mean. The 'spin' of an electron (and in fact every other particle) is a purely quantum property that has nothing to do with magnetism in the fundamental sense. So while the electron's spin has a magnetic interaction, the neutron's spin, for example, does not. So your term 'magnetic spin' is a misnomer. Hope this helps.
@@qiwi111 no, it's not this simple. This video makes it look like actual rotation to align with a magnetic field. Electron spin is a very unfortunate name. It is a purely quantum effect and does not have a classical equivalent. It's not a tiny solar system. For example I read a paper somewhere that said that a 360 degree rotation of a magnetic field was not enough to return the state of an electron to its original state. It took another 360 degrees. So 720 degrees rotation to return to its original state. But I bet I'm oversimplifying it too, I'm still learning about it and I have a degree in physics. Your professor probably didn't want to lose the whole class and have everyone throwing their desks and walking out. Read more about this topic it gets crazy.
This guy made me realize that my top-of-the-class awards in first to fourth grade didn't really reflect my intellectual capacity. The more he explained it the farther I got from grasping it.
That’s good. Hang in there man
the summary:
the quantum computation - for now- can be used only for huge amount of data processing but can not be translated into our classical computational systems
that means we need an inter-translation system to connect between both worlds
I believe that what are we going to see in the near future
A quantum computer can do everything a classical computer can do, just much slower and using much more energy.
And, somewhat ironically, to program a QC and read out its results, you need a classical computer anyway. (Well, not really, but it does make things easier.)
The hope is that the qubits can hold superpositions of bit vectors, and resolve them into defined results.
This would be done via quantum gates, which are logic gates for quantum states.
And because a qubit can hold a superposition which is only stochastically defined until it is measured, you don't need to backtrack through all possible permutations to find a consistent configuration. At the end it can only hold a state that the computation allows. And that means it should be able to do some computations over large sets faster than a classical computer ever could. Quantum computers are basically their own class of computational complexity.
But until quantum supremacy is confirmed, it might not actually be any faster. In practice you have to deal with decoherence, which happens whenever a quantum particle interacts with anything else. And the more qubits you have, the more it will happen when you don't want it to happen, which is why you also need quantum error correction. You have to do the calculation several times, compare the results, and test them. And it might just be that that erases quantum supremacy. In theory it shouldn't, and if it does, new physics is required.
On the other hand, if quantum supremacy is confirmed, new maths will be required. And so far, a few pre-existing fields of mathematics have been sufficient to serve physicists well.
Well Google made it happen
@@007fearmonger We can't confirm there claims though.
Really cool how he made this video 10 years ago with a broad level of understanding and it's just gaining major relevance in 2024.
I didnt think it would be that hard to move a compass needle but his "ehh" just proved it was hard
yo mang i've scoured youtube and not a single actual in-depth explanation of why it is able to solve problems faster. namely how is an algorithm designed such that measuring the states yields an actual answer and how do you even know how to interpret that answer. Everyone is just like "well herp derp I got an intro and now I know how they work"
Tiwaking Tiwaking They solve certain SPECIFIC problems like integer factorization faster. Problem is no one has ever come up with a concise yet thorough explanation as to how they achieve this.
***** that's exactly what i mean by bs hand wavy explanation. It doesn't really explain how it happens. And though you may have followed the gist of the explanation, you still really have no idea exactly why an exponential growth in possible superposition states actually translates to faster problem solving. it is obviously more involved than just increasing the amount of working memory.
+Max Loh Its BS and handwavy because its exceptionally complicated. If you want a thorough understanding of how quantum computers work, you aren't going to get it from a 6-7 minute youtube video. You probably need at least a masters degree. This video is VERY simplified because otherwise nobody outside of the field would have any hope of understanding what was going on.
Dan Albrecht Perhaps so. I was hoping for something that might help us actually understand the idea, without getting technical. Because really all they did was explain that the number of possible states stored grows exponentially, not explain how that's even leveraged to solve anything faster, and then some people are claiming that they "understood" the explanation just because they grasp the concept of exponential growth.
For example, for "why does diffusion happen" the hand-wavy explanation would be "particles tend to state of disorder", but it's much more informative to explain that the reason this happens is there are so many more disorderly states over time so it's just a matter of probability -- then a layperson can completely understand how this works without delving into any equations.
Similar explanations are available for many other physical phenomena and computer science problems. But maybe for quantum computing there is no such middle ground.
+Max Loh Totally agree! I've been casually looking for this explanation myself for years and have yet to come upon it. Namely how does one setup an optimization problem to leverage qbits and how do you then read qbits to get the answer. I don't think the reason is hasn't been explained is because "it's technical" and I really don't want a watered-down non-technical answer. I'm sure that equally complex things have been explained both in video and in literature. For some reason this one has been elusive. I mean no one has even tried to explain it. Weird.
I still don't understand how the randomness of superposition helps in calculation.
Y'know what? This is magic. That's the only reasonable explanation. I'm gonna go look for witches now.
“The speed is not in the amount of operations that it can perform but just in the amount of operations it needs to get to the mathematical result”. He couldn’t have been explained better, this is awesome.
"The bit can be a zero or a one at the same time"
I don't get it.
Mike G you can blink and breathe at the same time.....understood??
@@Michael12HM you can breathe but breathe at the same time
I always struggled with the quantum physics theory that says an object can be two places at the same time.
Having a background in computer hardware and software, it's hard for me to imagine a bit that can be both high and low at the same time.
@@mikeg4972 There is a famous thought experiment called Schrödinger's cat. In this experiment, a cat is in a box with a radioactive source and a poison. If an atom from the radioactive source decays, a Geiger counter detects the decay, triggers a relay, and releases a poison that kills the cat. If an atom doesn't decay, the cat is alive. Because the situation is dependent on the superposition of the radioactive element, an observer can't really know whether the cat is alive or dead until they look in the box. Therefore, the cat is both alive and dead at the same time. Now, there are lots of interpretations that make this paradox a non-issue (and you can read about them on the wikipedia page) but the point is that you can't know the state of a subatomic particle until it's observed, but the act of observing it affects its state. This makes sense because if you want to measure the position of a particle, as an example, the methods needed to observe it (i.e. a visual observation depends on photons) will move the particle as they interact with it.
It's not that a qubit is actually both one and zero at the same time, it's just that we don't know what the state is until it's observed, but observing it will give us the wrong answer because our observation modifies its state. So instead we think about it as a probability space. If we know the probability that a given qubit is 1, and that probability is, say, 85%, and we rerun the algorithm maybe a few dozen times and the mean probability that qubit is 1 is still around 85%, then we can be reasonably certain that the value really is 1, and we never need to look at it.
Everyone is talking about their understanding or the accent, can we appreciate the fact that Derek dressed up as a phosphorus atom? ❤️❤️❤️
yes indeed! and ive never seen him comment on that choice ever since.
So, I still don't know how quantum computers work.
How does the computer actually use the quantum states to do calculations?
Or is it just a metaphor, and it still works with normal electric switches?
Because theoretical information is useless, we need ACTUAL information!
Really wish someone would explain it to me ....
For as far as I understand. When you have 2 qbits, they are not yet defined as up up, down down, up down, down up, until you read them. So on 2 qbits you can store 4 combinations. The only problem, and that is what is keeping us from having quantum computers, is that we can not yet be sure that we read the state that we want to read.
***** Thanks, I think I get it a bit more now ^.^
Maybe I can ask a physics major or professor about it some day ...
Nerobyrne :') I wouldn't trust my answer to much, I haven't finished highschool yet. It's just what I think it means.
This might make the explanation a bit easier for some: The classical bit system allows for information to be stored in binary, or base 2. This means for each bit there can only be 2 possible entries, 0 or 1. The qbits allow for 4 possible entries making it a base 4. The decimal system we use every day is base 10 (0-9). This means that more information can be stored in a smaller space. As for how to calculate anything with the data, I still can't get my head around being in both possible states at once. Quantum mechanics is beyond me for now.
I can't tell you exactly HOW these calculations are performed. To do that, I would have to teach you quantum mechanics. This is not that surprising, even teaching you how a classical computer works would require teaching you about logic gates and stuff. But I'll try to answer your questions in a sort of "high-level" pop science way. I'll try to explain how a classical computer works in a language that will make it easier to "get" what a quantum computer does.
"How does the computer actually use the quantum states to do calculations?"
Every computer is essentially a physical system that realizes some model of a computer. The computer you have on your desk uses classical states to represent zeros and ones -- specifically, it uses electrical currents. If you can distinguish between there being a current and there not being a current, for instance, you have two states that you can label "zero" or "one" as you wish. For definiteness, let's say that "there is a current" is the "one" state.
The second step is to apply some operation on these states: in the abstract they're called "logic gates", and they have simple rules like "and", "or", "not", etc: these are realized by a circuit. An "and" gate is a circuit that outputs a current if both of its inputs receive a current. A "not" gate outputs the opposite of its input, and so on. With these basic operations you can build any classical calculation on binary numbers.
The third step is to measure the result: let's say your computer isn't very sophisticated and doesn't have anything like a screen. Then you'd take an ammeter to the "outputs" and you measure if there is or isn't a current on each of the output bits. Then you'd know the answer!
A quantum computer follows the same basic paradigm: you start with a physical realization of a computer model, but you can't use anything quite as simple as electric currents. You have to use objects that obey the laws of quantum mechanics. The video talked about spins, I think spins are way too complicated to explain to the layman. So let's just say you have a quantum switch that can be "on" or "off", or any superposition of the two. Don't think too hard about what a "superposition" is, in practice it just means that there's some probability that when you measure, the switch will either be on or off. This is not a classical probability though, it's not like a coin toss where if you just knew a lot about the dynamics, you could in principle predict the answer. This randomness is fundamental to the system. That's all.
Then, just like on a classical computer, you encode a problem by "preparing" a quantum state. You're just going to mess with these switches in a way such that they contain the input to a question you want answered.
Then, just like on the classical computer, you apply operations that correspond to "logic gates". But they're not quite as simple as "and" or "not". Just think that these are a way of changing the quantum switches -- and the probabilities -- in a definite way. You have to be a bit cleverer in choosing what operations to apply because quantum algorithms by nature are a bit harder to design.
In the end, the third step, if you mucked the probabilities correctly, the answer will pop out when you measure the "output".
"The only problem, and that is what is keeping us from having quantum computers, is that we can not yet be sure that we read the state that we want to read."
That's not true, because quantum algorithms are designed in such a way that you don't need to be lucky: you get the right answer with some good probability. Designing such an algorithm can be very hard, as I said, but people can be very smart. The reason we don't have quantum algorithms is that quantum correlations tend to be very delicate and the interaction with the environment spoils the states before the calculation's done. It's a bit like expecting your computer to function inside a thunderstorm (assuming it wouldn't be outright destroyed, of course). There are many ways to protect the states from interference or compensate for the noise, but so far we've only been able to make quantum computers with a handful of qubits. So we can only do silly sounding things like factor the number 15.
"The qbits allow for 4 possible entries making it a base 4."
That's not true either, although I see why you're confused. The researcher on the video actually misspoke, he didn't mean 4 classical bits. He meant 4 real numbers (well, really 3, since one of them doesn't matter, but okay). Technically to store one real number, let alone 3, you need an infinite number of classical bits. So qubits can carry a *lot* more information, in that sense. The trouble is though, you can't read them out. It's a theorem that the maximum number of classical bits that can be reliably read out of N qubits is... N. So no free lunch there.
So, to tl;dr, it's not that quantum computers allow you to store more information, they just allow you to "shuffle" it around in ways that make certain calculations more efficient.
More clarity:
According to the principles of quantum mechanics, an electron can exist in a state of "superposition," where it has both spin-up and spin-down states simultaneously. This means that until a measurement is made, the electron's spin is not in a definite state, but rather is described by a probabilistic distribution that includes both spin-up and spin-down states.
The concept of superposition arises from the fact that in the quantum world, particles like electrons do not have a definite state until they are observed or measured. Before measurement, their properties exist as a range of possible outcomes that are described by a wave function. The wave function gives the probabilities of different states that the electron could have when it is measured.
Therefore, an electron can be in a superposition of spin states, which means it has both spin-up and spin-down states simultaneously until a measurement is made, at which point the superposition collapses into one of the two possible states. This is a fundamental aspect of quantum mechanics and is responsible for many of the peculiar and counter-intuitive properties of the quantum world.
Andrea Morello is an amazing teacher! Thanks for that!
von quvantum state!
two quvantum state!
tree quvontum state!
EHEHEH!
This is probably the best explanation from all the comments I dug through......
"ideally each q-bit should be in either state 0 or 1. That state can be reached by aligning it with or against the magnetic field. They align it by giving each q-bit a coefficient of something, which they didnt explain. Once it is in state 0 or 1, it's just like a classical computer from there onwards."
My replay....
Yes, he did not explain the coefficient. But your mentioning of alignment is related to the coefficient does make sense. I just hope that he explains that further. He said, in classical bits, all you have to do is giving 2 number of information (i.e. 00, 01, 10 or 11), but in two qubit system, you need to give 4 coefficient numbers in order to define the state of the two qubit system. Ah, I think you are right....the coefficient is related to the magnetic field. And that's how it is measured or stored. By having a certain amount of magnetic coefficient, the N-qubit system will be forced to be a certain state and that is how it's measured. AHHHHH....
Amazing job!
Greatly appreciate all the work you do, I've never seen anyone who would be able to explain such a complex topic in just 6 minutes
How would you debug a quantum computer?
From what I understand; classical bits can have two positions and can carry the information with combination of these positions. But in quantum computation every qubit can have a unique position(unique mathematical numbers for its position) which can indicate a spesific result, and which can have incredible amount of different combinations. That means a singular qubit by itself can carry the meaning of big amounts of classical bit combinations. As qubits increase and the number of computation steps increases, those dozens or maybe hundreds of qubits can carry the weight of impossible amount of classical bits. But only on that spesific task.
(please if you got it better correct me)
Hey, that is good. Much better explanation than the 4 coefficients he is talking about. I didn't get what he says about having those 4 coefficients for the two qubit, but you explaining it better by saying a single qubit can have many states other than UP or DOWN. For example, if a qubit can have 3 states, i.e., UP, DOWN, MIDDLE, then three number can be represented by these states with a single qubit..... (0, 1, 2). And if you put 2 qbits together then, and each can represent 3 states, then there are 3^2 = 9 possible combinations. If a qubit can have 4 states, i.e. UP, DOWN, MIDDLE UP, MIDDLE DOWN, then 4 numbers can be represented by these 4 different states .... (0, 1, 2, 3). And 2 qubits with 4 states each can have 4^2 = 16 combinations. Then the question is...I thought qubit once it's measured, it is in either UP or DOWN state and other states cannot be measured ?
I like how the professor explained part of quantum states, but it really doesn’t help explain WHY it’s better, etc. Thanks for your additional thoughts on the subject 😊
It's amazing how they are using spin quantum nos and superposition principle to determine binary digits.
First time I can better understand what's quantum computing
Julie Coupard one ticket 2 CERN plz
I want to study quantum mechanics to get a long hair like him
Bingo! I finally understand the basic difference between classical and quantum computers. The different states of a quantum bit allow fornanwhole lot more variables, etc. to be set for comparison, and much faster because you’re doing stuff at the molecular level informed now. Good stuff. So much more to learn, but after seeing this vid and a few others, I’m much more
Are we in a simulation? Yes. Did this simulation come from a super powered quantum computer that our ancestors used? Maybe. A different being? Who knows. Whys life in the ratio 1.618? Some entity computed that. What programmed that entity?
Correct me if I am wrong, but what I gathered here are two things.
1. Quantum computers are more like Quantum storage devices where they store 2^n bits in a space of n bits (thanks to super position)
1.a Also you don't want to read the values in those bits during a calculation.
2. You also need a specially tailored algorithm that can store the bits and perform certain sets of classical operations using a proxy quantum operation, like instead of adding a number 5 times, you are able to directly multiply with 5.
2.b Once the operation is done, you can measure the output and get the answer and the data is lost.
3:19 when The Count was counting qubits I was beside myself. "Thrrreee bits, ha ha ha."