Saw quantum time? Google claims that dnen potae are pekonny

The Google team claims that a completely new type of quantum pot is in some ways irreversibly superior to the classic pot. Apparently this is true, even though the quantum pots are in their infancy and Google’s research team has been mistaken for almost ten thousand years.

Google had a media bomb ready for June 23 this year. On that day, Nature published a paper in which the company claimed that there had been a point where the daytime classical electronic (and spintronic wall) potas were grown in some beds to become quantum pots. The fulfillment of the grand publication did not go exactly according to full, and NASA and Google’s algorithms themselves are to blame.

Roughly before the release of work, a team member at NASA’s Ames Research Center apparently inadvertently uploaded a copy of the scientific text of the shower to NASA’s Technical Reports Server. He quickly corrected the mistake and deleted the text.

With this search (specific search crawler), Google captured the text on NASA’s small server almost immediately. The system immediately sent a copy of cadmium, who had a check mark in the service of Google Scholar’s scientific publications, that he wanted to receive news from the field of quantum pots. At first, the unstoppable outburst broke out and the bombing of the bomb in June ended up with a significantly smaller effect, which was probably not hoped for by the PR department of Google.

If you are not interested in all the computer technology, you can easily find here more important information before the team escapes everything that escapes me. We invite all the others according to them to either learn something new, or at least have fun with how (gave) the newspaper hard to fight the darkness of quantum pots.

In the world behind the mirror

Quantum potae have a rich history of pepper. They were dreamed of by the famous physicist Richard Feynman before the age of felt. He seems to be the first to realize that, using quantum mechanics, it is theoretically possible to build a device with tremendous power.

It is thanks to which several seemingly really strange contexts of the quantum world are connected. The first of these k, that very small things can be in several states at once (so-called quantum superposition state). Classic bit me can be either 1 or 0; quantov bit (or qubit) me bt both at once. It makes no sense to think about dark, as and for it’s mon. In our macroscopic world it is not possible, in the world of elementary states it is.

In one qubit, it is not a big difference compared to the classic bit, in fact only a double amount of information. But again, the normal rules do not apply in the quantum world. Quantum potae, in addition to the principle of ambiguity, ie the superposition of the state, use the so-called magic bond of quantum interconnectedness.

This is another non-intuitive phenomenon: at this point it is very simplified that a magical bond is formed between two atoms and atoms. When you can distance them and hit them, you can then immediately affect the state of the same species by acting on one of them for theoretically any distance.

So when you make two quantum bits, that is, two qubits, you can have so much information at once. Theoretically, there is no reason not to take them to go more: you, ten or a hundred. And here is the incredibly fast amount of information in the system grows with the power of the number of connected qubits. In the case of two it is 2², ie 4. For pt it is 25, ie 32. And for 53 qubit it is at 253, tedy 9 007 199 254 740 992 (devt billiards).

You probably want the vitality of the idea. Unfortunately, its implementation is not easy. By creating a connected qubit system and running the sweat, it never ends. The result then you have to get out of the pot so that you don’t damage it. If you simply look at the qubit state during the pot, they would collapse into a boring classic stitch with a single bit of information. The magic of the quantum pot would disappear immediately.

Therefore, special techniques (based on the principle of so-called quantum interference) are used to achieve the result. As the name suggests, it is basically a matter of the individual possibilities within each other being either mutually reinforcing or, on the contrary, weakening slightly. So some are becoming more likely because they are supported, while others are more likely because they are different. The correct construction of the quantum pot (and the correctly set algorithm) gradually amplifies the first one, and at the end of the calculation the first one remains.

Quantum potae will never be fast in all types of lye. Not always a similar parallel calculation is such a big advantage. But there are problems that we are absolutely sure of that will make them unbeatably the best. Probably the most famous of these concerns potato ifrovn and is named after its discoverer, mathematics Peter Shor, the so-called Shorv algorithm.

It was published in 1994 and it was the first test of the practical usability of quantum pots. From the layman’s point of view, this is a completely uninteresting guide on how to look for the prime movers of large peas; in fact, Peter Shor showed that quantum potency, thanks to its properties, should be able to break the whole series of then (and now still) used ifers.

There are other famous ideas on how to bring quantum computers to it, which classical computers cannot handle (especially the Groverv algorithm on a transparent database), but the year 1994 was a turning point. Since then, the advent of quantum systems has been a mixture of fears and gifts. For decades, I have been talking about the feelings of the Platonic, because it was clear that quantum potencies are far away. But how (among other things) proved the result of Google, asu ubv.

Pro 53?

You may have noticed that we chose a rather special number 53 for the interpretation of the extremely fast number of quantum pots.

The Sycamore processor was the first to work with 53 qubits, and with its help, Google announced the advent of quantum dominance in the Nature magazine, so let’s repeat this moment when the quantum computer reln of some bugs faster than even the fastest classical computers.


Fotografie ipu Sycamore

Sycamore is a device based on today’s much-favored technology of creating qubits in shears of superconducting conductors (let’s add that not only because of the superconductivity cooled by liquid helium to a temperature close to zero). This specific device was developed directly by Google, in the form of a mesh surface of a superconduct with 54 nodes, ie qubits. Yes, the actual 54 knots one qubit was defective and it was only possible to work with 53. As you can see, quantum potes are not exactly the most reliable device.

54 qubit

Schematic representation of the Sycamore IP: 54 qubits (edits) of connected hyphens (blue rectangles). The faulty qubit is also displayed in the middle row roughly in the middle.

The least interesting thing about the whole thing is perhaps the problem on which Google has shown quantum dominance. Only the sawdust was chosen so that quantum sweat could prove its strength on it, wrote John Preskill of California State of Technology, author of the term quantum dominance, in Quanta Magazine, adding: Otherwise, it is not a problem that would be interesting from a practical point of view.

A very simplified one then executed a randomly selected set of instructions, and then the leaders changed the result. After that, they do not have an obvious application and serve primarily as proof that the team has mastered quantum hardware enough to handle relatively reliably some problems that classical computers simply cannot.

He has shown Google that today we can build a quantum machine that is large and accurate enough to solve a problem we could not solve before, Preskill wrote in an article.

IBM vrac the

According to Google, the difference between the results of quantum and classical potency was indeed dangerous. Kvantov Sycamore coped with the sweat in about two hundred seconds, the fastest days of the day would be the same problem in about ten thousand years.

Between no one of the line and its release, the IBM team got involved. On June 21, before the official publication of his competitor’s work in the Nature magazine, IBM’s team published professional work and text on the company’s blog, according to which Google was very prominent. According to them, the strongest sweat in the world would count the bed much faster in the first ten thousand years.

IBM is thinking of one particular pot, the Oak Ridge National Lab Summit, which is currently the strongest pot in the world (approximately 200 petaflops) and which, in this case, is extremely 250 petabytes. Such a loit (simplified one) will fit a full listing of all the possible states of the quantum 53-bit Sycamore ip that Google used. Thanks to that, I work faster, not anticipated by the simulation.

Therefore, IBM estimates that the calculation of the fastest classic day of the day would not take ten thousand years, but about two and a half days. And according to n, this is the worst result, in fact it should be faster. It’s just an estimate. The calculation did not take place, because IBM would not prepare the test software for the test to go there on the Summit schedule. By the way, as Scott Arronson remarked, if the deadline is released, Google will undoubtedly have no hope in its work at all, because it could directly verify its returns, ie verify the actual implementation of the comments. Their own estimates were based on simulations and checks on a small sample of sweat.

Experts who comment on the situation quite agree that the IBM proposal seems completely rational and practically feasible. According to Google, the team overlooked a relatively obvious alternative. The question is whether the situation is something to me. Apparently not, quantum dominance is inevitable at least in the case of this type of function.

For? As we have described, the Sycamore ip works with 53 qubits and its simulation will fit into the 250 petabyte loit superpotae Summit with a small margin. But the power of a quantum pot does not grow liner. This means that if Sycamore had you 55 qubit, this would not be possible. The simulation would not fit in the disk array. And if the quibit was 60, it would take about 33krt more space. It seems that we are really roughly in the area where quantum potase begin to bounce back very fast.

On the other hand, the case shows that the classic potae certainly did not say the last word. A number of people will be able to think more intensively about algorithms, which could solve the same length more elegantly and faster even on a classic computing machine, and I will succeed. But the advantage of quantum technology seems so great, over a time horizon of five years and decades, someone else will keep up with them.

Let’s add that skeptics will be found. Those interested can take a look at the blog in the last days of the very active mathematician Gil Kalai, who has been questioning the possibility of quantum dominance for a long time and is very diligent (he would not fail) to look for mistakes in the newly published work of darkness from Google.

Kind of Mr. Shore

The Google machine did not have a big problem with quantum sweat. That is their error rate. Day qubits are very sensitive and I can easily influence them by external influences. And as soon as an error in the quantum sweat appears, it must be eliminated quickly, because just like the power input, its effect on the swell grows, and it will soon collapse. It seems that this problem can never be completely eliminated.

This is a consequence of the quantum operation of the environment. They have the unpleasant feature that they usually establish similar relationships with their surroundings as with each other. Inevitably, mistakes get into the quantum sweat, which, by their nature, cannot be prevented. With the classic technique, you had to be careful not to place a magnet on the disc or put a floppy disk on the speaker, but otherwise you could be sure that your data is perceived safely in the world and will not just change on its own. . (The classical logical unit is distinguished from zero by five volts, which is a difficult constraint in normal circumstances.) The quantum system is much keh, because we simply do not know how to completely separate it from the environment. We can’t build quantum walls.

Nevertheless, today there is a price that is generally considered feasible, unfortunately (including) inexpensive. It was designed by the well-known Peter Shor, who considers me to be all cryptographers after quantum sweats. Piel with a proposal of a way to correct the quantum algorithm according to the initial stored task. At first glance, it’s a natural thing: just don’t make a copy, eh? Perhaps it occurs to kadho…

In the quantum world, the bag is unfortunately something like that. When you want to make a copy, you have to pest originl. And if you read the quantum algorithm exactly, it will collapse. The top and a series of vdc after them eventually saw a relatively sophisticated system that can back up the original load to other qubits in the system without even damaging the original. It works even in some experiments (see for example this work), but only small ones.

So far, a working standard has never been shown on a system similar in size to Google’s for its evidence of quantum dominance. So on the pot, which is 50 and more qubit. The reason is that to achieve the required level of reliability of error protection, several times the number of control quibits would be needed. And no one has anything left today.

In the event of future futures, the problem will be more severe. For example, to break some significant stills ifer you will need a pot capable of working with imported qubit printers. But if there is to be a number of protections against the occurrence of errors that occur during the spontaneous operation of spikes with the environment, then, according to experts, millions of additional control spikes will be needed. And something like this can be hardly imagined with the same technique.