IBM Delivers Its Highest Quantum Volume to Date

[PowerPC]

When you reply using Don't know what kind of conspiracy theory that is. what sort of response did you expect?
It's a form of attack. Have some manners when responding in future, You might get a better response.

You literally wrote "they" won't let us have this technology. It's the pure definition of a conspiracy, and it's your theory...

Maybe actually read my whole post as I gave your post the same favor, before answering with MEH after only reading one sentence of what I wrote. Don't talk about manners, if you want to ignore the smallest criticism of what you wrote so quickly. This is a public forum, so be prepared that what you write will be scrutinized. And no need to take these things immediately as a personal attack.

 

[Hemmingstamp]

That article over-simplifies things to a great extent. It isn't meant to demonstrate an actual risk, only to present a theorized concept. A quantum processor would need to have a base algorithm for the compute needed to break any given encryption scheme. Each variation would need it's own algorithm. Such is not easily achieved as a directed attack on a cypher has to be built in a way similar to conventional methods. The more complex the cypher, the longer it will take to crack, even for a quantum processor. For example, a lone 256bit cypher would take a quantum processor with 128qbits approximately 3 days to break(using a proposed theory as a basis of estimation). Using a double cascade cypher set would take 8 years. A triple cascade, 73 years. Increase the cypher strength to 512bit or even 1024bit and the time needed still falls outside the range of a complete human lifetime. Increasing the number of qbits per quantum processor does not guarantee an equivalent reduction in time.

Cracking encryption does not become a breeze with a quantum computer, it just becomes faster depending on the parameters and requirements at play. Quantum computers do not and will not make today's modern encryption irrelevant. Things just don't work that way.

It was meant to be simple. I wasn't going to post an article that some may find difficult to get their heads around. Science evolves daily and has many who think differently. And with corporations wanting to sell their wares the threat of breaking that encryption is very real. Not my words, see Microsoft's Brian LaMacchia as a starter.

One question is, when is the deadline? When will an internet-breaking computer actually be available? Today’s best machines can manipulate a few dozen qubits. Brian LaMacchia, who runs the security and cryptography team at Microsoft Research, thinks a “cryptographically interesting” quantum computer might be able to handle somewhere between about 1,000 and 10,000 of them. Predicting progress is hard. But Dr LaMacchia reckons such a machine might be ready some time between 2030 and 2040.

Full article here: https://www.businesstelegraph.co.uk/quantum-computers-will-break-the-encryption-that-protects-the-internet/

It's doable, and to be be told otherwise is madness. I also stand by what I said earlier regarding the dangers of QC in the wrong hands.

You literally wrote "they" won't let us have this technology. It's the pure definition of a conspiracy, and it's your theory...

Maybe actually read my whole post as I gave your post the same favor, before answering with MEH after only reading one sentence of what I wrote. Don't talk about manners, if you want to ignore the smallest criticism of what you wrote so quickly. This is a public forum, so be prepared that what you write will be scrutinized. And no need to take these things immediately as a personal attack.

You litteraly wrote'........What??? I didn't. I will word it for you to understand. It isn't readily available for the masses. like a home computer.
And in the hands of those wishing to carry out evil on others....You get the point.

I didn't take your post as a personal attack, but your choice of opening lines was an attempt to smear shit on what I suggested because you didn't like it.

 

[PowerPC]

I doubt this technology will be available to the average Joe, do you? It's clearly in the hands of those running / or are part of the circus, hence my reply.

So saying "It's clearly in the hands of those running" and therefore "this technology will (not) be available to the average Joe"

... isn't the same as saying there is a huge conspiracy around this stuff? Do you really know how many people would have to be involved for this to happen? Besides the fact that people can still figure it out as I explained in my first post? Sorry that I'm skeptical... Because things like this are usually called conspiracy theories if you don't have anything to back it up beside the claim.

And as I said before, the overwhelming evidence, like with most conspiracy theories, is on the other side. The sane side. The more mundane side, if you like. Just the fact that they are bringing out these press releases on every little development for many years now, one of which we are literally reacting to right now, contradicts your claim directly. Plus, it's not even hard to find academic papers on this. I know it's easy to string up wild claims in any which way possible when it suits your narrative, but when someone points out the obvious flaws and tries to show you the holes in your reasoning where you just have to connect two points in a straight line, it's suddenly smearing shit on your suggestions and you can call it whatever kind of attack you like. Still doesn't make it an attack. Wasn't even hostile. Just drink some tea with a bit of Namaste.

 

[lexluthermiester]

It's doable, and to be be told otherwise is madness. I also stand by what I said earlier regarding the dangers of QC in the wrong hands.

The problem with that statement is that he projects a 2030-2040 time frame to crack TODAYs encryption, maybe. By that time encryption will become so strong and ubiquitous that quantum processors will have only a marginal advantage. The problem remains the same, to crack encryption you need the right tools and a platform fast enough to do the job in a reasonable amount of time. That isn't happening anytime in the next decade to 15 years. Even then it will take a great effort and amount of resources that only the most important of tasks will be justified for.

Quantum computing is not the "Magic Wand" that will make everything faster. It functions a specific way for a specific kind of computing. Cracking encryption is not one of them.

 

[PowerPC]

The problem with that statement is that he projects a 2030-2040 time frame to crack TODAYs encryption, maybe. By that time encryption will become so strong and ubiquitous that quantum processors will have only a marginal advantage. The problem remains the same, to crack encryption you need the right tools and a platform fast enough to do the job in a reasonable amount of time. That isn't happening anytime in the next decade to 15 years. Even then it will take a great effort and amount of resources that only the most important of takes will be justified for.

Quantum computing is not the "Magic Wand" that will make everything faster. It functions a specific way for a specific kind of computing. Cracking encryption is not one of them.

Besides that, encryption can be made theoretically infinite today already. It's not hard to double our 128 bit encryption, that has never been cracked before, to 256 bit, which is already being used. But you could just keep doubling and doubling it as much as you want. We just take the speed of decrypting into account, that's the only reason we don't already have 2048 bit encryption or even more. But 20-30 years from now it won't matter anyway. We'll probably just keep extending encryption to what is far, far stronger than anything any quantum computer could crack in the next 50 years, which is already the case today.

 

[Deleted member 185088]

This is exactly right. A quantum computer will never, ever make it into a PC (as in a Personal Computer) or anything that is supposed to be a generic number cruncher. One explanation (and I have no idea if this is accurate) is that a quantum machine is great at quantum calculations, meaning stuff that happens on a quantum level. Like accurately calculate properties of molecules or magnetism. Simulations. And a bunch of other very specific stuff. At this point it's hard to know exactly what they will be great at as they are still very much beta machines. So you won't use one as a personal computing device, but one day you might task one with work, especially if machine learning and quantum machines work well together, which .. I have no idea if they do or not, but I guess they might. Machine learning is going through endless data to reach a point and then go again and quantum computers can be exceptionally good at that. This is where the next decade or so gets really interesting.

From what I managed to find they use two key properties of quantum mechanics, superposition and entanglement, superposition means that a particle can be in two states at once till you verify then it'll choose one state, entanglement is a state where two particles are in entangled basically if you separate them and check the state of one the other will instantaneously in the same state regardless of the distance.
While I can see now some use, but the algorithms need to be very unique and the whole system should be in a very controlled environment, because any interference will interfere with the particle state thus ruin the result.

 

[Easo]

I doubt this technology will be available to the average Joe, do you? It's clearly in the hands of those running / or are part of the circus, hence my reply.

Just a question of time until quantum computing goes into mainstream (unless we find out that this all is big pipe dream and we cannot make it work the way we thought we can). Then yes, the average script kiddie will have it at his home, just like the ordinary PC's became mainstream in a relatively short period of time.

 

[Hemmingstamp]

So saying "It's clearly in the hands of those running" and therefore "this technology will (not) be available to the average Joe"

... isn't the same as saying there is a huge conspiracy around this stuff? Do you really know how many people would have to be involved for this to happen? Besides the fact that people can still figure it out as I explained in my first post? Sorry that I'm skeptical... Because things like this are usually called conspiracy theories if you don't have anything to back it up beside the claim.

And as I said before, the overwhelming evidence, like with most conspiracy theories, is on the other side. The sane side. The more mundane side, if you like. Just the fact that they are bringing out these press releases on every little development for many years now, one of which we are literally reacting to right now, contradicts your claim directly. Plus, it's not even hard to find academic papers on this. I know it's easy to string up wild claims in any which way possible when it suits your narrative, but when someone points out the obvious flaws and tries to show you the holes in your reasoning where you just have to connect two points in a straight line, it's suddenly smearing shit on your suggestions and you can call it whatever kind of attack you like. Still doesn't make it an attack. Wasn't even hostile. Just drink some tea with a bit of Namaste.

There is no conspiracy, it's well documented the revolviong door between big tech and government exists. You're really going to have to try harder.

I'll have to play catch up on this thread later. Life is calling. Laters folks. (I will respond)

 

[Vayra86]

Besides that, encryption can be made theoretically infinite today already. It's not hard to double our 128 bit encryption, that has never been cracked before, to 256 bit, which is already being used. But you could just keep doubling and doubling it as much as you want. We just take the speed of decrypting into account, that's the only reason we don't already have 2048 bit encryption or even more. But 20-30 years from now it won't matter anyway. We'll probably just keep extending encryption to what is far, far stronger than anything any quantum computer could crack in the next 50 years, which is already the case today.

That is the point though. Its not like encryption today is really fast to decrypt. That happens only at the level of our basic 128~256 bit encryptions really. Theoretically infinite is not relevant. What is ALWAYS relevant in terms of theft whether physical or digital, is the time you need to be in and out of the building with your loot.

Access times matter, and the period of time you can keep data secure until it is decrypted also matter and those two are at odds with one another. NO, you can't just up your encryption to infinite length because it imposes all sorts of limitations on the data you are securing, with one certainty: it won't be getting faster. In the meantime however, we need more and more data and we DO need it as fast as possible. We don't like waiting and in many cases, waiting costs money.

Security is always at odds with usability. We can secure everything incredibly well. But the user experience always suffers. That goes from your simple password hash to 2FA to a myriad of other measures we implement. You can easily expect that, if we do not have sufficient mitigation against available quantum technology, you will find yourself physically attached to your data to secure it. With an implanted chip, personal codes, token logins and/or smartphone. It is already happening in many businesses, by the way, and its crossing a line between private atmosphere and work. Imagine if we had no way to rely on encryption anymore... this would become mandatory in no time.

As far as 20-30 years from now, I think its becoming clear that binary systems are reaching a practical limit and any steps beyond that are extremely costly, difficult, and time consuming. I don't expect infinite growth in binary systems like we've seen it, even for the simple fact that nodes just won't get much smaller. We are already hitting many situations with diminishing returns, while Quantum will develop exponentially. With that, a cost-benefit scenario also comes into play. If you need whole data centers full of binary systems to fight a few square meters worth of QC systems, I think its clear who's going to win the battle.

Think of it like the Cold War arms race. As long as both sides keep placing enough missiles on either side of the fence, nobody's shooting. But if one side has overwhelming odds to win, they will push the button and probably won't ever suffer the counterpunch.

 

[PowerPC]

There is no conspiracy, it's well documented the revolviong door between big tech and government exists. You're really going to have to try harder.

Well, if you are going to ignore the logical arguments and just keep repeating your claim, again and again, this discussion is over.

That is the point though. Its not like encryption today is really fast to decrypt. That happens only at the level of our basic 128~256 bit encryptions really. Theoretically infinite is not relevant. What is ALWAYS relevant in terms of theft whether physical or digital, is the time you need to be in and out of the building with your loot.

Access times matter, and the period of time you can keep data secure until it is decrypted also matter and those two are at odds with one another. NO, you can't just up your encryption to infinite length because it imposes all sorts of limitations on the data you are securing, with one certainty: it won't be getting faster. In the meantime however, we need more and more data and we DO need it as fast as possible. We don't like waiting and in many cases, waiting costs money.

Security is always at odds with usability. We can secure everything incredibly well. But the user experience always suffers. That goes from your simple password hash to 2FA to a myriad of other measures we implement. You can easily expect that, if we do not have sufficient mitigation against available quantum technology, you will find yourself physically attached to your data to secure it. With an implanted chip, personal codes, token logins and/or smartphone. It is already happening in many businesses, by the way, and its crossing a line between private atmosphere and work. Imagine if we had no way to rely on encryption anymore... this would become mandatory in no time.

As far as 20-30 years from now, I think its becoming clear that binary systems are reaching a practical limit and any steps beyond that are extremely costly, difficult, and time consuming. I don't expect infinite growth in binary systems like we've seen it, even for the simple fact that nodes just won't get much smaller. We are already hitting many situations with diminishing returns, while Quantum will develop exponentially. With that, a cost-benefit scenario also comes into play. If you need whole data centers full of binary systems to fight a few square meters worth of QC systems, I think its clear who's going to win the battle.

Think of it like the Cold War arms race. As long as both sides keep placing enough missiles on either side of the fence, nobody's shooting. But if one side has overwhelming odds to win, they will push the button and probably won't ever suffer the counterpunch.

I'm also not in the boat of saying that QCs won't replace binary systems in 50 years. A known computer scientist (forgot his name) said not long ago on the Lex Fridman podcast that QCs are about at the same level of development today as computers with vacuum tubes were 70 years or so ago. Computers have increased in transistor capacity literally tens of millions of times since then. This is what will probably happen with QCs when the cat is out of the bag so to speak and when many more competing companies start developing this technology.

To the encryption part. Obviously, you'll have to live with some kind of usability penalty for encrypting data at higher and higher levels of security. Or as lexluthermiester said, there are probably different ways to encrypt data than the current system, which QCs won't even be able to crack. I'm not an encryption expert because only very few people actually are. It's one of the hardest subjects to roll on your own. It's very math-heavy but I believe some people will figure it out. All the cries about QCs breaking encryption seem to be only about current day encryption. For less sensitive data, you'll probably still be good just using a higher bit encryption and dealing with higher decryption times, which for example wasn't that big of a deal going from 128 bit to 256 bit at least. I don't think people ever noticed the switch since you can also combine it with increased usability in many other areas at the same time.

 

[$ReaPeR$]

Skynet gets closer by the day ..

 

[R-T-B]

Don't know what kind of conspiracy theory that is. By that logic, we wouldn't have access to modern digital CPUs because they're so much more powerful than any mechanical computation devices they had 100 years ago. Doubt it or not, if it's out and it's that powerful, it'll be out.

It requires absurd near absolute zero cooling and that's not likely to change.

What else are they (IBM and others) developing it exactly for in your mind?

Cloud (big money) and governments.

By that time encryption will become so strong and ubiquitous that quantum processors will have only a marginal advantage.

Encryption is a slow moving target. Most of the standards we use were ratified decades ago.

 

[PowerPC]

It requires absurd near absolute zero cooling and that's not likely to change.

That's why the first QCs will be part of large data centers, available in the cloud, and not just for governments.

They also said computers will never be smaller than a large closet.

Cloud (big money) and governments.

Yes, it will be only in the cloud at first. Doesn't change much. Amazon, Google and IBM will offer this technology as soon as it's out. That's my point. And they aren't just developing it for governments, why should they? I don't think they are piling so much money into the development just to make money off of governments in the end and cut off the other half of paying customers. If you have the money, they want to sell it to you. You already know how a business works.

 

[R-T-B]

Yes, it will be only in the cloud at first. Doesn't change much.

Cloud is not average joe.

They also said computers will never be smaller than a large closet.

That wasn't a fundemental law of physics though. Quantum states are only manageable at extremely low temps.

 

[PowerPC]

Cloud is not average joe.

Well, then I'm the counterexample. I'm using the cloud to host all my sites. In fact, that's how most internet businesses are getting started today. Average Joes who use the cloud. Netflix started out with a couple of Joes using the cloud to host videos and are still doing it. That's just one "small" example.

That wasn't a fundemental law of physics though. Quantum states are only manageable at extremely low temps.

Today, they are only manageable today with extremely low temps. It's kind of strange to assume there could never be some new way to do this. They said many things in quantum physics were impossible just ten years ago. Some of those things are done routinely today.

 

[R-T-B]

Well, then I'm the counterexample. I'm using the cloud to host all my sites

That's average joe cloud. And where the encryption-breaking qubits will be going will, at least at first, not be there.

You're talking something that there might be two or three machines capable of doing when we hit that milestone. They are NOT going to rent it to you.

 

[PowerPC]

That's average joe cloud. And where the encryption-breaking qubits will be going will, at least at first, not be there.

You're talking something that there might be two or three machines capable of doing when we hit that milestone. They are NOT going to rent it to you.

It's not even relevant who the first QCs will be rented to because they will be so primitive at first. Sure, the Harvard Mark II (the first supercomputer I believe) wasn't rented out to average Joes. But does it matter? It's hilariously slow compared to any computer I can carry in my pocket today. When these machines are actually worth it and as I said, "the cat is out of the bag", you'll DEFINITELY be able to rent them. Again, Google, Microsoft, IBM, or ACME Corp. all just want to make money with this and they aren't government property yet as far as I know. There are so many potential applications for this technology, that keeping it away from 99% of the population is utter non-sense and borderline a certain kind of conspiracy thinking.

 

[lexluthermiester]

Encryption is a slow moving target.

Not really. Encryption development has become increasingly important and momentum has jumped forward.

Most of the standards we use were ratified decades ago.

The sub 256bit variants, yes. 256bit and above have had recent changes and improvements. The algorithms themselves are in a constant state of refinement and improvement.

 

[R-T-B]

Not really. Encryption development has become increasingly important and momentum has jumped forward.

The sub 256bit variants, yes. 256bit and above have had recent changes and improvements. The algorithms themselves are in a constant state of refinement and improvement.

I know, remember I work this field. However, nearly none of these developments has seen widespread deployment or even minor deployment, hence my concern.

 

[Yukikaze]

Quantum computers pose a threat to current encryption schemes, as mentioned in this thread, somewhere in post-2030, but this date is being extended by doubling the length of keys used (RSA2K->RSA4K->RSA8K->RSA16K). Of course, fighting exponential growth (as expected, a Moore's Law equivalent for Qbits is already presenting itself) with linear means will not last long (at most 1 or 2 doubling cycles, so 3-5 years, depending on just how fast the progress is), but it can allow current technology to remain effectively resistant for another half-a-decade, which is what we (the industry, and the rest of humanity, as an extension) are currently banking on.

Theoretical work on quantum-resistant (or even quantum-proof) cryptography has been undergoing for a long time. Currently NIST is in the process of selecting its post-quantum recommendations, so we, as humanity, will have the capability to replace our cryptography with algorithms that remain valid even in the face of multi-mega-Qbit quantum computers. Some hardware vendors are already preempting this process and adopting some of these (or alternates) for their own security mechanisms.

What will be interesting is to see what kind of data that has been captured, stored, but could not be decrypted will become readable. Stuff that was encrypted 20 years ago may still have a lot of value (consider how many secrets and technologies are kept classified for dozens of years AFTER they were recorded). This will affect both things like national security, but also privacy, and has impact on the criminal underworld as well (depending on the statute of limitations for certain offenses where evidence may be encrypted).

As for this technology being accessible to the average joe: It will be. Heck, it already is. Log onto IBM's Quantum experience, put together a small quantum circuit and run it on one of their quantum computers. For free. Today. This is the foundation of Quantum-Computing-as-a-Service. Yes, the government will have access to the higher-end (read, higher Quantum Volume) systems first, but this is not going to be something that enterprises, academics, and even private citizens have no access to. This access is how companies intend to make money off this, after all.

 

[R-T-B]

As for this technology being accessible to the average joe: It will be. Heck, it already is. Log onto IBM's Quantum experience, put together a small quantum circuit and run it on one of their quantum computers. For free. Today.

And break conventional encryption? You MIGHT need a bit more qubits.

This access is how companies intend to make money off this, after all.

And my whole point is average joe is not the target audience.

 

[Yukikaze]

Yes, of course you aren't breaking RSA with 20 qbits. But there is nothing to indicate that as the Quantum Volume increases these machines won't be just as accessible as free AWS instances, or at a relatively low cost.

The average joe is never the audience for anything, unless it can be carried in a pocket and display ads/watch netflix. That doesn't mean the technology is not accessible for hobbyists, lone developers, independent researchers, and other people who aren't governments and multi-billion dollar corporations. These aren't quite the average joe, but they are still common people like you and me.