These machines do not provide any computational speedup, we have plenty of papers coming from academic physicists clearly showing this. Why these people insist with their narrative is beyond my comprehension at this stage (maybe, artificially increase the value of their market shares?)..
P.S. I was involved in semiconductor physics and quantum computing way before it became a buzz word.
Yes but if you’re involved in physics you know that physics papers are rich in discourse with correct and incorrect ideas alike.
With Phoenix IBM has a system with over 100 qubits close to 99.9 2Q Fidelity and though that does not sound like much it’s already enough to run operations that can’t be computed classically. The system is so well built that most of these qubits can perform well without crosstalk errors hindering execution, and with phoenix they solved mid circuit reset which didn’t work well before, blocking experiments for meaningful runtime error correction.
I also have to add that, so far, we know only a handful of algorithms which might run on a general quantum computer (if we had one!). Among those, only two performs something which would actually be useful (Grover and Shor). And we have only those, not for a lack of trying.. it's because coming up with quantum algorithms is intrinsically complicated (it's definitely not like writing an algorithm for a classical machine). So those quantum bits could be, in theory, the most stable thing in the universe, I'd still see them as irrelevant in everyday applications.
I agree with the “in everyday applications” qualifier. But there are plenty of applications that are research focused where having access to a proper quantum computer would be useful. As a Quantum Circuit fundamentally models a Unitary transformation on some quantum state it is hugely useful to model the time evolution of quantum systems which are experimentally difficult to access using a quantum computer.
Oh the ordering and lack of commas in my phrasing made it my point unclear. What I mean is: using a quantum computer enables simulating quantum systems that can otherwise not be simulated in a reasonable time (using a conventional computer) and that is especially problematic if those systems can’t be experimentally looked at easily.
The trouble is that greed is a force right now which obscures in both directions. If a company didn’t have an algorithm set and hardware that will print billions they might imply they have it. If they did they might imply they don’t for fear of competition. Likewise government customers want to to enjoy extending their harvest now decrypt later investments and pressure companies to say less, for example Google and ECDSA.fail
Because peeps are doing HNDL (harvest now, decrypt later) today. They are intercepting your traffic and storing it until they can break crypto. They're doing that today. And they've been doing it for a while.
There are contexts in which that's an unacceptable level of risk.
HNDL can be worth it without the "threat" of quantum computing; for instance we see currently a lot of interest in AI-assisted math, and maybe it could lead to a breakthrough that affect crypto algorithms that are currently considered safe.
Therefore it seems weird to use QC as a benchmark. Rather, if your communications being decrypted in the future is unacceptable, then the logical conclusion is that you must not transmit it in such a way that a third party can intercept it.
Admittedly, it is inconvenient or even impractical, so people do accept the risk anyway without admitting it. They are sweeping the issue under the carpet of supposedly QC-proof crypto - because new quantum algorithms can be discovered too.
I'm personally more worried about flaws in the quantum-safe crypto schemes than in RSA, which is based on the age-old factorization problem of prime numbers.
They can harvest all they want, but if no one ever invents a usable, practical and scalable quantum processor all this will be moot.
What are the odds that a new QS crypto scheme has a fatal flaw or is otherwise intentionally backdoored?
Amping everyone up with fears over quantum breaking classical seems like a fantastic way to get another round of vulnerabilities injected into everyone's infrastructure.
I don’t understand who the target audience of these ads is. For most people, including myself, this is like seeing an ad for “New and Improved Time Machine! (Untested)” or “Magic Beans - Indica”.
These ads are aimed at the board members of IBM's customers. IBM has been doing this for ages. I might think IBM is a high margin body shop with a veneer of innovative-ish projects like Watson that are almost comically out of step with the rest of the tech industry. But that's not the golf course discourse.
The purpose is to make it safe to remain an IBM customer.
And I'd add that there are a ton of quantum startups out there. Maybe they are all idiots. But, if you're IBM with a large research organization that's been working on some related foundational technologies for literally decades, it's not an obviously wrong-headed move for IBM Research to be working in this space.
Everyone who is curious? I mean did you know that you can run something on a real quantum computer for free? Go try out something new, or are we only getting excited about the newest AI slop machine release these days?
They are currently targeting more material science, such as quantum chemistry for drugs discovery or new material discovery. They can’t and do not aim to do prime factorization on those.
Currently the two qubit gate error rate is listed as ~2.7e-3 (see https://quantum.cloud.ibm.com/computers?system=ibm_phoenix but note it will vary from day to day). That implies you can expect ~400 entangling gates before experiencing an error. A zero-shenanigans factoring of 21 takes ~2000 entangling gates. So it's not good enough to factor 21 in any meaningful sense.
Really what matters about this chip is that it is better suited for error correction. They have a square grid connectivity (previous IBM chips incurred fatal overheads from their heavy hex connectivity), they can actually reset their qubits now (previous IBM chips couldn't scale to long running computations because once leakage arose nothing could remove it except waiting a long time), and they have tunable couplers instead of fixed frequency ones (previous IBM chips were just doomed whenever some problem happened to land on their operating area in frequency space).
Most interesting quantum computations take millions or billions of gates. Factoring classically intractable numbers takes tens of billions of gates. No one is going to do that many quantum gates without error correction. Whenever a quantum computing company says anything other than "it made error correction better" or "it will make error correction better", you can ignore that. It's just some side quest. Maybe it's interesting in and of itself, but it's not what matters for progress.
FWIW, Scott Aaronson has likened the question on factorization of concrete instances to what happened during the Manhattan Project:
> Joe #11: Then you’ll be waiting a long time, because that’s a stupid metric. (Or rather: anyone willing to use that metric today, is probably someone you shouldn’t trust.) It’s almost exactly analogous to judging the Manhattan Project in 1942, 1943, 1944 by the metric “how big an explosion can you make, today?” I.e., it’s something that’s going to be basically flat for a long time and then undergo a step change, under the assumption that everything is progressing as it should—because you need a critical mass in the one case and to exceed the fault-tolerance threshold in the other.
I agree and I was being facetious since I know the answer is nothing much. The point is that IBM write all this breathless marketing copy that seems detached from the actual utility of the device.
Although I believe there's a difference with the Manhattan project. The original atom bomb was "useful" even at its small scale. Whereas there's likely to be a lot of time between a quantum computer that can factor 69 and a quantum computer that can factor RSA 2048.
> IBM write all this breathless marketing copy that seems detached from the actual utility of the device
What are the actual utility according to IBM? Did they say you can order one today on their website? Or even "call us to get the pricing info"? Are there case studies of what their customers use it for? Or only potential uses?
> Although I believe there's a difference with the Manhattan project. The original atom bomb was "useful" even at its small scale.
No it was not useful, otherwise it'd have been paraded and hyped up on the media. The trinity test was in July 1945. Hiroshima happened in August. They were practically used as soon as they were deemed ready. Hundreds of thousands people died.
Quantum software engineer here - yes! I use Claude quite extensively with quantum SDKs like Qiskit and Pennylane. It tends to be very good at scaffolding and architecture, but can spiral or be outright incorrect when it comes to implementing certain algorithms like QAOA.
The real benefit comes from real-world integration - if I'm exploring e.g. a logistics application, it can source real-world data and models very quickly, and suggest quantum appropriate-formulations. Nevertheless, a lot of hand-holding is required.
Yes, I haven't looked at Qiskit for a while but I played with it for a bit when I was an IBM Quantum Ambassador. I don't know all the details off the top of my head but I know there has been AI integration work with OpenShift, Ansible, and RHEL.
These machines do not provide any computational speedup, we have plenty of papers coming from academic physicists clearly showing this. Why these people insist with their narrative is beyond my comprehension at this stage (maybe, artificially increase the value of their market shares?)..
P.S. I was involved in semiconductor physics and quantum computing way before it became a buzz word.
Yes but if you’re involved in physics you know that physics papers are rich in discourse with correct and incorrect ideas alike.
With Phoenix IBM has a system with over 100 qubits close to 99.9 2Q Fidelity and though that does not sound like much it’s already enough to run operations that can’t be computed classically. The system is so well built that most of these qubits can perform well without crosstalk errors hindering execution, and with phoenix they solved mid circuit reset which didn’t work well before, blocking experiments for meaningful runtime error correction.
There’s an open scoreboard ibm runs open to all to submit problems demonstrating algorithmic breakthroughs https://quantum-advantage-tracker.github.io/
I also have to add that, so far, we know only a handful of algorithms which might run on a general quantum computer (if we had one!). Among those, only two performs something which would actually be useful (Grover and Shor). And we have only those, not for a lack of trying.. it's because coming up with quantum algorithms is intrinsically complicated (it's definitely not like writing an algorithm for a classical machine). So those quantum bits could be, in theory, the most stable thing in the universe, I'd still see them as irrelevant in everyday applications.
I agree with the “in everyday applications” qualifier. But there are plenty of applications that are research focused where having access to a proper quantum computer would be useful. As a Quantum Circuit fundamentally models a Unitary transformation on some quantum state it is hugely useful to model the time evolution of quantum systems which are experimentally difficult to access using a quantum computer.
You mean "... using a conventional computer"? (Excuse me if you didn't mean that.)
Oh the ordering and lack of commas in my phrasing made it my point unclear. What I mean is: using a quantum computer enables simulating quantum systems that can otherwise not be simulated in a reasonable time (using a conventional computer) and that is especially problematic if those systems can’t be experimentally looked at easily.
The trouble is that greed is a force right now which obscures in both directions. If a company didn’t have an algorithm set and hardware that will print billions they might imply they have it. If they did they might imply they don’t for fear of competition. Likewise government customers want to to enjoy extending their harvest now decrypt later investments and pressure companies to say less, for example Google and ECDSA.fail
If two AIs talk to each other does the rest of the forest hear?
It's still mostly a research tool. Don't be too harsh on an early system like this. One day it (may) become widely available and used by millions.
[flagged]
I too am convinced that RSA-4096 will remain secure for the foreseeable future. Why everyone's mental about quantum-safe encryption is beyond me.
Because peeps are doing HNDL (harvest now, decrypt later) today. They are intercepting your traffic and storing it until they can break crypto. They're doing that today. And they've been doing it for a while.
There are contexts in which that's an unacceptable level of risk.
HNDL can be worth it without the "threat" of quantum computing; for instance we see currently a lot of interest in AI-assisted math, and maybe it could lead to a breakthrough that affect crypto algorithms that are currently considered safe.
Therefore it seems weird to use QC as a benchmark. Rather, if your communications being decrypted in the future is unacceptable, then the logical conclusion is that you must not transmit it in such a way that a third party can intercept it.
Admittedly, it is inconvenient or even impractical, so people do accept the risk anyway without admitting it. They are sweeping the issue under the carpet of supposedly QC-proof crypto - because new quantum algorithms can be discovered too.
[dead]
I'm personally more worried about flaws in the quantum-safe crypto schemes than in RSA, which is based on the age-old factorization problem of prime numbers.
They can harvest all they want, but if no one ever invents a usable, practical and scalable quantum processor all this will be moot.
It would be ironic if the holes in the quantum-safe crypto meant that they were weaker than the old school options.
Generally people encrypt with both post-quantum crypto and regular crypto in series to prevent that from being a problem.
You can sell them new software because of this. ;-)
Classic odds/stakes. What do the odds have to be of a surprise quantum breakthrough before the stakes are serious? 1 in… a million? A billion?
What are the odds that a new QS crypto scheme has a fatal flaw or is otherwise intentionally backdoored?
Amping everyone up with fears over quantum breaking classical seems like a fantastic way to get another round of vulnerabilities injected into everyone's infrastructure.
[dead]
I don’t understand who the target audience of these ads is. For most people, including myself, this is like seeing an ad for “New and Improved Time Machine! (Untested)” or “Magic Beans - Indica”.
These ads are aimed at the board members of IBM's customers. IBM has been doing this for ages. I might think IBM is a high margin body shop with a veneer of innovative-ish projects like Watson that are almost comically out of step with the rest of the tech industry. But that's not the golf course discourse.
The purpose is to make it safe to remain an IBM customer.
And I'd add that there are a ton of quantum startups out there. Maybe they are all idiots. But, if you're IBM with a large research organization that's been working on some related foundational technologies for literally decades, it's not an obviously wrong-headed move for IBM Research to be working in this space.
I agree that they should be doing a press release and a separate technical deep dive.
Here they mix both and don’t get the public. It has always been IBM problem to correctly target the product to the audience.
We see it here again
Everyone who is curious? I mean did you know that you can run something on a real quantum computer for free? Go try out something new, or are we only getting excited about the newest AI slop machine release these days?
Investors
What's the largest semiprime it can factor?
You can attempt 15 yourself if you want:
https://towardsdatascience.com/where-are-we-with-shors-algor...
If you run it over and over again you will eventually get 3 or 5...
They are currently targeting more material science, such as quantum chemistry for drugs discovery or new material discovery. They can’t and do not aim to do prime factorization on those.
Good
Currently the two qubit gate error rate is listed as ~2.7e-3 (see https://quantum.cloud.ibm.com/computers?system=ibm_phoenix but note it will vary from day to day). That implies you can expect ~400 entangling gates before experiencing an error. A zero-shenanigans factoring of 21 takes ~2000 entangling gates. So it's not good enough to factor 21 in any meaningful sense.
Really what matters about this chip is that it is better suited for error correction. They have a square grid connectivity (previous IBM chips incurred fatal overheads from their heavy hex connectivity), they can actually reset their qubits now (previous IBM chips couldn't scale to long running computations because once leakage arose nothing could remove it except waiting a long time), and they have tunable couplers instead of fixed frequency ones (previous IBM chips were just doomed whenever some problem happened to land on their operating area in frequency space).
Most interesting quantum computations take millions or billions of gates. Factoring classically intractable numbers takes tens of billions of gates. No one is going to do that many quantum gates without error correction. Whenever a quantum computing company says anything other than "it made error correction better" or "it will make error correction better", you can ignore that. It's just some side quest. Maybe it's interesting in and of itself, but it's not what matters for progress.
FWIW, Scott Aaronson has likened the question on factorization of concrete instances to what happened during the Manhattan Project:
> Joe #11: Then you’ll be waiting a long time, because that’s a stupid metric. (Or rather: anyone willing to use that metric today, is probably someone you shouldn’t trust.) It’s almost exactly analogous to judging the Manhattan Project in 1942, 1943, 1944 by the metric “how big an explosion can you make, today?” I.e., it’s something that’s going to be basically flat for a long time and then undergo a step change, under the assumption that everything is progressing as it should—because you need a critical mass in the one case and to exceed the fault-tolerance threshold in the other.
https://scottaaronson.blog/?p=4121#comment-1804101
I agree and I was being facetious since I know the answer is nothing much. The point is that IBM write all this breathless marketing copy that seems detached from the actual utility of the device.
Although I believe there's a difference with the Manhattan project. The original atom bomb was "useful" even at its small scale. Whereas there's likely to be a lot of time between a quantum computer that can factor 69 and a quantum computer that can factor RSA 2048.
> IBM write all this breathless marketing copy that seems detached from the actual utility of the device
What are the actual utility according to IBM? Did they say you can order one today on their website? Or even "call us to get the pricing info"? Are there case studies of what their customers use it for? Or only potential uses?
> Although I believe there's a difference with the Manhattan project. The original atom bomb was "useful" even at its small scale.
No it was not useful, otherwise it'd have been paraded and hyped up on the media. The trinity test was in July 1945. Hiroshima happened in August. They were practically used as soon as they were deemed ready. Hundreds of thousands people died.
Their cloud offering lists this with some specific calibration data:
https://quantum.cloud.ibm.com/computers?system=ibm_phoenix
Does this mean they literally have a single one? (Or at least a single one they offer publicly)
But can you run:
IBM Bob
On:
IBM Quantum Nighthawk R2
?
Tangential question: do quantum researchers benefit from LLMs in any way?
Quantum software engineer here - yes! I use Claude quite extensively with quantum SDKs like Qiskit and Pennylane. It tends to be very good at scaffolding and architecture, but can spiral or be outright incorrect when it comes to implementing certain algorithms like QAOA.
The real benefit comes from real-world integration - if I'm exploring e.g. a logistics application, it can source real-world data and models very quickly, and suggest quantum appropriate-formulations. Nevertheless, a lot of hand-holding is required.
Yes, I haven't looked at Qiskit for a while but I played with it for a bit when I was an IBM Quantum Ambassador. I don't know all the details off the top of my head but I know there has been AI integration work with OpenShift, Ansible, and RHEL.
From IBM: "The Quantum Nighthawk Ultimate Extreme Subatomic Interstellar Whizbang Mark II"
Also from IBM: "Bob"
they made the pic look like it can be installed on an intel motherboard.