Although quantum processors exist today, they are still a long way off from becoming practical replacements for classical computers. This is due to many practical considerations, not the least of which are factors such as the need for cryogenic cooling and external noise affecting the system necessitating a level of error-correction which does not exist yet. To somewhat work around these limitations, IBM has now pitched the idea of a hybrid quantum-classical computer (marketed as ‘quantum-centric supercomputing’), which as the name suggests combines the strengths of both to create a classical system with what is effectively a quantum co-processor.
IBM readily admits that nobody has yet demonstrated quantum advantage, i.e. that a quantum computer is actually better at tasks than a classical computer, but they figure that by aiming for quantum utility (i.e. co-processor level), it could conceivably accelerate certain tasks for a classical computer much like how a graphics processing unit (GPU) is used to offload everything from rendering graphics to massively parallel computing tasks courtesy of its beefy vector processing capacity. IBM’s System Two is purported to demonstrate this when it releases.
What the outcome here will be is hard to say, as the referenced 2023 quantum utility demonstration paper involving an Ising model was repeatedly destroyed by classical computers and even trolled by a Commodore 64-based version. Thus, at the very least IBM’s new quantum utility focus ought to keep providing us with more popcorn moments like those, and maybe a usable quantum system will roll out by the 2030s if IBM’s projected timeline holds up.
it was about 25 years ago i went to a colloquium jointly hosted by the physics and CS departments. my physics knowledge is rudimentary but i perceived such an abundance of handwaving in the presentation that i was confident no one actually had an idea how ‘superposition of states’ and ‘prime factorization’ are equivalent. so i feel pretty smug seeing how the last quarter century of development has gone :)
but it reminds me of another presentation i saw in the same era. analog computing. basically within the realm of classical electromagnetism there are a lot of circuits you can make out of simple components that can represent simple equations, some of which are quite thorny. it’s faster and less power hungry and doesn’t accumulate numerical error the way that like discrete integration / differentiation does. on the other hand, i don’t know if anyone actually wants to solve large numbers of the particular differential equations that are accessible that way. but you could imagine analog computing having a lot of different use cases, like synthetic neural networks.
so i think there is a lot to be explored there and i would consider quantum computing as just a subset of analog computing — a different set of equations that can be evaluated.
the other angle is that this sort of research is exposing a much larger audience to the tools necessary to ponder the really deep philosophical questions raised by quantum mechanics. i feel like the question is still undecided: how spooky is it? i have seen 3 explanations of quantum mechanics: the interactions themselves are spooky; the interactions are classical but the way god (??) decides which interaction happens is spooky; or everything is classical but measuring and statistical effects make it impossible for us to properly understand the state of the system before the experiment.
i don’t know how to differentiate these possibilities in an experiment but the more people playing with spooky effects, the more likely someone stumbles onto a more fruitful way to think about it.
I like your explanation of the scenario. I recall the hubbub as well and have a similar feeling about the current state of it all. Hopefully its first task is world peace lol. It it is really so smart then do that. Otherwise, I just feel like it is another vaporware MLM thing where no one can explain what it is what it does or why it is good kinda like crypto currency. Garbage in garbage out.
Yeah I agree, that goal is appropriately BS for a BS machine
“the really deep philosophical questions raised by quantum mechanics. i feel like the question is still undecided: how spooky is it?”
It’s not spooky at all. The spookiness comes from trying to shoehorn a preconceived idea of the way the world works into it and using limited vocabulary to do it.
It’s just wave mechanics of a field with quantized interactions. A state decays, it generates two entangled photons: the spookiness is thinking “how are these two things I call photons linked” rather than “I interacted with the field and created a restricted 2 particle state.”
The reason it gets poorly communicated is that sometimes it’s easier to think of it as 2 particles entangled, and you pass the entanglement around, etc. like it’s something real. But it’s not – it’s just your knowledge that the original state created (the original entanglement) wasn’t a “full” two particle state – it had an internal correlation.
any understanding of the universe sufficiently divorced from what i learned playing with blocks as a kid is spooky :)
So let me get this straight. IBM says no one has proven a quantum computer to be superior to a traditional computer but are willing to sell you a quantum co-processor. That is going to be an interesting marketing challenge. I hope IBM has a killer application ready or it is going nowhere.
“I hope IBM has a killer application ready”
The killer application is “learning how to work with a quantum computer.”
The killer application for quantum computing: “Quickly crack end-to-end public key cryptography that has a significantly large key”. If you are the first to prove you can accomplish that, either the world is your oyster, and/or your life expectancy is rather short.
“Whistler, get me the diagnostics”
TOO MANY SECRETS
And Terry gets the extra marks for the perfect film reference. My observation based on experience is, the current risk is that you end up needing a multi-petaflop GPU cluster just manage the quantum error corrections, to get the quantum computer just to stay up long enough to do something useful – so is that quantum advantage, or not?
Why the heck is everyone so focused on breaking cryptography whenever they hear “quantum computers”? It’s a total red herring.
We’ve known about cryptographic algorithms that are still NP-hard with quantum computers for decades, quantum cryptography is secured by the known laws of physics, and even current cryptography schemes are nasty enough that leveraging a quantum computer to break them is just not going to be practical.
Think about it. We don’t actually care about factoring large numbers. No one runs into your office and says “quick, I need the prime factorization of 83251098!”. The only reason we care about it is we use it for securing transactions, and the only reason we do that is because it’s currently hard. If it’s not hard… no one cares about it anymore. So no, that’s not a “killer app” for a quantum computer, it’s a “killer app” for new cryptography providers.
If you are talking about a killer app for a fully working, large number of qubit quantum computer, it’s simulating quantum systems. That’s what you need them for.
“quantum-centric supercomputing’), which as the name suggests combines the strengths of both to create a classical system with what is effectively a quantum co-processor.”
I respectfully disagree. For me “quantum centric” means quantum system with classical coprocessor – quantum should do most of the work or at least significant part of it. Since quantum gives no real advantage we should call it “quantum supported/aided/accelerated super computing” right?
“For me “quantum centric” means quantum system with classical coprocessor – quantum should do most of the work”
It’s in the article. The issue is simulating a quantum system, not necessarily outperforming a classical computer on a classical task. Stuff like “let’s factor this big number” is a quantum computer trying to act like a classical computer, but fast: this is different from simulating quantum systems, which is a classical computer trying to act like a quantum computer, but slow.
The whole idea of thinking about a quantum computer as something that can do classical math tasks somewhat misses the point – you don’t need a classical math problem to make a quantum computer useful, you have the entire Universe to simulate.
That’s cool if that is what “quantum centric” means to you.
But that would be a really bad computer.
Most things we use a computer for are stuff that a Qbit is either bad at, or can’t do at all. The quantum part SHOULDN’T be the main part of the machine.
It is similar to (though still very different than) how a GPU is good for certain tasks, but unsuitable for others.
You probably COULD write an OS that ran entirely on a GPU instead of a general purpose CPU, but it would be worse at being a “computer” in many ways, because a GPU does things in a very different way than a CPU.
snort It’s due to more than “practical implementation issues”.
That’s like saying we don’t build our homes out of clear glass because there are “practical implementation issues”.
Most of what we use a computer for isn’t what a Qbit is useful for.
Even if we COULD build a fully “general purpose” computer out of Qbits, we wouldn’t have a good reason to.