The Case Against Quantum Computing
spectrum.ieee.org
spectrum.ieee.org
> So the number of continuous parameters describing the state of such a useful quantum computer at any given moment must be at least 2^1,000, which is to say about 10^300. That's a very big number indeed. How big? It is much, much greater than the number of subatomic particles in the observable universe.
> To repeat: A useful quantum computer needs to process a set of continuous parameters that is larger than the number of subatomic particles in the observable universe.
> At this point in a description of a possible future technology, a hardheaded engineer loses interest.
Comparing bits of information to number of particles in the universe is a fancy card trick to impress students, but to pretend that 1kb of state makes "hardheaded engineers" lose interest is laughable. People will keep trying because it's the holy grail for certain types of computational power that will still be there even as Moore's law wanes. For fields like bioinformatics and other natural sciences, it's probably our only way to get reasonable sized simulations regardless of worldwide digital processing power.
It's the computing equivalent of fusion, we know it can be done and that's enough to keep trying, whether we get there this decade or next century is somewhat irrelevant while the impetus remains.
It's not though? If we were actually storing the parameters describing the state we'd need 8x10^300 bytes = 8x10^288 Pb (assuming each "continuous parameter" is stored as a double precision float).
I do think the author exaggerates the difficulty though. The impossibility of even storing a complicated quantum state on a classical computer just shows how different the world is from a classical computer. The world is quantum. The state space of the world cannot be stored on or manipulated by a classical computer.
However, the authors point does show that if we build a quantum computer you will not be able to place it in every possible state (like you can with a block of memory in a classical computer). The space of states in a qc is just too large. That's very different from computer systems we're used to.
Edit: I found this interesting: https://www.hpcwire.com/2019/01/09/the-case-against-the-case...
IBM's at the 400 qubit mark, and claims to be on-track to deliver 1k within a year. If this were to happen, it'd be like going from the Gemini Guidance Computer to a 386 in 1 year. This means HFT platform$ will be able to use more near-real-time simulation capabilities to gamify the markets to capture greater tranches of total capitalization. I don't see this trend ever declining except when wandering into increasingly-frequent longtail, LTCM-like collapses whenever externalities of the real world catch up to the so-called "efficiencies" of an under-regulated casino masquerading as a bazaar.
https://newsroom.ibm.com/2022-11-09-IBM-Unveils-400-Qubit-Pl...
this literally isn't even known for a fact
>On the relationship of BQP to NP, little is known beyond the fact that some NP problems are in BQP
"In the physical world, continuous quantities (be they voltages or the parameters defining quantum-mechanical wave functions) can be neither measured nor manipulated exactly."
"Could we ever learn to control the more than 10^300 continuously variable parameters defining the quantum state of such a system? My answer is simple. No, never."
These were exactly the main arguments against quantum computing in the 90s. Quantum error correction solved them, by observing that errors can be thought of as discrete, much like a digital computer. It's subtle, but it's also covered in basically every textbook on quantum computing (Nielsen and Chuang is particularly readable).