Will QC make LLMs faster/better?
The answer is Yes and No
Until you look; then it collapses to one or the other.
Microsoft doubles down on controversial quantum computing claims
41–50 of 61 posts
Re: Microsoft doubles down on controversial quantum computing claims
#42Earlier quoted context omitted.
Any kind of quantum workload would live in servers since their capabilities will likely never be latency sensitive. I think what is more interesting is finding a way to utilize the fact that quantum information is capable of traveling faster than the speed of light, obviously observation of the data is bound the same limitations as we have today, but we could scale bandwidth by encoding data and "teleporting" it to t…
No quantum process currently known allows for information transfer faster than the speed of light.
Wasn't the entire conflict behind entanglement that it appeared to have the capability to violate this princible with the only two explanations being: quantum state travels faster than the speed of light or we're in a many-worlds universe where state propagates backwards which means the information was there from the very beginning?
Re: Microsoft doubles down on controversial quantum computing claims
#43Earlier quoted context omitted.
It can't become mainstream, it is a very narrowly specialized hardware for a very limited set of tasks.
One of those tasks though is "modeling reality".
If quantum computers turn out to be able to do that, the api would be very similar to what dwave has today.
That is one will use normal computer to reformulate the task into a predefined form, submit that to quantum computer, and then use normal computer again to process the results.
Re: Microsoft doubles down on controversial quantum computing claims
#44Earlier quoted context omitted.
No quantum process currently known allows for information transfer faster than the speed of light.
Wasn't the entire conflict behind entanglement that it appeared to have the capability to violate this princible with the only two explanations being: quantum state travels faster than the speed of light or we're in a many-worlds universe where state propagates backwards which means the information was there from the very beginning?
The key word here is "information". No known quantum effect results in information being transferred faster than the speed of light (which might be more correctly known these days as a the speed limit of information). Entanglement, even at great distance, does not violate this principle as that cannot be effectively used to transfer information.
Re: Microsoft doubles down on controversial quantum computing claims
#45Earlier quoted context omitted.
No quantum process currently known allows for information transfer faster than the speed of light.
Wasn't the entire conflict behind entanglement that it appeared to have the capability to violate this princible with the only two explanations being: quantum state travels faster than the speed of light or we're in a many-worlds universe where state propagates backwards which means the information was there from the very beginning?
The historical debate around Bell's Theorem and Einstein's "spooky action" often leads to this exact confusion, but quantum mechanics has an ironclad mathematical guardrail against this: the No-Communication Theorem.
Entanglement cannot be used to transmit messages, amplify bandwidth, or achieve "infinite compression" for a few foundational reasons:
1) The Classical Bottleneck: In quantum teleportation, you aren't actually moving information through space faster than light. To reconstruct the state of a teleported qubit at the destination, the sender must transmit two classical bits of data over a standard, traditional channel (limited by the speed of light, c).
2) The Randomness Vector: Without those two classical bits, the receiver's particle looks like completely randomized entropy (a maximally mixed state). You could spin your entangled particle right now, and the person on Mars would see their particle change state instantly—but to them, it just looks like a random coin toss. They cannot know what you chose to measure or what your result was until your classical radio signal arrives to break the encryption.
3) Holevo's Bound: From an information theory perspective, Holevo's theorem proves you cannot extract more than one bit of classical information from a single qubit. While superdense coding lets you pre-share entanglement to send two classical bits using one physical qubit, it still requires physically moving that qubit through space at or below the speed of light.
Whether you favor Copenhagen, Many-Worlds, or Pilot Wave theory, the physical reality across all interpretations remains identical: local causality is never violated. Entanglement shows us that nature is non-local, but it completely forbids us from weaponizing that non-locality to send a signal faster than c.
Re: Microsoft doubles down on controversial quantum computing claims
#46We failed at AI, let see if Quantum sticks
Re: Microsoft doubles down on controversial quantum computing claims
#47Earlier quoted context omitted.
Wasn't the entire conflict behind entanglement that it appeared to have the capability to violate this princible with the only two explanations being: quantum state travels faster than the speed of light or we're in a many-worlds universe where state propagates backwards which means the information was there from the very beginning?
The historical debate around Bell's Theorem and Einstein's "spooky action" often leads to this exact confusion, but quantum mechanics has an ironclad mathematical guardrail against this: the No-Communication Theorem. Entanglement cannot be used to transmit messages, amplify bandwidth, or achieve "infinite compression" for a few foundational reasons: 1) The Classical Bottleneck: In quantum teleportation, you aren't ac…
That's the bit I was thinking about the most: "pre-share entanglement to send two classical bits using one physical qubit" I did mention that we cannot achieve data transfer faster than the speed of light, but with the use of "teleportation" we can increase bandwidth that still holds true doesn't it?
But that does clarify quite a few things thanks!
Re: Microsoft doubles down on controversial quantum computing claims
#48When quantum computing becomes mainstream, what will be the first standard programming language? I wonder what the 'C' of quantum computers will be
I feel like Qiskit is the standard nowadays, and maybe because of the early mover advantage it will prevail after NISQ (Noisy Intermediate-Scale Quantum). It's actually just a Python framework.
Re: Microsoft doubles down on controversial quantum computing claims
#49Now, they have replaced aluminum with lead to increase the strength of the binding of the pairs in the semiconductor, and the key quantum states last 20 seconds
Anyone want to give a layman-ish explanation of why aluminum versus lead makes such a huge difference?
Re: Microsoft doubles down on controversial quantum computing claims
#50Earlier quoted context omitted.
My prediction would be that it won't become mainstream. Even if it will be practically possible to build quantum computers for average users (given they currently rely on complex physical experiments, one can doubt that), there's the question of whether there's a need for "mainstream" quantum computing. As has often been said, quantum computers aren't some magical thing that makes every computation faster. They are f…
>You may be able to rent your quantum computing time if that gets cheap enough to be practical, but I doubt many people will ever own one. You already can rent time on one - IBM and others offer it - but they are not cheap.
They are not cheap, but at least they are useless.