Sabine
patreon
Quantum Computers cross 1000 qubits threshold. What does this mean?
🕑 Added 2023-12-17 16:00:08 +0000 UTC[This is a transcript.]
IBM just announced they produced the first quantum processor with more than 1000 qubits and I got a lot of question about what this means, so let’s have a look.
Last week, IBM revealed the IBM Condor processor which has 1,121-qubit. This is up from 433 qubits in its previously largest chip and in line with the quantum roadmap that IBM has set for itself.
Qubits are the operating units of quantum computers. If you manage to entangle them, you can use them to perform certain calculations dramatically faster than on conventional computers. However, the number of qubits in and of itself isn’t all that relevant.
You see, the type of qubit that IBM is using is tiny superconducting circuits. They’re etched on silicon wafers similarly to normal microprocessors. Lumping many of them together isn’t really the problem. I mean, not like I can do it, but IBM has been doing stuff like this for decades.
No, the question is whether you can actually operate all these qubits as desired. In particular you want to know how faithfully they perform operations and how large the error rate is. IBM hasn’t released data on that, other than somewhat vaguely saying the new processor has a “similar performance” as their previous quantum processor with 433-qubits.
Indeed, it’s odd how little IBM has been saying about either this new chip or the previous one. They’ve preferred talking about an older, even smaller chip with just a little more than 100 qubits.
This is even odder if you take into account that in 2018 people from IBM argued that the number of qubits is not a good measure of performance and introduced a better measure called the “quantum volume” that would be more suitable. But now they seem to not want to use this measure. their quantum volume updates end in 2022 with a chip that has merely 16 qubits.
So what can you do with this new 1000 plus qubit chip? No one really knows. But we do know that realistically for commercially interesting applications we’ll need at least a million well-working qubits, so the quantum revolution will have to wait a bit longer.
Comments
I assume that this chip is not usable by most engineers, such as the 8051 and derivatives or general purpose CPU, because it needs to be cold, right? I can create an embedded system with regular semiconductors in my home lab, but haven't the capability of using this except in a fully funded research or commercial facility due to the cost of specialize gear, such as super-cooling, right? Their article doesn't mention temperature requirements!
Or be developed at some advanced government-run laboratory at taxpayers' expense, but not privatized to make some who are rich super rich in the easiest possible way. If they do, as it does happen, that would be bad politics, not good governance.
D Brown
/s: In a related post, IBM's Vaporware™ division has reported it is operating at full capacity. I believe QC, Fusion, and room temperature superconductance will have similar ripple effects when they truly arrive in that the first company that develops a viable scalable version of the tech will be on its way to becoming the richest company on the planet, or be nationalized by a government.
