So, the paper describes how some logical operations could be carried out comparatively simply, whereas just a few others require manipulations of the error correction scheme so as to work. (These manipulations have names like lattice surgical procedure and magic state distillation, that are good indicators that the sphere does not take itself that significantly.)
So, in sum, Microsoft feels that it has recognized an error correction scheme that’s pretty compact, could be applied effectively on {hardware} that shops qubits in photons, atoms, or trapped ions, and allows common computation. What it hasn’t performed, nevertheless, is present that it truly works. And that is as a result of it merely does not have the {hardware} proper now. Azure is providing trapped ion machines from IonQ and Qantinuum, however these prime out at 56 qubits—effectively beneath the 96 wanted for his or her favored model of those 4D codes. The most important it has entry to is a 100-qubit machine from an organization referred to as PASQAL, which barely suits the 96 qubits wanted, leaving no room for error.
Whereas it needs to be attainable to check smaller variations of codes in the identical household, the Azure workforce has already demonstrated its capacity to work with error correction codes primarily based on hypercubes, so it is unclear whether or not there’s something to achieve from that strategy.
Extra atoms
As a substitute, it seems to be ready for one more accomplice, Atom Computing, to area its next-generation machine, one it is designing in partnership with Microsoft. “This primary technology that we’re constructing collectively between Atom Computing and Microsoft will embody state-of-the-art quantum capabilities, may have 1,200 bodily qubits,” Svore stated “After which the subsequent improve of that machine may have upwards of 10,000. And so that you’re taking a look at then with the ability to go to upwards of 100 logical qubits with deeper and extra dependable computation out there. ”
