Hardware savings also stems from careful implementation of the operation known as CA does not a gate, which is performed during error correction. Amazon researchers showed that CA does not work is not disproportionately introducing errors essentially overturns. This meant that after each round of error correction, the quantum computer continues to mainly make errors in phase flip, so that a simple, hardware correction code could still use.
AWS started working on the design for the ocelet on 2021. years, says painter. His development was a “complete problem”. To create high-quality Qubits that could ultimately make errors, researchers had to understand a new way to cultivate Tantalum, which are made of silicone chips, in silicone chip with less atomic damage.
It is a significant progress that AWs can now create and control multiple cats in one device, Puri says. “Any job that goes according to the scaling of new types of Qubitsa, I think it’s interesting,” she says. However, there are years of development. Other experts predicted that quantum computers will require thousands, if not millions, quubits to perform a useful task. Amazon’s work “is the first step,” Puri says.
He adds that the researchers will need further reducing the stake of errors due to important turns while jumping on the number of Qubits.
However, this announcement signifies Amazon’s path forward. “This is the architecture we believe in,” says a painter. Previously, the company’s main strategy should continue conventional pretonal quubits like Google and IBM, and they treated this Cat Qubit project as “Skunkworks,” he says. Now they have decided to give the priorities of the cats of Qubits. “We really convinced that it was supposed to be our main engineering effort and we will still be doing some research things, but this is the direction we go.” (Starting Alice & Bob, based in France, also build a quantum computer made of cats.)
As, the ocelat is basically a demonstration of quantum memory, says painter. The next step is to add multiple Qubits to chip, coder more information and make real calculations. But they have a lot of challenges in front, from the way you fasten all the wires to connect more chips together. “Scaling is not trivial,” he says.
