A team just squeezed twice as many reliable quantum bits out of the same hardware

[Dark Hawk]

Researchers from QuEra, Harvard, and MIT published a result this week showing they could encode more than 1,100 reliable logical qubits using just over 2,300 physical qubits, a ratio of roughly 2 to 1. To understand why this matters: most current approaches need anywhere from hundreds to thousands of physical qubits just to make a single reliable one, because quantum bits are fragile and errors pile up fast. The team achieved error rates so low they land in what researchers call the teraquop regime, meaning roughly one mistake per trillion operations. This is still a simulation result rather than a live hardware demonstration, but the method is specifically designed around how the best neutral-atom machines already work.

QuEra, Harvard, and MIT Demonstrate 2:1 Physical-to-Logical Qubit Ratio - Quantum Computing Reportquantumcomputingreport.com

[FrostBear]

Two physical qubits per logical qubit sounds almost too good. Current approaches need hundreds. This is a big deal if it holds up on real hardware

[SilverRider]

One error per trillion operations is wild. That is the kind of reliability you need before you can run anything actually useful

[MJF]

Still a simulation. Let us see them run this on the actual machine before the headlines start writing checks the hardware cannot cash

[Ellie_28]

The fact that the code is designed around how neutral atom hardware already moves qubits around is smart engineering not just clever maths