By Mark Buchanan PHYSICISTS turned to drugs this year to build the first quantum computer. Standard computers process data bits in the form of 0s or 1s. But things are different in a quantum device, which uses single particles to store more complex multi-states called “qubits”. An electron with spin “up” or “down”, for example, can store either a 0 or 1. But since an electron can lead a split existence, and be spin “up” and “down” simultaneously, a qubit can represent both 0 and 1 at once. This oddity makes a quantum computer enormously powerful—at least in theory. While a string of ordinary bits stores a single sequence of 0s and 1s, the coexisting branches of a string of qubits can represent every possible sequence at one go. Switch on a quantum computer, and it would splinter into a huge army of “ghost computers”, each doing its own calculation. After a decade of trying, researchers finally made the idea work in April, using atomic spins in molecules of the anaesthetic drug chloroform to store qubits. Isaac Chuang of the IBM Almaden Research Center in San Jose, Neil Gerschenfeld of the Massachusetts Institute of Technology and Mark Kubinec of the University of California at Berkeley used pulses of radio waves to flip two qubits and carry out a simple computation. Their computer did the equivalent of searching through four strings of letters such as XAEXI, AIEXX, XAXEI, XIXAE and successfully located the entry containing “XE”. Not so stunning—except the machine didn’t check the entries one by one. It inspected them all at once. But to make a really useful computer requires more qubits. If two qubits can check four items in parallel, a computer with 32 would have more than 4 billion calculating arms, and could fly through mathematical problems that bog down the fastest supercomputers. The trouble is, quantum parallelism tends to collapse if disrupted by any little bit of noise—even a stray atom. In September, David Cory and his MIT colleagues, working with physicists from the Los Alamos National Laboratory in New Mexico, took a step towards solving this problem. The most promising approach is “error correction”. By spreading information over several qubits, errors in some of them can be corrected by using the information that still resides in the others. In a simple quantum computer with three qubits, the team showed that even if all three were degraded by noise during an operation, the residual information in two could be used to bring the third qubit back to its proper state. “However, it’s a long way from three qubits to a quantum computer powerful enough to solve significant problems,” warns Peter Shor of AT&T Bell Laboratories in Murray Hill,