Scott Aaronson has a very nice piece in yesterday’s New York Times entitled “Quantum Computing Promises New Insights, Not Just Supermachines”, which I’d encourage any non-specialist with an interest in quantum computing to read. He does a great job dispelling myths about quantum computing and explaining why we should be interested in the quantum computing.
I agree with virtually everything Scott says in the article, and I think he has done a fantastic job of accurately explaining quantum computing and its merits without resorting to the kind of fudges found in many popularizations. The one minor point I would disagree with Scott on is where he suggests that decoherence is the limiting factor that currently prevents from building large scale quantum computers. In my view the situation is more complicated, and depends on the physical implementation you have in mind. Several implementations have demonstrated all the unitary operations for fault-tolerant computation with accuracies significantly exceeding what is necessary for fault-tolerant computation. However, these systems usually suffer from other problems such as an inability to individually address and manipulated qubits or difficulties with cooling. Indeed it is problems with cooling to a fixed (non-random) initial state that has limited liquid state NMR quantum computing, which provided the platform for many of the early demonstrations of quantum computing. This is only a very minor point though, and does not at all detract from Scott’s excellent article.