Thursday, November 15, 2007

And now - a physical impossibility?

It's a week for strange science. For two years I've been following a company which claims its radio system can deliver information faster than the physical limits described by the Shannon-Hartley theorem. This is stranger than D-Wave's quantum claims, since what it's claiming is a physical impossibility by existing science.

The company also has a much higher valuation: its share price puts the company's worth at more than a billion dollars. And it still hasn't shared its technology, or delivered a service. This month, phones using the technology will apparently start trials in Florida, and in advance of that, a financial company has given it a glowing assessment.

I think the comments following on from Techworld blog on the subject tell you what you need to know about this one.

Tuesday, November 13, 2007

Quantum computing - keep waiting

D-Wave, the quantum computing company I wrote about in February, is back, this time with a Google scientist in tow.

They've got Google's image-search expert, Dr Hartmut Neven, showing an algorithm on their new quantum computer, at SC07, a supercomputing show in Nevada. I've written it up for ZDnet and the Guardian.

But unfortunately, the quantum computing research establishment is even more sceptical this go round. "The interest of D-Wave for me is mainly one of psychology, business and the margins of science," professor Andrew Steane, of the centre for quantum computation at the University of Oxford, told me. "My conclusion is that I suspect they are misleading themselves."

While people like me still understand the original naive view of quantum computing (millions of computers in parallel universes), real researchers who have now spent twenty years trying to produce more qubits and keep them isolated from the world for longer have moved onto subtler considerations.

I have a suspicion that "proper" quantum computing may be impossible for somewhat similar reasons to perpetual motion. We can see a theoretical possibility that we might exclude interactions with the outside world long enough to keep the qubits coherent, but there's an eventual interaction with the outside world which, like friction in a proposed perpetual motion machine, imposes a limit which might turn out to be not as far beyond conventional computers as we had hoped.

Real quantum researchers say there is a distinction to be made between an "experiment" and a "computer". A computer takes an understood process, to work out new results. In an experiment, the results themselves don't signify anything except the extent to which they confirm or deny the theories behind the experiment.

My feeling is that quantum computing experiments may not in the end produce vastly more powerful computers, but they will push forward our understanding of the nature of reality.