22 Jul
2010
22 Jul
'10
2:36 p.m.
The layman's approach to solving NP-complete problems by postselection is to assert that the "many worlds" interpretation of quantum mechanics is true, bounce N photons off a semisilvered mirror, use the N-bit pattern detected on the other side as the input, and kill yourself if it doesn't work. In 2^{-N} of the worlds, you solve the problem and continue living.
So quantum computing only works if every user agrees to participate in what is essentially Russian roulette? Does the intended market know this? I can see why progress has been so slow; we keep losing all our researchers. -tom