I’ve spoken and written repeatedly about random number generation. In computing, it underlies several critical fields, including cryptography and simulation. Given the considerable limits of generating random numbers on a deterministic, classical computer, we take for granted that physical sources of randomness must be truly random.
Not so, according to a paper discussed at the Scientific American web site. The author of the paper, Antonio Acín, a physicist at the Institute of Photonic Sciences in Spain, clarifies that many physical systems could be predicted with sufficient information and compute time. He details what I prefer to think of as a quantum random number generator where that is not the case. The description of the experimental rig sounds very similar to the components used in quantum computers.
It makes sense that a probabilistic system, like one made of vacuum tubes and ion traps all geared towards preserving the quantum characteristics of particles, would much better defy attempts to predict the series of numbers generated. This prediction is the primary means of foiling security systems based on classical random number generators.
The efforts of Acín and his colleagues builds on an existing field, one of which I was unaware, bringing similar observations about degrees of randomness to quantum systems. Many already existing for producing random numbers but of varying quality. The new rig operates with provable randomness, acting as a black box that does require the same sort of tear down other devices might to be sure of the quality of its output. The biggest hurdle for using the technique in commercial applications is its efficiency. Since it uses similar apparatus to quantum computers, maybe recent improvements in that field may prove helpful here.