Monday, researchers at the California Institute of Technology reported that they were able to program the pathways by which DNA molecules self-assemble, a step in a series of steps for inexpensive DNA computers. Though it may seem like science fiction, DNA computing has existed as a proof-of-concept for nearly fifteen years, when Leonard Adleman of the University of Southern California (also of the Golden state) solved the Hamiltonian Path Problem for seven points.
Since then, more capable machines (such as the Turing machine, the model of all modern-day computers) have shown to be constructible. In fact, in 2002, a DNA computer that could perform 100,000 times as many operations per second as the (then) fastest PC was created. What makes DNA computing so interesting is that parallel processing on a DNA computer is much more natural to achieve than on a regular PC. On your Windows or Linux box, multithreading (both in hardware and software, which mimics hardware) is more of a kludge and suffers a performance hit that rises polynomially with the number of separate threads created.
Intriguingly, while DNA computing does not offer a sufficient counter to computational complexity theory (other technologies such as quantum computing (which may also sound like science fiction) do provide such interesting capabilities), it does offer the speed increase and parallel processing abilities that many security solutions will suffer from as they become mainstream. In fact, many solution tactics, from brute force to more sophisticated techniques, are amazingly separable. As the price of DNA computers slide down the scale to the level of “professionals”, perhaps more novel security implementations will have arisen to cover the slack.