Space of possible problems is huge. Only a few can be efficiently solved by Digital computing (DC). We know that a further fistfull of problems can be efficiently solved by Quantum computing (QC). Unfortunately, the most interesting problem, that of fast FFT, also known as Shor algorithm, will never work for any practical purpose because of inherent imprecision of QC. On the other hand, for problems where approximate sulutions are acceptable, such as adaibatic/minimization problems, QC readily shows exquisit performance and will continue to improve in future.

The question is then: is Random pulse computation (RPC) capable of efficiently solving some problems which are hard for both DC and QC? Such as: edge detection, stutter-free (streamline) computation with fundamentally fastest response to sensory information (data/in - result/out) or learning? The RPC is also imprecise as QC is, but consumes less power than DC or QC, has low cost in hardware, uses massive paralelism for speed up, and is robust to noise and system damage.

Article Entropy considerations in improved circuits for a biological...

https://www.semanticscholar.org/paper/The-Promise-and-Challenge-of-Stochastic-Computing-Alaghi-Qian/8db2fec2d07933fc083aabdd21feec60c707fa23