Neural networks are significantly concurrent in nature; many tasks can be processed at once. The 'normal computers' processes tasks sequentially; one task after the other. Though it might seem like the normal computer is multi-tasking, in actuality two processes are not actually happening simultaneously. The artificial neural network on the other hand is a essentially built on a multiprocessor architecture. The artificial neural network was designed to behave as close as possible to the human brain and it is easy to see why it was designed to be capable of parallel processing from the onset.
Neural networks and 'normal computers' function differently. Neutral networks function with images, and concepts. 'Normal computers' on the other hand can function logically with a set of rules and calculations. While the 'normal computers' can learn by predefined rules, artificial neural networks can learn only by examples, by doing something and then making its own rules based on the patterns observed.
'Normal computers' have to be programmed before they can get an output from an input. Neural networks on the other hand can program themselves; they can learn and make patterns of a given input data set which can be used to improve the processing algorithm. This is something a 'normal computer' simply cannot achieve.
Leave a comment below if you found this explanation helpful.