In computer science, one of the classic problems is called P vs. NP. The basic idea is that P is the class of problems where the answer can be computed in time that is less than some polynomial of the length of the input, and NP is the class of problems where an answer can be guessed and verified in polynomial time. (It is not known if these classes are identical or not.)
For example, suppose we wanted to know if the minimum of a list is less than a constant C. We could simply scan the list, and report if we find an item less than C or not. This requires looking at each item once, so the time is linear in the length of the list, so this problem is in P.
Similarly, suppose we are given a list of cities and distances between them, and want to know if there is a path that visits each city exactly once that has a total length less than C. If we make a lucky guess of a path, it is a simple matter to add up the total length of the path and verify that it is less than C. The number of additions is linear in the number of cities, so this "travelling salesman" problem is in NP.
If you want to solve one of the hardest problems in NP with a computer program, there are few options besides guessing each possible answer in turn and then attempting to verify it. Since the number of possible answers is exponential in the length of the input, this procedure will be exponential in the input, not polynomial.
However, the hope is that by using superposition, we can use superposition to do a calculation on *all* of the possible guesses at the same time, and as a result, we would have a method of solving NP-hard problems in polynomial time.