The author did not really want you to solve the problem, but to see that this kind of problem is undesireable and not satisfying. To see this you should consider the full bock of exercises of that chapter "Notes on the Exrcises".
For one, solving the base problem consists of merely computing a simple expression, for example by multiplying $13\cdot 13\cdot 13$. Next, there is nothing remarkable about this specific result. Then the "Genaralize your answer" strikes us in the face: What should one generalize here? That $13^4$, $13^5$ and so on can als be computed? So while the difficulty score 14 suggests that you should be able to solve the problem in a few minutes, Knuth's parenthesized comment should tell you not to take this specific problem too seriously.
In fact, the fourth exercise (to prove FLT) is rated M50, i.e. as mathematically oriented (open at the time of his writing) research project, thus rather being an example of how the rating system works than a seriously intended exercise.
So, once you have mastered the (recommended by the author!) exercise 1, and maybe also exercise 2, the remaining two exercises of this chapter (after all it is titled "Notes on the Exercises" - so we could expect something meta going on here) are not to be considered as exercises, but rather as examples of how exercises might look (or in these cases maybe: should not look).
A final obseration in support of this: These exercises belong to the front matter, not to the text per se! (Which doesn't keep Knuth from adding answer hints for exercises 1 and 4 in the answers section)
I base the above on what I have in the shelf, which is the second edition. But I'm pretty sure it applies to all editions.