For n ≥ 3, x^n + y^n = z^n has no nonzero integer solutions. Conjectured in 1637 and proved in 1995 via elliptic curves and modular forms, it exemplifies deep mathematics needed to resolve elementary-sounding problems.
Study special cases n=3 and n=4 via infinite descent. Understand the connection to elliptic curves and modularity conceptually.
The proof is elementary (it requires deep algebraic number theory and geometry). It contradicts the Pythagorean theorem (it applies only to n ≥ 3).
Fermat's Last Theorem asks whether the Pythagorean equation — which you know has infinitely many whole-number solutions like 3² + 4² = 5² — can ever work for cubes, fourth powers, or any higher exponent. The answer is no: for any exponent n ≥ 3, the equation x^n + y^n = z^n has no solution where x, y, and z are all positive integers. Pierre de Fermat claimed to have a proof of this in 1637, scrawling in a book margin that the proof was too long to fit there. For 358 years, every attempt to reconstruct it failed, and the problem became one of the most famous unsolved questions in mathematics.
The difficulty of the problem lies in a profound asymmetry: it is easy to state in elementary terms but requires tools that didn't exist when Fermat wrote it. The Fundamental Theorem of Arithmetic, which you've studied, gives us the unique prime factorization of integers. Number theorists in the 19th century tried to generalize this to larger rings of "integers" (like the Gaussian integers a + bi), hoping to factor x^n + y^n in ways that would force a contradiction. This strategy partially worked — it proved the theorem for many specific values of n — but it failed in general because unique factorization breaks down in these extended rings.
The breakthrough came when mathematicians noticed a surprising bridge between two apparently unrelated areas. Elliptic curves are smooth cubic curves defined by equations like y² = x³ + ax + b. If Fermat's equation had a solution (a, b, c, n), one could construct a specific elliptic curve from those values — the Frey curve y² = x(x − aⁿ)(x + bⁿ) — that would have bizarrely pathological properties. Specifically, it would fail to be modular, meaning it would not correspond to a modular form (a highly symmetric complex-analytic function). The Modularity Theorem (formerly the Taniyama-Shimura conjecture) asserts that every elliptic curve over the rationals is modular. So if the Frey curve is not modular, the original Fermat solution cannot exist.
Andrew Wiles proved the Modularity Theorem for the class of semistable elliptic curves in 1995, completing the argument. The proof runs to over 100 pages of graduate-level algebraic number theory, Galois representations, and analytic techniques. It is a landmark not only for resolving a 358-year-old conjecture but for unifying distant areas of mathematics — number theory, algebraic geometry, and complex analysis — in a single sweeping argument. The lesson is that apparently elementary problems can encode deep structural facts about mathematics, and that sometimes the right question is not "how do we solve this equation?" but "what kind of mathematical object would a solution produce, and does such an object exist?"