DeepSeek vs Kimi

The most significant unsolved problems in mathematics are deep, fundamental questions that have resisted solution for decades—sometimes centuries—despite immense effort by generations of brilliant mathematicians.

Here are some of the most famous, along with reasons why they remain unsolved.

---

1. Riemann Hypothesis

• Problem: All non-trivial zeros of the Riemann zeta function have real part $\frac{1}{2}$.
• Why it’s hard: The zeta function connects prime numbers (discrete, arithmetic) to analytic properties (continuity, complex analysis). Proving that all zeros lie on the critical line requires new insights into the interplay between number theory and complex analysis. Many partial results and numerical verification (trillions of zeros) support it, but no global proof exists.
• Why resisted: There’s no known approach to constrain all zeros at once; current techniques (like zero-density estimates) fall short of the full statement. It would likely require a new theory linking primes and analysis more deeply.

---

2. P vs NP Problem

• Problem: Is every problem whose solution can be verified quickly also solvable quickly?
• Why it’s hard: It’s about the fundamental nature of computation and problem difficulty. Almost all known proof techniques in computational complexity relativize (are valid relative to any oracle), but there exist oracles for which P=NP and others for which P≠NP, so such techniques can’t resolve it.
• Why resisted: We lack techniques to prove super-polynomial lower bounds for general algorithms on Turing machines for natural NP-complete problems like Boolean satisfiability. The problem seems to require non-relativizing, non-naturalizing proof methods.

---

3. Navier–Stokes Existence and Smoothness

• Problem: In 3D, do smooth solutions to the incompressible Navier–Stokes equations always exist for all time, or can they develop singularities?
• Why it’s hard: Nonlinear partial differential equations in 3D are extremely complex; possible cascade of energy to infinitesimally small scales (turbulence). We don’t have mathematical tools to rule out finite-time blowup for all initial conditions.
• Why resisted: The equations are fundamental to fluid mechanics but mathematically “unforgiving”—energy conservation and viscosity balance in a way that is hard to fully analyze. Progress often requires harnessing very subtle cancellations in the nonlinear term.

---

4. Birch and Swinnerton-Dyer Conjecture

• Problem: The rank of an elliptic curve’s Mordell–Weil group equals the order of vanishing of its L-function at $s=1$.
• Why it’s hard: It links arithmetic data (rational solutions) to analytic data (L-function). For curves of rank > 1, we don’t fully understand how to systematically produce rational points or control the Selmer group.
• Why resisted: Deep issues in Iwasawa theory and the p-adic L-function; still unknown in general even for analytic rank 2 or more whether the Tate–Shafarevich group is finite (a key part of the conjecture).

---

5. Hodge Conjecture

• Problem: On a smooth projective complex variety, which cohomology classes are represented by algebraic cycles?
• Why it’s hard: Algebraic cycles are geometric objects (subvarieties), but cohomology classes are topological/analytic. The bridge—Hodge theory—gives necessary conditions; conjecture says those conditions are sufficient. Difficulty increases with codimension > 1.
• Why resisted: We lack general construction methods for algebraic cycles with given Hodge type. Known techniques (e.g., Abel–Jacobi maps) fail in higher codimension.

---

6. Yang–Mills Existence and Mass Gap

• Problem: Prove that quantum Yang–Mills theory in 4D has a mass gap (lowest excitation above vacuum has positive energy).
• Why it’s hard: Quantum field theory in 4D is mathematically non-rigorous in physics; constructing an interacting gauge theory satisfying Wightman axioms and showing mass gap is far beyond current analysis/PDE and probability tools.
• Why resisted: Requires nonperturbative definition of the path integral, renormalization at strong coupling, and control over large-scale behavior in a nonlinear setting. Only simpler lower-dimensional or non-interacting cases are well-understood mathematically.

---

General Reasons Problems Stay Unsolved

1. Lack of bridge between fields: Many conjectures connect two distant areas (analysis & number theory, geometry & algebra). Only when one field advances enough to provide tools for the other can progress occur.
2. Insufficiently developed theories: Some problems need entirely new frameworks (e.g., Wiles’ proof of Fermat used modularity lifting, a theory not available earlier).
3. Extreme complexity or infinite behavior: Problems about all primes, all smooth functions, all algorithms involve “universal” statements that are hard to pin down.
4. Counterexamples hard to find or rule out: For P vs NP, we can’t find a fast algorithm for SAT, nor can we prove none exists.
5. Computational verification only partial: Massive numerical evidence (like for Riemann Hypothesis) doesn’t replace a proof for infinitely many cases.

---

These problems continue to inspire new mathematics because each attempt at a solution often leads to powerful new theories—even if the original problem remains open.

The landscape of unsolved mathematics is vast, but certain problems have achieved particular notoriety due to their age, their implications for other fields, and the peculiar stubbornness with which they resist attack. Below are the most significant unsolved problems, categorized by their domains, along with the specific technical and conceptual barriers that have prevented their resolution.

---

1. The Millennium Prize Problems (Clay Mathematics Institute)

These seven problems (six remaining after Perelman solved the Poincaré Conjecture in 2003) represent the "Mount Everests" of modern mathematics, each carrying a $1 million prize.

The Riemann Hypothesis (1859)

The Problem: All non-trivial zeros of the Riemann zeta function $\zeta(s)$ have real part equal to $1/2$. This implies the "most regular possible" distribution of prime numbers.

Why it resists proof:

• The Analytic Gap: Current tools in analytic number theory (sieve methods, harmonic analysis) can control the average behavior of primes, but the Riemann Hypothesis requires controlling microscopic fluctuations. We lack techniques to rule out "conspiratorial" arrangements of zeros off the critical line.
• Physical Analogy without Proof: The zeros behave statistically like eigenvalues of random Hermitian matrices (quantum chaos). This suggests the problem is connected to physics (Hilbert-Pólya conjecture), but we have no rigorous bridge between quantum mechanics and number theory.
• Critical Line Barrier: We can prove that $\approx 40%$ of zeros lie on the critical line, and that "most" are very close to it, but crossing from "almost all" to "all" requires eliminating pathological configurations we cannot yet conceptualize.

P versus NP

The Problem: Are problems whose solutions can be verified quickly (NP) necessarily solvable quickly (P)? Formally: Is $\text{P} = \text{NP}$?

Why it resists proof:

• The Barrier Theorems: Three major results prove that entire classes of proof techniques cannot resolve the question:
  1. Relativization (Baker-Gill-Solovay): Techniques that work regardless of oracles cannot separate P and NP.
  2. Natural Proofs (Razborov-Rudich): Computational lower bound proofs that satisfy certain "natural" properties cannot exist unless cryptography fails.
  3. Algebrization (Aaronson-Wigderson): Extensions of relativization that include algebraic operations also fail.
• This means resolving P vs. NP requires a proof technique unlike any we currently possess—one that is inherently non-relativizing and non-naturalizing.

Navier-Stokes Existence and Smoothness

The Problem: Do solutions to the Navier-Stokes equations (governing fluid flow) exist for all time and remain smooth, or do singularities (blow-ups) form in finite time?

Why it resists proof:

• Turbulence: The equations describe turbulent flow, which involves energy cascading across infinitely many scales. To prove regularity, one must control the vorticity at all scales simultaneously.
• Supercriticality: The problem is "supercritical"—we can control solutions at scales where the nonlinearity is weak, but the energy concentration that causes blow-up (if it exists) happens exactly where our estimates break down.
• Scaling Symmetry: The equations have a scaling invariance that makes it difficult to distinguish between a genuine singularity and a "near-singularity" that resolves itself.

Yang-Mills and the Mass Gap

The Problem: Prove that quantum Yang-Mills theory (the foundation of the Standard Model) exists mathematically and exhibits a "mass gap" (the mass of the lightest particle is strictly positive).

Why it resists proof:

• Infinite Dimensions: The problem requires rigorous measure theory on infinite-dimensional spaces of connections (path integrals).
• Renormalization: Physicists use perturbation theory and renormalization to calculate observables, but these are not mathematically rigorous constructions. Proving the mass gap requires non-perturbative control over the quantum field theory.
• Confinement: The mass gap is related to the physical phenomenon of quark confinement, which is analytically intractable in 4D.

Birch and Swinnerton-Dyer Conjecture

The Problem: Relates the algebraic rank of rational points on an elliptic curve to the order of the zero of its associated L-function at $s=1$.

Why it resists proof:

• Bridge Between Worlds: It connects discrete arithmetic (the group of rational points) with complex analysis (L-functions). The two languages do not translate easily.
• The Tate-Shafarevich Group: The conjecture's fine form involves the Tate-Shafarevich group Ш (Sha), which measures the failure of the Hasse principle. This group is conjectured to be finite, but we cannot even prove it is finitely generated in general.

Hodge Conjecture

The Problem: On projective algebraic varieties, certain topological "Hodge cycles" (rational cohomology classes of type $(p,p)$) can be expressed as linear combinations of classes of algebraic subvarieties.

Why it resists proof:

• The Algebraic/Topological Divide: The conjecture asks when a topological object (a cohomology class) must come from an algebraic object (a subvariety). We lack criteria to detect algebraicity at the level of cohomology.
• Singularities: The obvious approaches fail because algebraic cycles can have singularities, and the Hodge condition is smooth.

---

2. Classical Number Theory

Goldbach's Conjecture (1742)

The Problem: Every even integer greater than 2 is the sum of two primes.

Why it resists proof:

• The Parity Problem: Sieve methods (the workhorse of additive number theory) can count numbers with specific prime factors, but they encounter a fundamental barrier: they cannot distinguish between numbers with an even or odd number of prime factors. Since semi-primes (products of two primes) have exactly two factors, sieves struggle to isolate them.
• **We can prove "almost Goldbach" (every sufficiently large even number is the sum of a prime and a semi-prime), but crossing to two primes requires breaking this parity barrier.

Twin Prime Conjecture

The Problem: Are there infinitely many primes $p$ such that $p+2$ is also prime?

Why it resists proof:

• The Gaps Barrier: While Zhang (2013) and Maynard proved there are infinitely many prime gaps of at most 246 (later improved), extending this to 2 requires controlling the "level of distribution" of primes in arithmetic progressions beyond the Bombieri-Vinogradov theorem.
• Pseudorandomness: The primes behave enough like random numbers that probabilistic models suggest gaps of 2 occur infinitely often, but also enough like deterministic objects that proving their specific recurrence is maddeningly difficult.

The Collatz Conjecture (3n+1 Problem)

The Problem: Given any positive integer, if even, divide by 2; if odd, multiply by 3 and add 1. Repeat. Does this process always reach 1?

Why it resists proof:

• Lack of Structure: The map mixes additive and multiplicative operations in a way that destroys algebraic structure. It behaves like a pseudorandom dynamical system.
• Tao's "Almost" Result: Terence Tao proved that "almost all" Collatz orbits attain almost-bounded values, but this is qualitatively different from proving boundedness for all starting values.
• No Modular Obstruction: Unlike Fermat's Last Theorem, which failed modulo certain numbers, the Collatz conjecture is consistent with all known modular constraints.

---

3. Foundational and Logical Limits

The Consistency of ZFC

The Problem: Can we prove that the Zermelo-Fraenkel axioms with Choice (the foundation of modern mathematics) do not contain a contradiction?

Why it resists proof:

• Gödel's Second Incompleteness Theorem: If ZFC is consistent, it cannot prove its own consistency. Any consistency proof must therefore rely on axioms stronger than ZFC (e.g., large cardinal axioms), which shifts the problem rather than solving it.
• Independence: The Continuum Hypothesis (CH) and the Axiom of Choice are independent of ZFC. This suggests many "natural" mathematical questions may be unsolvable within our current axiomatic framework.

---

Common Themes Why Proofs Fail

1. The Wrong Toolkit: Many problems (Goldbach, Twin Primes) require methods that transcend current sieve theory; others (P vs. NP) require proof techniques barred by meta-mathematical theorems.

2. Scale Mismatch: Problems like Navier-Stokes and Riemann involve controlling behavior across infinitely many scales (turbulence/zeros) simultaneously, while our tools work locally or on average.

3. Interdisciplinary Blindness: The hardest problems (BSD, Hodge, Yang-Mills) sit at the intersection of multiple fields. They resist solution because mathematicians in one specialty lack the technical vocabulary of the other.

4. Computational Intractability: Some problems (finding exact solutions to chaotic systems, or optimal algorithms) may be solvable in principle but require computation beyond the capacity of the observable universe.

5. Axiomatic Independence: Some questions may be literally undecidable from current axioms, requiring new foundational axioms (e.g., large cardinals) to resolve—changing the question from "Is it true?" to "Should we assume it?"

A Note on Progress: While these problems remain unsolved, the partial results are often revolutionary. The methods developed to fail at proving the Twin Prime Conjecture (sieve theory) revolutionized Diophantine approximation. The attempt to understand the Riemann Hypothesis birthed algebraic geometry and modern number theory. In mathematics, the resistance of a problem often creates more value than its solution would.