Day 60 · 2026.08.21

The Frontier of Mathematics

The four sharpest questions in mathematics today are not "is this true?" but "what entitles us to say so?"
"The frontier is not on the far side of harder theorems. It is where our answer to 'what counts as a proof' is being rewritten."

The abc Conjecture

Addition and multiplication cannot conspire
Number Theory
Intuition

Two structures sit on top of the integers: addition and multiplication. Each is simple on its own, yet they barely speak to each other — knowing the prime factorizations of $a$ and $b$ tells you almost nothing about the factorization of $a+b$. Nearly every hard problem in number theory grows out of that crack.

The abc conjecture is the strongest single sentence anyone has written about it: addition and multiplication cannot conspire. A number is "tidy" multiplicatively when it is built from a few primes raised to high powers — a lot of size from very little raw material, which is extravagant. The conjecture says: once $a+b=c$ ties three numbers together, they are not allowed to be that extravagant all at once.

Formal definition

The radical $\operatorname{rad}(n)$ is the product of the distinct primes dividing $n$ — flatten every exponent to 1. So $\operatorname{rad}(72)=\operatorname{rad}(2^3\cdot 3^2)=6$. It records which primes took part, never how deeply.

$\gcd(a,b)=1,\; a+b=c \;\Longrightarrow\; c < K_\varepsilon\cdot\operatorname{rad}(abc)^{\,1+\varepsilon}$

Here $\varepsilon$ is any small positive number and $K_\varepsilon$ a constant depending only on $\varepsilon$. A small $\operatorname{rad}(abc)$ is exactly the measure of "all three are multiplicatively tidy." So the inequality says: $c$ cannot greatly exceed the raw material it was built from. Exceptions do exist — $1+8=9$ while $\operatorname{rad}(1\cdot8\cdot9)=6<9$ — but only finitely many for each $\varepsilon$.

8 = 2³ 2 2 2 9 = 3² 3 3 rad: flatten 2 3 = 6 1 + 8 = 9, yet rad(1·8·9) = 6 < 9 too tidy + an additive relation = anomalous, and rare
Why it is beautiful

It is a load-bearing beam: once it stands, everything hanging from it follows for free — Fermat's Last Theorem (for large enough exponents), Roth's approximation theorem, the Mordell conjecture, the Szpiro conjecture. One inequality subsumes half of Diophantine analysis. Most conjectures are leaves; abc is a trunk — and its content is plain enough to explain to a schoolchild.

Applications

An effective form of abc would give computable bounds on the solutions of Diophantine equations — "check this far and you are done," precisely what computational number theory has always lacked. The Szpiro conjecture it implies constrains the ratio of discriminant to conductor of an elliptic curve, the very quantities weighed when choosing elliptic-curve cryptography parameters.

More famous is the deadlock it produced. In 2012 Shinichi Mochizuki released five hundred-odd pages of Inter-universal Teichmüller theory claiming a proof; in 2018 Scholze and Stix, after meeting him in person, identified an unbridgeable gap in Corollary 3.12, which Mochizuki rejects; in 2021 the papers appeared in a journal he edits. The mainstream still has not accepted it. This is not "we don't know whether it's right" — it is a community unable to reach consensus, a rare failure mode in mathematics.

The essential line + a question
Addition and multiplication cannot both be simple at once — and if that one line holds, half of number theory collapses into its corollaries.
Question: why should an inequality imply so many statements of the form "this equation has no solutions"?

Formalized Proof

Machine-checked mathematics · proof written for a machine
Foundations
Intuition

A proof in a paper is a persuasive text written for humans: the author is betting the reader will nod at every "clearly." A formalized proof is written for a machine that knows only axioms and inference rules — every step must be mechanically checkable, and the word "clearly" does not exist. The first is an essay; the second is a program that compiles.

The key design is the trusted kernel. However elaborate the automation, however clever the AI search, all of it must ultimately emit a proof term handed to a small kernel — a few thousand lines, read and re-read by many people — which checks it line by line. So you need not trust the AI. You need only trust that little box.

Formal definition

The foundation is the Curry–Howard correspondence:

$p : P \qquad\Longleftrightarrow\qquad p \text{ is a proof of the proposition } P$

Read it this way: the proposition $P$ is simultaneously a type, and the proof $p$ is simultaneously a program term of that type; "$P$ is provable" becomes "the type $P$ is inhabited." The payoff is decisive: "is this proof correct?" degenerates into "does this type-check?" — something computers are extremely good at, and can get deterministically right.

human tactics automation / search AI generation proof term (untrusted — merely a candidate) trusted kernel ≈ few kLOC → ✓ / ✗ all trust rests on that one small box at the bottom
Why it is beautiful

It carves out of social consensus a mechanically verifiable core of mathematical truth — and, in the same motion, measures how much human proofs leave out. The ratio of formalized text to original is called the de Bruijn factor; once in the dozens, it is now down to single digits in mature areas. That falling number is the machine gradually learning what humans leave unsaid — a philosophical question you can put a number on, which is very rare.

Applications

The 1976 four-colour theorem rested on a machine enumerating 1,936 configurations and triggered the argument "does that even count as a proof?" After four years of refereeing, Hales's proof of the Kepler sphere-packing conjecture drew only "99% certain" from the referees; he then spent eleven years formalizing it (Flyspeck, finished 2014). The turning point was Scholze's 2021 Liquid Tensor Experiment: unsure of the single most delicate step in his own new proof, he handed it to the Lean community, and it checked out within six months — the first time formalization served as live quality control on frontier research.

Industry runs on the same machinery: the correctness of the CompCert compiler and the seL4 microkernel are machine-checked theorems. On the AI side, Lean supplies something reinforcement learning rarely gets — a hallucination-free reward signal: the proof either compiles or it does not.

The essential line + a question
Turn a proof into a program and "is it right?" becomes "does it compile?" — trust moves from people to a kernel small enough to read line by line.
Question: do you trust a piece of code because you read it, or because the tests pass? Mathematics is switching sides on the same question.

Experimental Mathematics

See it first, prove it after
Discovery
Intuition

You compute a number: $1.20205690315959\ldots$, and have no idea what it is. Feed it to an integer-relation engine and the machine replies that it is very likely $\zeta(3)$. You have proved nothing — but you now know what to try to prove.

This is how physicists work, and the axiomatic narrative of the twentieth century hid it from view. Euler computed $\zeta(2)$ to many digits, recognized $\pi^2/6$, and only then went looking for a proof. Experimental mathematics merely re-legitimizes the logic of discovery; it never pretends to be the logic of justification.

Formal definition

The core tool is integer relation detection: given reals $x_1,\dots,x_n$ (computed to hundreds of digits), find integers $a_1,\dots,a_n$, not all zero, with

$a_1x_1+a_2x_2+\cdots+a_nx_n = 0$

The PSLQ and LLL algorithms find such a relation in polynomial time, when one exists and the precision suffices. The use is direct: let the $x_k$ be powers of the constant under study, or a basket of known constants; the moment the machine emits integer coefficients, you have guessed a closed form. Precision is the microscope here: too few digits and real relations stay invisible; enough digits and spurious ones become all but impossible.

Why it is beautiful

The prettiest catch is the 1995 BBP formula, found by PSLQ and proved only afterwards:

$\pi=\sum_{k=0}^{\infty}\dfrac{1}{16^{k}}\left(\dfrac{4}{8k+1}-\dfrac{2}{8k+4}-\dfrac{1}{8k+5}-\dfrac{1}{8k+6}\right)$

The $16^{-k}$ is the whole point: it lets you compute the $n$-th hexadecimal digit of $\pi$ without computing any of the digits before it. In two thousand years nobody imagined this was possible — because nobody had a reason to. What found it was not insight but search: search too can strike structure lying outside human imagination.

Applications

The OEIS (On-Line Encyclopedia of Integer Sequences) does this daily: throw in a string of numbers, get back the combinatorial object behind it. In theoretical physics, closed forms for Feynman integrals are often guessed by numerical identification (multiple zeta values) and derived rigorously afterwards. In 2021 DeepMind, with mathematicians at Oxford and Sydney, used neural networks to surface a strong dependence in large datasets of knot invariants, from which a new theorem was conjectured and then humanly proved — the machine is not the prover but the conjecture generator: it points a direction through high-dimensional data, the human supplies the reason.

The essential line + a question
Numerical evidence is not proof, but it decides where you spend your proving effort — always the scarce resource in mathematics.
Question: how strong must the evidence be before you would bet? How much should you believe a statement verified up to $10^{10}$?

What Counts as a Proof

True, provable, accepted — three different things
Philosophy
Intuition

We treat "proof" as an objective object. In practice a proof is accepted because enough competent people read it and believed it — a social process. Three pressures are squeezing that process at once. Proofs are too long (the classification of finite simple groups spans hundreds of papers and nearly ten thousand pages, some authors now dead); too machine-made (nobody can read the enumerations behind four colours or Kepler); too private (a mere handful claim to understand IUT).

Formal definition

Separate three things that get conflated:

$\models\varphi\;$ (true in every model) $\quad$ $\vdash\varphi\;$ (a legal symbolic derivation exists) $\quad$ $\mathsf{Acc}(\varphi)\;$ (the community accepts it)

Gödel's completeness theorem gives $\vdash\varphi \leftrightarrow \models\varphi$ inside first-order logic. But $\mathsf{Acc}$ is never $\vdash$: no real paper is written as a formal derivation. All of mathematical practice lives in that gap.

⊨ φ semantically true ⊢ φ formally provable Gödel completeness: equivalent Acc(φ) community accepts the gap: omission, trust, authority formalization = the first tool for calibrating Acc to ⊢
Why it is beautiful

Admitting the gap does not weaken mathematics; it explains its efficiency. Omission is exactly what makes human mathematics readable — a fully formalized text is unimpeachably correct and nearly impossible to read. Mathematics has always chased two goals, certainty and understanding, and for centuries a single act (writing a proof) delivered both as a side effect, so nobody noticed they were two things. Machines have pulled them apart for the first time — that is where the frontier really lies.

Applications

The shape is already visible. AlphaGeometry and AlphaProof reach silver-medal level on IMO problems, and their reliability comes from Lean's mechanical checking rather than the model's confidence. Terence Tao's Equational Theories Project settled over twenty million implications in a matter of weeks using Lean plus crowdsourcing plus AI — collaboration at that scale simply was not a viable form of organization in the pen-and-paper era. Tao calls it "industrial-scale mathematics": humans ask, decompose and exercise taste; machines enumerate and verify.

The essential line + a question
Mathematics has always wanted two things — certainty and understanding. Machines are perfecting the first, and forcing us to admit that the second must now be pursued on its own.
Question: if a machine produced a billion-step proof that is certainly correct and unreadable by anyone, would you consider the problem solved?

Going Deeper

Does the abc deadlock show that mathematics is not objective?
No. The truth value is objective; our access to it runs through a social process. The IUT impasse is not "some say yes, some say no" — the two sides cannot even align on what the crucial step asserts. Failure of shared meaning comes before any verdict on correctness. This is precisely formalization's strongest use case: written into Lean, the disagreement would degenerate into a decidable technical question. Why nobody has done it is equally clear — formalization presupposes understanding. You must grasp it before you can write it down.
How unreliable can numerical evidence be?
Extremely. Pólya's conjecture looks flawless on small data; its least counterexample is $n=906{,}150{,}257$. Mertens's conjecture was disproved in 1985, with the counterexample bounded below somewhere past $10^{30}$. The crossover of $\pi(x)$ and $\mathrm{li}(x)$ — the Skewes number — sits far beyond anything computable. The lesson: in number theory "small" can run all the way to $10^{30}$. Numerical evidence is good at ruling out directions and hinting at structure, not at supporting universal claims.
If AI can prove any theorem, what do mathematicians do?
Whatever machines do not: ask questions worth asking, judge which definition is the right one, decide what is interesting. The real leaps in mathematical history came from definitions more than theorems — group, sheaf, scheme, category all began with someone deciding "this is how it should be seen." Such judgements rest on taste and cross-domain intuition, and there is no objective function to optimize for them. The risk is real too: if the next generation grows up with "the machine will fill in the details," the intuition that only comes from grinding through computations may atrophy across a whole cohort.
Will formalization change what mathematics looks like?
Almost certainly. Branches that formalize readily — algebra, combinatorics, logic — will accelerate first, while those leaning on geometric intuition and "the reader will see it" lag behind: a selection bias induced by tooling. The reverse happens too: to write something into Lean, people are forced to polish definitions, and often discover ambiguity hiding in the original statement. Tools do not merely record mathematics; they select it.