Lehmer conjectured that Ramanujan's tau-function never vanishes. In a related direction, a folklore conjecture asserts that infinitely many primes arise as absolute values of Ramanujan's tau-function. Recently, Xiong showed that these prime values form a subset of the primes with density at most $2/11$. Assuming the $abc$ Conjecture, we prove the stronger upper bound \[ S(X):=\#\{\ell\le X:\ \ell\ \text{prime and } |τ(n)|=\ell \text{ for some } n\ge 1\} = O(X^{9/10}\log X), \] which implies that Ramanujan's tau-function misses a density 1 subset of the primes. We give a heuristic suggesting that $S(X)$ should nevertheless be infinite, with predicted order of magnitude \[ S(X)\asymp \frac{C X^{\frac{1}{11}}}{(\log X)^2}. \] The main engine in this note was formalized and produced automatically in Lean/Mathlib by AxiomProver from a natural-language statement of the problem.
Alex: Welcome to another episode of ResearchPod. Today we're diving into a paper that asks a sharp question about numbers from a famous math function. Sam, can you set the stage for us?
Sam: This work is titled "ABC Implies that Ramanujan’s Tau Function Misses Almost All Primes." It tackles a longstanding puzzle: does Ramanujan's tau function produce infinitely many prime numbers in its values? The authors show, assuming a major unproven idea called the abc conjecture, that it actually skips almost all primes—meaning the primes it hits become a vanishingly small share as numbers get larger.
Alex: So this paper is basically saying Ramanujan's tau misses a density one subset of primes under abc?
Sam: Yes, exactly. Previously, work by Xiong capped the density at most 2/11 in certain cases, but here they push it to zero density overall.
Alex: Right, so tau(n) is this sequence of integers from some special generating function. Walk me through what that function is, like why it even matters.
Sam: Picture the tau function as the building blocks of a power series tied to a symmetric pattern in the complex plane—like coefficients in an infinite expansion that encodes deep properties of numbers. Ramanujan spotted it, and it's central because its values hint at prime patterns. The old guess was that infinitely many of these values are prime, up to sign.
Alex: Okay, got it—special coefficients with prime curiosity. But how do they count those prime hits? It can't just be checking every n.
Sam: They reduce the search using Hecke relations, which are recursive rules for tau on prime powers. Odd primes in tau(n) must come from even powers n = p^{2k} since tau is odd only on odd squares. For k=1, tau(p^2) = tau(p)^2 - p^11 = ±ℓ implies y^2 = x^11 ± ℓ with x=p, y=tau(p).
Alex: Wait—so that's points where two integers satisfy y squared nearly equals x to the 11th, off by a small prime?
Sam: Precisely. It's like a high-degree twist where y squared nearly equals x to the 11th. For k=2, it leads to a similar but higher power equation. Xiong handled higher k≥3 with bounds, leaving the counts of large-x points near these curves to bound tightly.
Alex: So the core is pruning these near-misses on the curves. And abc comes in to limit how many such points exist?
Sam: Yes. The abc conjecture says for coprime a+b=c nonzero integers, the size of c is controlled by the distinct primes dividing a, b, and c. For these curves, they set a = x^11, b= ±ℓ or small k, c=y^2. This forces x to stay small, so few points.
Alex: That prunes the search space sharply—like abc acts as a fence keeping large x out. So overall, the count of such primes is far sparser than the total number of primes.
Sam: Correct. Xiong had density ≤2/11 in some cases; here abc kills the rest, implying tau misses density 1 primes. They note explicit primes exist, like tau(251^2) a huge negative prime, but sparse.
Alex: But it's conditional on abc, right? Unproven, though famous.
Sam: Yes, fully conditional—the main engine, those curve bounds, assumes abc. Higher k≥3 uses Xiong unconditionally.
Alex: And for infinity? They don't prove it, but hint?
Sam: In section 4, a heuristic using Sato-Tate: tau(p) behaves like 2 p^{11/2} cos θ_p, equidistributed; for fixed m, primes p up to some size give values ~ p^{11m}, suggesting the total count grows like X^{1/11} times log terms—infinite but ultra-sparse.
Alex: So even if density zero, still potentially endless, just very slow growth. Neat counterpoint. And they formalized the engine in Lean/Mathlib via AxiomProver—from natural language to verified code, assuming abc and Xiong's results.
Sam: Indeed. Input informal statements, config for Lean 4.26, Xiong papers; output autonomous proofs.
Alex: Wrapping up, this shows tau likely misses almost all primes under abc, yet heuristics say infinitely many lurk. A precise conditional advance on a classic puzzle. Thanks, Sam—clear as always.
Sam: Pleasure unpacking it. The logic holds tightly on those curves. Thanks for listening to ResearchPod.