This is my current read on it, and we also ran the claim through our review framework.
There is a real mathematical object behind this discussion, but the image overstates what its plotted construction actually shows.
1) The correct analytic object
Define the Gaussian-smoothed arithmetic field
where ∧(n) is the von Mangoldt function.
Now define the completed spectral object
This is a legitimate analytic object. In that completed sense, the Gaussian-smoothed prime/zero correspondence is real.
Using
together with the Gaussian Mellin inversion identity,
one obtains formally
That identity is real. But it is not a new discovery. It belongs to the classical explicit-formula tradition: its origin is in Riemann’s 1859 work, with rigorous formulation and proof developed by von Mangoldt in 1895 and later explicit-formula theory.
2) Where the image goes wrong
First: the plotted Z(T) is not the same object as Q(T)
The image’s Z(T) is a truncated cosine sum over zeros. That is not automatically the same thing as the completed analytic object QH(T).
So while
belongs to the classical explicit-formula setting, it does not follow that the plotted finite surrogate satisfies the same equality unless that extra step is separately derived and justified.
Second: “high correlation guarantees zeros generate prime locations” does not follow
That is too strong.
A high correlation score is a numerical observation. It is not, by itself, a proof of prime recovery, and it is not evidence that the zeros “generate” prime locations in the strong causal sense implied by the image.
The explicit-formula framework gives a deep structural relation between primes and zeros. But that is an analytic identity, not a conclusion that follows from a large Pearson or cosine score.
Third: leaving Q(T)=??????? is rhetorical, not mathematical
In this context, Q(T) is not some newly discovered unknown object. It can already be defined analytically, as above.
So presenting Q(T) as though it were the missing final piece of a new equivalence is misleading. The deeper framework is classical; the novelty claim is not.
3) The honest interpretation
The mathematically honest version is narrower:
- the prime-zero correspondence is real,
- the completed Gaussian-smoothed identity is real,
- but the image’s plotted Z(T) is only a truncated zero-only surrogate,
- and the stronger claim is therefore overstated.
So the right conclusion is not
in the sense the image suggests.
The right conclusion is:
while the plotted finite Z(T) still requires additional justification and cannot simply be promoted to that exact identity.
4) What the patched code actually tests
The patched code is still useful, but only for a narrower purpose.
It does not test the full completed explicit formula.
It tests a more modest question:
- build PH(T) independently from von Mangoldt data,
- build a truncated zero-only surrogate independently from actual zeta zeros,
- then ask whether near-perfect agreement appears automatically.
That is a legitimate non-circular sanity check.
What it shows is not that the classical explicit formula is false.
What it shows is that the stronger numerical claim was overstated, and that the earlier circular construction should not be treated as proof.
5) Non-circular sanity-check code
import numpy as np
from mpmath import mp, zetazero
mp.dps = 25
def von_mangoldt(n: int) -> float:
"""
Returns log(p) if n = p^k for some prime p and integer k >= 1,
otherwise returns 0.0.
"""
m = n
p = 2
while p * p <= m:
if m % p == 0:
while m % p == 0:
m //= p
return np.log(p) if m == 1 else 0.0
p += 1
return np.log(n) # n itself is prime
def zscore(x: np.ndarray) -> np.ndarray:
x = np.asarray(x, dtype=float)
s = x.std()
if s == 0:
return np.zeros_like(x)
return (x - x.mean()) / s
def pearson(a: np.ndarray, b: np.ndarray) -> float:
a_n = zscore(a)
b_n = zscore(b)
return float(np.corrcoef(a_n, b_n)[0, 1])
# Parameters
T = np.linspace(1.0, 10.0, 300)
H = 0.3
N_max = 1500
M_zeros = 100
# 1) Arithmetic source: P_H(T)
P = np.zeros(len(T))
norm = 1.0 / (np.sqrt(2.0 * np.pi) * H)
for n in range(2, N_max + 1):
lam = von_mangoldt(n)
if lam > 0.0:
P += lam * norm * np.exp(-((np.log(n) - T) ** 2) / (2.0 * H ** 2))
# 2) Independent zero-only truncated spectral surrogate
# This is NOT the full completed explicit formula Q_H(T).
gammas = [float(zetazero(k).imag) for k in range(1, M_zeros + 1)]
Z_tilde = np.zeros(len(T))
for g in gammas:
Z_tilde += np.exp(-0.5 * H * H * g * g) * np.cos(g * T)
# Honest metrics
raw_corr = pearson(P, Z_tilde)
diff_corr = pearson(np.diff(P), np.diff(Z_tilde))
print(f"Pearson r (z-scored raw signals): {raw_corr:.6f}")
print(f"Pearson r (first differences): {diff_corr:.6f}")
print("\nInterpretation:")
print("- This is a non-circular comparison.")
print("- P is built from von Mangoldt data.")
print("- Z_tilde is built independently from actual zeta zeros.")
print("- This does NOT test the full completed explicit formula.")
print("- It DOES test whether a truncated zero-only construction")
print(" automatically reproduces the arithmetic signal strongly")
print(" enough to justify the original claim.")
6) How to read that code correctly
This code is valid as a non-circular sanity check.
It does not disprove the classical explicit formula.
It does not prove RH.
It does not show that zeros directly generate prime locations.
What it does show is narrower and more honest:
an independently constructed truncated zero-only cosine sum does not, by itself, justify the stronger claim made in the image.
7) Final one-sentence conclusion
My view is that the image repackages a classical explicit-formula correspondence as though it were a newly solved mystery, and then upgrades numerical resemblance into a causal claim. In the sense the image implies, P(T)≡Z(T)≡Q(T) does not hold.