Hi HN,
I'm an independent researcher and I built an interactive tool that lets you explore structures in three classical number theory problems Collatz sequences, Riemann zeta zeros, and twin prime distributions all running client-side in a single 192 KB HTML file.
What it actually does:
Collatz: Constructs numbers with a target step count via inverse tree navigation (instead of testing billions of candidates). You pick a target like T=50, it builds the number in seconds and verifies it forward.
Riemann zeros: Navigates to zero positions at arbitrary heights (tested up to 10⁶⁸) using Gram points + a quasiperiodic correction model. GUE spacing statistics verified at 11 heights with a single calibration constant. Uses BigInt for arbitrary-precision fractional parts.
Twin primes: Visualizes the CRT structure mod 240 — why exactly 24 of 240 residue classes can hold twin primes, and how the Hardy-Littlewood error term decays as a power law (measured across 14 decades of OEIS data, 808 trillion pairs).
The underlying framework connects all three through the E₈ lattice modulus K₈ = 240. I'm not claiming proofs of these conjectures, this is a computational/visualization tool that demonstrates structural patterns. The measured exponents and statistics are empirical with honest confidence intervals.
Would love technical feedback.
Live demo: https://aidoctrine.github.io/uct-navigator/
What it actually does:
Collatz: Constructs numbers with a target step count via inverse tree navigation (instead of testing billions of candidates). You pick a target like T=50, it builds the number in seconds and verifies it forward.
Riemann zeros: Navigates to zero positions at arbitrary heights (tested up to 10⁶⁸) using Gram points + a quasiperiodic correction model. GUE spacing statistics verified at 11 heights with a single calibration constant. Uses BigInt for arbitrary-precision fractional parts.
Twin primes: Visualizes the CRT structure mod 240 — why exactly 24 of 240 residue classes can hold twin primes, and how the Hardy-Littlewood error term decays as a power law (measured across 14 decades of OEIS data, 808 trillion pairs).
The underlying framework connects all three through the E₈ lattice modulus K₈ = 240. I'm not claiming proofs of these conjectures, this is a computational/visualization tool that demonstrates structural patterns. The measured exponents and statistics are empirical with honest confidence intervals. Would love technical feedback. Live demo: https://aidoctrine.github.io/uct-navigator/