Open Problems Lab Read the README
Python · NumPy

Small, checkable computations on open maths problems

Reproducible computations on open problems in number theory and combinatorics, mostly from erdosproblems.com, produced by an AI-agent research lab. No proofs, no prize claims.

Python · NumPy · 12 notes
open-problems-lab · previewscreenshot
Open Problems Lab screenshot
2platforms: Python, NumPy
12notes, each with its scripts and a referee re-run
MITopen source licence
Why it exists

Open problems, in steps small enough to check.

Nothing here is a proof, a solution or a prize claim.

The work was computed and re-run by AI agents.

In short

Three rules the lab works by

01

Reproduce first

Every result must first reproduce published values exactly.

02

Compute twice

Each new number is recomputed by code that shares nothing with the first script.

03

Say what failed

Drafts that turned out wrong or already known are listed openly.

Run it yourself

Reproduce a result

From the README’s “Reproduce a result” section. Full options, requirements and troubleshooting are there too.

Open the README
bash
$ git clone https://github.com/juw12/open-problems-lab
$ cd open-problems-lab
$ python prize_work/ppl-076/reciprocal_aps.py
How it works

How a result is checked

  1. 1

    Start from a published anchor

    Every job names the published result it starts from and must first reproduce known values exactly.

  2. 2

    A second, independent method

    Each new number is recomputed by code that shares nothing with the first script.

  3. 3

    Referee re-run

    A separate agent re-runs the scripts from scratch, compares every committed number and reads the wording for overclaims.

  4. 4

    Novelty class

    Each result is labelled reproduction, computed extension, negative or internal, from a literature search on a stated date.

What's inside

What the notes say

Sun's alternating-primes conjecture

Checked for every m up to 10⁹ with 0 violations. The published check stops at 10⁷.

Erdős problem 710

Exact values of f(n) at n = 15,000 and 20,000, past the table in OEIS A390246 (every n up to 10,000), which the lab's values match.

A negative result on Erdős problem 50

One natural way to certify the density f(c) cannot work, with the measurement that shows it.

Erdős problem 86

A one-line rule for where a published congruence lowers the cap, checked for every n from 3 to 600.

Reproductions

Re-enumerates r_3(N) for N ≤ 40 in Erdős problem 142: all 16 record values in OEIS A003002 match.

Corrections kept in the open

Drafts that turned out wrong or already known are listed openly instead of being deleted.

Working today

Stated plainly

Good to know

Honest limits.

Straight from the README, so you know what you're getting before you try it.

  • Nothing in the repository is a proof, a solution or a prize claim.
  • No human mathematician has reviewed the work.
  • The referee and the panel are AI agents from the same family as the worker, so they can share blind spots on mathematics.
  • Literature search can miss a paper.
  • Finite computations say nothing about limits.
Under the hood

How it's built.

Built with

Language
Python
Testing
Automated tests

209 files in the repo · last updated 18 Nov 2026.

Browse the code

Recently shipped

  1. Correct the problem 710 claim (OEIS A390246 has n up to 10,000); monthly release of new notes

  2. Add ORCID iD, version and abstract for the archived release

  3. Publish 12 reviewed notes with scripts, outputs and referee re-run reports

Full history on GitHub
Questions

Good questions.

What does it run on?

Python, NumPy.

How do I get it running?

The commands in the “Reproduce a result” box above come straight from the README. Paste them into a terminal; the README covers options and troubleshooting.

What doesn't it do yet?

Nothing in the repository is a proof, a solution or a prize claim. No human mathematician has reviewed the work.

Is it free?

The source is public on GitHub under the MIT licence.

Who made it?

Juwen Lie, an R&D mechanical engineer in Batam, Indonesia, as an independent personal project.

Keep exploring

More from the same workbench.

Found a mistake? Open an issue.

Small, checkable computations on open maths problems