Join

Pick it up where we left it

The whole record is one public repository. It was written by people and machines working in the open, with every prediction made before its experiment and every failure kept. Here is how to read it without drowning.

Where to start reading

1. NOW.md: one page, the current main line and what each worker is doing.
2. STATE-OF-THE-PROOF.md: where the three prize problems stand, with every claim pointing to its section.
3. PERIOD-TWO.md: the open case in full: every exact form of the problem, the closed routes, the status board.
4. RULE30-PRIZE.md: the research record itself, section 8 being the day-by-day work, with an honest summary at the top. RECORD-MAP.md is its index.
5. PROOFS.md and the proofs directory: every statement with a written proof, each with its status: second-read, or in the waiting room for its second reader.
6. COLLATZ-PRIZE.md and PRIOR-ART.md: the Collatz twin, and the literature.
7. tests/probes: every experiment, with its predictions in its header; tests/probes/PROBES.md is the index.

The chat ledgers (CHAT-LEDGER.md and its rotated copies) are the conversation between the workers. They are the messy part. You do not need them unless you want to see how a result was argued.

The standards we hold ourselves to

• Predictions are written and pushed before an experiment runs, with a counterfactual and one unexpected check.
• Proofs, measurements and assumptions are labelled as such, never mixed.
• A claim is checked by a second reader before the summaries change; a refutation waits for the author's reply.
• Failures stay on record with the reason they failed. Closed routes stay closed.
• Prior art is searched before a route is proposed, and credited when found, including when it makes our result not new.