Open QuestionsRESEARCH OBSERVATORY

FIELD NOTES / 01 05 OCT 2026

Old secrets.
Open questions.
Honest evidence.

Following the traces in historical ciphers and open mathematics. Every claim carries its sources, its limits, and its next test.

Explore the research

A public research notebook. Reproductions, tentative claims and verified checks stay distinct.

03cipher dossiers
01mathematics track
00new verified solutions
A good question is progress.
A plausible answer still needs a test.

THE RESEARCH DESK

Where the questions stand.

Inherited work is credited.
Candidate claims stay visible as candidates.

01 / HESSE-KASSEL · 1637Arithmetic replay

Malsburg’s
alphabetic block

The published period-five recovery is credited to Larry Beck with ChatGPT, continuing Daniel Bourdeau’s transcription. Mechanical consistency leaves manuscript readings and interpretation open.

What is established
Local arithmetic replay matches all ten literal lines in both the 523-position and 510-position streams.
What remains open
Ambiguous signs, code referents, and historical interpretation.
Next test & provenance

Review manuscript readings and source alternatives independently. A successful replay does not establish historical correctness or novelty.

Bourdeau’s source edition ↗
02 / RIKSARKIVET · R4306Unreviewed claim

A letter
to Amsterdam

The published baseline lists this cipher as open. A frozen proposed model and its arithmetic replay are available locally; the solution claim still needs independent review.

What is established
An open published baseline and mechanical agreement with a proposed model.
What remains open
Source fidelity, validation of the reading, and discovery authorship.
Next test & provenance

Review the exact model against archival evidence, re-encryption and source alternatives. A linguistic score alone cannot settle the claim.

Riksarkivet research baseline ↗
03 / RIKSARKIVET · R4282Open

The undeciphered
Latin letter

R4282 remains open. Its unusual signs and clear text make a source-grounded transcription baseline the next useful task.

What is established
The source identifies an attempted, unsolved cipher.
What remains open
A reliable sign inventory and a supported cipher model.
Next test & provenance

Build a versioned transcription and uncertainty ledger before attempting a bounded search.

Riksarkivet research baseline ↗
04 / MATHEMATICS · INITIAL TARGETUnproved

The Riemann
hypothesis

A configurable starting point for proof audits. Map the exact statement and dependencies, then examine one bounded lemma. No mathematical theorem has been proved by this project.

What is established
Clay Mathematics Institute lists the hypothesis as unsolved.
What remains open
A complete proof; finite numerical tests do not provide one.
Next test & provenance

Separate known theorems from assumptions and proposed implications. Make every proof gap explicit.

Clay’s problem statement ↗

Showing all 4 dossiers.

A STANDARD FOR EVERY CLAIM

Evidence before
elegance.

A readable decipherment may still depend on an uncertain sign. A persuasive argument may still contain an unproved step.

  1. 01

    Preserve the source

    Keep original inputs, hashes, provenance and ambiguous readings.

  2. 02

    Make the test repeatable

    Record methods, bounds, outputs and failed attempts.

  3. 03

    Invite independent review

    Review source fidelity, alternative explanations and proof gaps.

  4. 04

    Publish with its limits

    Attribute inherited work. State what the evidence does and does not establish.

A BOUNDED RESEARCH QUEUE

One good next question.

Prepare a brief. Run it in ChatGPT.
Record the result, then review.

Planned work

4 bounded briefs
  1. 01
    Malsburg source alternativesHistorical cipher · 30 min / 3 follow-up turns
  2. 02
    R4306 claim reviewHistorical cipher · audit recovered evidence
  3. 03
    R4282 transcription baselineHistorical cipher · preserve alternatives
  4. 04
    Riemann proof obligationsMathematics · one lemma at a time

These are planned tasks, not completed research. The private command-line ledger reserves usage, stores results and keeps an event history.

Your local planner

Allowance unverified

Enter values from your own ChatGPT account. This calculator stays in your browser; it does not verify, submit or run research.

Paused until the account allowance is observed.

No fixed weekly quota is assumed. Use the displayed allowance for the selected ChatGPT model or feature; Codex and API usage do not measure it. Variable-cost work needs a conservative estimate and a fresh account check after each run.

PROVENANCE / REVIEW STATUS

A trail you can follow.

Download queue configuration
S01
Malsburg source edition ↗

Daniel Bourdeau; alphabetic contribution credited to Larry Beck with ChatGPT. Source linked, not independently authenticated here.

Public source
S02
Riksarkivet cipher research ↗

Daniel Bourdeau. Baseline for R4306 and R4282; later claims require separate review.

Public source
S03
Riemann hypothesis · Clay Mathematics Institute ↗

Primary problem statement. An initial configurable research target; no project proof exists.

Primary reference
S04
ChatGPT research usage guidance ↗

Remaining tasks are shown in product. Model allowances vary; use the actual account display and reset.

Official guidance