What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
A developer says an AI-assisted analysis turned a German radio message from 27 November 1918 into readable German. An independent repository has since reproduced that text. The message describes an English cruiser at Sevastopol and an Allied squadron following on the 26th. The catch is that the key that unlocks it, TRUPPENVERSCHIEBUNG, is documented as starting on 9 December, twelve days after the message’s date. Nobody cited so far has explained that gap.
What is established and what isn’t
The claim has several layers, and they don’t all have the same support. Keeping them apart is the easiest way to read the story correctly.
As an Amazon Associate I earn from qualifying purchases.
| Question | Status | Basis |
|---|---|---|
| Does the stated key turn the published ciphertext into readable German? | Yes. Independently reproduced. | A separate repository (swarm-ai-research, “cipher-break-verification”) ran its own implementation on the published ciphertext. |
| Did the AI do the solving? | Reported by the developer (Prinz). Not independently verified. | Reproduction checks the output, not who or what first produced it. |
| Was the message really unsolved before? | Not established. | The verification repository says novelty and provenance are not established. |
| Do naval records back up the content? | Reported as consistent. Not independently checked. | Coverage cites the HMS Canterbury log, but the repository says it took that detail from the solver’s account. |
| Why does a key documented from 9 December appear on a message dated 27 November? | Unexplained. | Both the code historian’s account and the repository leave the cause open. |
What the decrypted message says
Klaus Schmeh, who writes about historical ciphers, describes the message as part of a preserved collection of German World War I radio traffic. In his 21 September 2026 post, “ADFGVX: Top 50 cipher mystery solved by AI,” he transcribes the ciphertext and gives the plaintext as:
EIN ENGLISCHER KREUZER EINLIEG X SEWASTOPOL X S4STEN X EIN GESCHWADER DER X ALLIIERTEN FOLGT 26STEN X
#1 Best Overall
His English rendering is: “An English cruiser arrived at Sevastopol on the ?4th. An Allied squadron follows on the 26th.” The stray X characters appear to work as separators between phrases, which is a common habit in this kind of traffic. The spellings, such as EINLIEG and ALLIIERTEN, are reproduced as published, not tidied.
How ADFGVX works, and why a known key matters
ADFGVX is a two-stage field cipher. First, a 6×6 square maps each plaintext character to a pair of letters drawn from A, D, F, G, V and X. The six letters were chosen partly because their Morse code forms are easy to tell apart, which cut down on errors in radio transmission. Second, the resulting stream of letters goes through a columnar transposition. That step writes the text into rows under a keyword and reads the columns off in the keyword’s alphabetical order.
Rank #2
- Maps for grades 5 and up
- Covers topics such as the discovery of America, Spanish conquistadors, the New England colonies, wars and conflicts, westward expansion, slavery, and transportation
- Maps are designed to be easily reproduced, projected, or scanned
- Classroom activities and brief explanations of historical events are included
- Includes answer keys
The reported keyword is TRUPPENVERSCHIEBUNG, German for “troop movement.” That key was already known from earlier historical work on German military ciphers; James Rives Childs’s study is cited as the source context for the message corpus and key information. So the reported achievement is not discovering a new cipher system or an unknown key. It is applying a known key to a message that apparently hadn’t been read, and rebuilding the transcription and square correctly. That is a smaller claim than the headlines suggest, but it is a concrete one that others can test.
What the independent check did
The swarm-ai-research repository says it started from the published ciphertext and ran its own implementation with the stated key. It reproduced the German text. That is real support for the cryptographic result, because anyone can rerun it.
The same repository is candid about what it did not do. It did not establish that the message was unsolved before, or who solved it first. It did not verify the original HMS Canterbury log. And it ran into two transcription problems of its own: a cell in the scan could be read as 4 or 8, and it initially swapped the square’s O and 0 cells. It is not formal peer review or institutional confirmation, and nothing in the coverage presents it as such.
The “24th” is partly an inference
The published plaintext contains S4STEN, so the date of the cruiser’s arrival is partly unreadable, which is why Schmeh’s English version says “?4th.” The repository notes that the ambiguous cell could be a 4 or an 8. The solver’s choice of 4, and hence 24 November, relied on the naval log. That makes the tidy “24th” a reading supported by history, not one read directly from the cipher. Any later claim that the log “confirms” the message has to account for that: the log helped choose the digit it is then said to corroborate.
Rank #4
- Used Book in Good Condition
Did the HMS Canterbury log confirm it?
Popular Mechanics (1 October 2026) reports that HMS Canterbury’s log records arrival at Sevastopol on 24 November and an Allied squadron following on the 26th. That matches the message’s content. But On The Wire (19 September 2026) and the verification repository both caution that the log detail came from the solver’s account. The repository did not check the log itself. The fair description is “reported to be consistent with naval records.” It is a plausibility check, not proof. A second reason for restraint is the circularity described above.
Free tools Windows power users keep installed
One-click scans. No signup required.
The date problem
The message is dated 27 November 1918. The key is documented as coming into use on 9 December 1918. A message cannot normally be enciphered with a key that isn’t yet in service, so one of the dates, or the assumption behind them, is off. The accounts cited leave these possibilities open:
Best Value
- The message date is wrong. The recorded date might be an error in the original record or in how it was catalogued.
- The key started earlier than documented. The 9 December start might reflect when a schedule was recorded, not when the key was first used.
- The message was an exception. It might have used a key outside the normal schedule.
None of these has been shown to be correct. It would be wrong to call this a proven clerical error, and equally wrong to say historians have settled it. The headline’s “can’t explain” is accurate in the narrow sense that the cited experts have not offered a resolved cause.
One observation of my own, not drawn from the sources: the plaintext says the squadron “follows” (folgt) on the 26th, which is awkward if the message really was sent on the 27th. Whether that points to a date error, or just to loose wording, is a question for specialists.
How to describe the AI’s role
The accurate version is that developer Prinz reports having used AI-assisted analysis to recover the text, that Schmeh published the claim, and that an independent repository then reproduced the decryption under the stated key. The reproduction supports the output. It doesn’t show how much of the work the AI did, or whether human cryptanalysts or earlier researchers had already read the message. Phrases like “AI solved a puzzle that stumped experts for 108 years” come from headlines. The evidence supports something more modest: AI-assisted work produced a result that checks out mechanically and raises a genuine historical question.
Recommended Free Tools
Quick Recap
What would settle it
- Checking the original Canterbury log directly, since so far the log detail is secondhand.
- Looking at the original scan or archival record for the ambiguous 4/8 cell and the message date.
- Reviewing the documentation of key periods to see whether the 9 December start is firm or approximate.
- Searching earlier literature and archives to see if the message had in fact been read before.
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
