How AI Solved a 370-Year-Old Cipher
Claude Fable 5.1 deciphered the Cyphral Distich by recognizing a book-cipher pattern humans overlooked. Here is what this means for document analysis and historical research.
Short answer
An advanced AI model, Claude Fable 5.1, successfully decoded the Cyphral Distich, a 64-number cryptogram left by Sir Thomas Urquhart in 1653. The model identified that the numbers were not a standard substitution cipher but a coordinate system pointing to specific words within the book itself. While this is a significant milestone for historical research, it does not indicate a new threat to modern business encryption, which relies on entirely different mathematical principles.
What the source actually says
The source article details how the AI was tasked with solving an unsolved historical puzzle. The Cyphral Distich consists of two lines of 32 numbers each, appearing at the end of Urquhart’s Logopandecteision. Previous human attempts failed because they applied standard cryptanalysis techniques like frequency analysis, which assume a static substitution alphabet.
The AI realized the key was internal to the text. The book contains 32 "Proquiritations" (paragraphs). The rule is: for the i-th number in the cipher, go to the i-th Proquiritation, find the word at that index, and take its first letter. This revealed a royalist prayer: "O GOD UPHOLD KING CHARLS THE SECOND AND MAKE HIM THE SUPREME RULER OF THIS LAND."
The source also notes the AI applied the same logic to a larger, 285-number cipher in Urquhart’s The Jewel (1652), decoding most of it by using page numbers as indices. The author emphasizes that the AI’s success came from persistence and pattern recognition rather than complex cryptographic math.
Why it matters
This event highlights a shift in how AI handles unstructured, historical, or ambiguous data. Traditionally, solving such puzzles required a human to spend hours or days cross-referencing obscure texts. The AI demonstrated the ability to hypothesize a non-standard rule (using the text as a key) and verify it quickly.
For small businesses, the relevance is not in breaking ciphers but in document analysis. Modern AI can now identify patterns in large datasets that humans might miss, such as inconsistencies in contracts, anomalies in financial logs, or hidden structures in legacy code. However, it is important to distinguish this from security risks. The Cyphral Distich is a book cipher, a simple substitution method. Modern encryption (like AES or RSA) used by businesses is mathematically resistant to this type of pattern-matching attack. AI cannot "guess" a modern encryption key by reading the encrypted file.
What to do this week
There is no immediate action required for your IT infrastructure. This story is about historical analysis, not a new vulnerability. However, you can consider how AI-assisted analysis might help your business:
- Legacy Data Review: If you have large volumes of unstructured text (old contracts, manuals, or logs), AI tools can help index and find patterns faster than manual review.
- Verify Sources: As with any AI output, verify results. The AI’s solution to the Octastich had minor errors due to transcription issues in the source text. Always check AI-generated insights against primary sources.
- Secure Your Own "Keys": While AI can’t break modern encryption, it can help attackers find weak passwords or social engineering hooks. Ensure your business uses strong, unique passwords and multi-factor authentication.
What practitioners are debating
Hacker News discussion focused on the nature of the AI’s achievement. Some users pointed out that human cryptographers had previously identified the cipher as a book cipher in 2014, suggesting the AI did not discover a new method but executed a known hypothesis efficiently. Others debated whether the AI’s success was due to the problem being "easy" in hindsight or if it represents a genuine leap in pattern recognition. The consensus is that the AI’s ability to self-direct and persist through dead ends is more impressive than the solution itself.
When to call IT
If you are concerned about the security of your business data, focus on proven security practices rather than AI-driven cryptanalysis threats. Call Simple IT SRQ if:
- You need to audit your current encryption standards.
- You want to implement multi-factor authentication across your team.
- You are looking to integrate AI tools for document analysis and need guidance on data privacy and security.
For more on securing your business data, visit our services page.