AI Solves 370-Year-Old Historical Cipher Using Internal Book Keys

Claude Fable 5.1 solved the 370-year-old Cyphral Distich and Octastich by discovering that the keys were embedded within the structure of the author's own books. By mapping numbers to specific word indices in the text, the AI revealed hidden royalist prayers for King Charles II. This breakthrough demonstrates how AI can overcome human attention bottlenecks to solve historical mysteries through persistence and pattern recognition.

Key Points

• Claude Fable 5.1 solved the 370-year-old Cyphral Distich by realizing the book's own structure served as the cipher key.
• The AI deciphered a second, larger cryptogram called the Cyphral Octastich using a similar page-based word index method.
• The recovered plaintexts are royalist prayers for King Charles II, which align perfectly with Urquhart's known political leanings.
• The AI's success highlights its ability to overcome the 'human attention bottleneck' by persistently analyzing obscure archival materials.
• The solution was simple in hindsight, demonstrating that AI can find patterns in historical data that humans have overlooked for centuries.

Sentiment

The community is mixed but engaged. Commenters generally accept that the solved cipher is interesting and that LLMs can be useful for neglected textual puzzles, but they are skeptical of expansive claims about AI intelligence and strongly concerned with attribution, prior art, and hype. The center of gravity is cautiously positive about the practical result and critical of the framing.

In Agreement

• The model appears to have found a specific internal key that earlier casual suggestions did not fully identify, so the result is more than merely repeating that the cipher might be a book cipher.
• LLMs can be valuable on obscure historical problems because they persistently test simple hypotheses, combine broad textual context, and pursue leads that humans may not have time to investigate.
• Even if the solution depended on scattered public hints, turning those hints into a complete, testable decipherment is useful synthesis rather than empty parroting.
• The case supports a narrower, defensible claim: AI can help unlock neglected archives and puzzles bottlenecked by attention, not necessarily outperform expert cryptanalysts at hard cryptography.

Opposed

• The article may overstate novelty because older online comments had already proposed that the solution should be found using the book itself.
• Some commenters suspect the model benefited from training data or internet-visible prior art without adequate attribution, making the result feel like repackaged human work.
• The cipher mechanism seems straightforward in hindsight, so the story may say more about limited human attention than about exceptional AI reasoning.
• The headline and surrounding framing risk feeding AI hype by implying deep cryptanalytic ability where the actual method was contextual pattern matching and trial of a simple clue.