Using LLMs to trace alchemical knowledge and decode 17th century letters
A historian explores using frontier AI models to solve concrete historical puzzles, such as decoding 17th-century alchemical anagrams, arguing for systematic collaboration between AI labs and researchers.
read it The piece provides rare, concrete case studies on directing frontier models at niche historical problems rather than relying on generic summaries.
summary
A historian shares case studies of using frontier AI models to tackle concrete historical puzzles, including analyzing John Dee's manuscripts and uncovering parallel anagrams in Samuel Hartlib's alchemical papers. While some automated analyses duplicated past scholarship, others successfully linked texts across languages and identified subtle lexical patterns. The author argues that unlocking these insights requires better digitized archives, dedicated compute, and systematic collaboration between historians and AI labs.
key points
- Frontier models successfully identified a shared anagram for Hungarian vitriol used by both Isaac Newton and Samuel Hartlib, revealing a previously unnoticed textual link.
- Effective historical AI applications depend on tractable problems with digitized primary sources, multilingual reasoning, and verifiable solutions like cryptography.
- The author advocates for AI labs and funding agencies to provide free compute and API access to historians to support systematic research.
who should read it
Historians, archivists, and technical readers interested in practical applications of frontier AI models beyond standard chat interfaces.
discussion warm
consensus
Commenters enthusiastically share personal successes using AI for genealogical research and historical text analysis, noting that frontier models excel at finding subtle, cross-referenced connections.
pushback
Some participants voice skepticism regarding high hallucination rates when using LLMs to transcribe difficult historical handwriting.
FWIW, I have tried my best to make the 10s of thousands of books on https://SourceLibrary.org ergonomic for both agents and people. All feedback welcome.
To date, it is the largest collection of agent-accessible translations on the web. The mcp pulls texts and illustrations — and the API provides access to the embeddings. It’s free.
We are based at the Embassy of the Free Mind in Amsterdam, a UNESCO-recognized library of alchemy, magic and mysticism (among other related topics)
And, it’s worth saying, I’m very much inspired by Res Obscura’s line of curiosity-driven humanist inquiry…
Dive …
It highlights a real-world library of alchemy and mysticism that has been explicitly optimized with APIs and MCPs for AI agents.
Generated by an LLM from the article and thread. Check the sources before relying on it.