LLMs decode 17th-century alchemical letters, tracing hidden knowledge

A piece published on the Substack newsletter “res obscura,” titled “Using LLMs to trace alchemical knowledge and decode 17th century letters,” was flagged via Hacker News on September 24, 2026.

The available material for this item consists of a headline, an original title, a URL, publication metadata, and a source attribution of “Hacker News.” No excerpt, transcript, or summary of the essay’s actual content, methods, or findings was supplied.

Based on the title alone, the piece appears to concern the application of large language models to historical correspondence from the seventeenth century, with some connection to alchemical subject matter. The URL slug suggests the author may also address funding priorities among AI labs with respect to historical scholarship, though this cannot be confirmed without the underlying text.

No claims can be verified regarding what the letters contain, what techniques the author used, whether existing transcription or translation methods were inadequate to the material, or what results, if any, were obtained. Likewise, no claims can be verified regarding the reception of the piece beyond its appearance on Hacker News, nor regarding any broader argument about AI labs and historical research funding.

The Specola notes this item for the record and awaits fuller source material before further characterization of its contents or claims.