GPT-6 Astra Cracks Napoleon’s 1809 Encoded Letter in About 6 Hours

Written by

in

AI engineer Carter Church gave OpenAI’s GPT-6 Astra a low-resolution scan of an encoded letter and a single objective. About six hours later, it produced an interpretation of Napoleon’s wartime briefing to a general. The model identified 1,300 handwritten symbols and 155 distinct characters across 24 lines of code, then combined a simulated annealing solver with surrounding letters to complete its analysis and verification.

Source:dailygalaxy.com · Read Original

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *