30-04-2026, 03:55 PM
(This post was last modified: 30-04-2026, 03:59 PM by Mark Knowles.)
(30-04-2026, 03:15 PM)DG97EEB wrote: You are not allowed to view links. Register or Login to view.Hi Mark,
I ran a combination GPT and Claude.. Hope it helps
Ed
Thanks for that.
When I look at these kinds of AI researches as with the previous stuff I tried with GPT and Gemini I don't know how to react. My initial reaction is that this is brilliant and very impressive. However when I start to work through the details I am not sure. First, I wonder to what extent it is feeding back to me in a slightly garbled/misinterpreted form what I have posted on Nick Pelling's blog as I have posted a lot about the ciphers that I have located. However, funnily enough I like it's confidence which I find encouraging. I find the reference to "Andrea Barbarigo’s private merchant cipher" intriguing. I have been in touch with Dr. Iordanou before, so I could follow that up. Paolo Bonavoglia says he has searched the Venice archives for diplomatic ciphers of that period thoroughly, so either he hasn't searched it as thoroughly as he thinks or there isn't any more to be found there. It seems to not have picked up on the 1397 Milanese cipher in the Chronicles of Lucca by Giovanni Sercambi as the earliest example of homophones we know of.
It is right that the State Archives of Mantua, Modena and Florence Archives need to be searched from top to bottom as they have already yielded much interesting materials, but there may be more to be found there.
It may be worth confirming that I have seen the following that it mentions:
1404–1418 enciphered correspondence
Jacobo de Faitinellis, Dino ser Paci, Iohannello Thieri, Guido da Pietrasanta, Nicolao da Moncicoli, Nicolao Arnolfini.
Foreign Correspondence, b. 25, sez. 3, Rome
Look for the 1426–1430 enciphered letters: Ricardo de Mutiliana, Pandolfo Malatesta, unsigned Rome letter, Francesco de Catalenis.
---
The tendency of these AI to hallucinate or just misunderstand what was previously been written makes me wonder whether it's references are genuine and correctly dated.I think it would be worth asking the AIs for the precise "shelfmark" for these ciphers and where they saw the references to them.
I like to think that I am not so proud that I care where a new lead comes from and that I will follow it up regardless.