Dear ALL!!
The hot weather in Japan has become a little more autumnal.
How are you getting on ?
Here is a report on the use of new AI.
I’m experimenting with Claude on Tinderbox 11.
Even using the free version with Claude ver.0.13.37
They did the note creation to Tinderbox 11.
Search for the contents of the 2025 Moyse.tbx you created, and
Are there any contradictions in the content written?
“Let’s take a look at the list of notes to see how it is structured.”
When running Claude, the note name is not recognized and the name is changed.
It does not recognize the French string Époque with an axon.
The answer was given, so in future analysis, French-style, which is cited in the context,
Perhaps the German umlaut could be a problem.
For now, we have improved the name to a simple one as above. Claude has recognized the Tinderbox11 note. It seems that when searching using Claude, they were searching except for [.tbx], which has the extension.
When you chat with other LLM (LM Studio & gemma-3-27b or mistral-3.2-24b), only the answers that are convenient for the person asking questions will come up. (The answer was that the program was programmed in a setting that gave a convenient answer and had no objections.)
The answer to this Claude question is
“There seemed to be a problem, so I tried doing something else.”
The following suggestions were presented and implemented:
—(Quote)
But there were limits:
The comparison table shows the correspondence of concepts, but as stated that “detailed correspondence analysis is required,” the specific philosophical content development was insufficient. So, I:
Deepen each correspondence philosophically
Added linkages to other related concepts (Proust, Deleuze, Hugo, etc.)
Specific relations with musical practice
This is what I did.
This means that this comparison served as a starting point, and from there it developed into a more detailed and comprehensive analysis.
Was this usage appropriate? Or did you want a different approach?
——(Quote end)
Claude’s point has revealed that there are some parts that were “lack of my work,” so is it possible to use it usefully?
Whether or not the proposed work from Claude will be adopted is
I can just make my own judgment, so can I take a step forward from the previous trial and error work on gemma-3-27b and mistral-3.2-24b? I think so.
The problems and questions
Since it is written in utf-8, there is no problem with recognizing it.
Claude responded, but
Q01: The “axan” type used in French, German, etc.
Anyone encountering the problem? Is there none?
Q02: I’m still able to access one note,
Is there a possibility that it will be accessible through multiple notes in the future?
Yours, WAKAMATSU
