Thanks a lot for your effort to make AI capable of programming JSL code.The improvements over the last few months and weeks are incredible!where did the improvements come from?
The new scripting index might be one of the levers:
https://jsl.jmp.com/
gemini3 pro can
All my current python is written with gemini3 pro.Interact directly with JSL.It was basically correct in one go.All are output to the JMP table
It doesn't work with gemini2.5 pro.
did somebody invest the time to compare the different models?some matrix: LLM vs. capabilities?- doesn't think that it's "JAVA Scripting Language"- basic JSL- Coding Concepts- Platform Calls- Display Boxes- Matrix Functions- Expression Handling- Dashboards- ...
This effect can be achieved the first time.The second fine-tuning was perfect.
gemini3 pro is really easy to use.
Ben Barroso-Ingham has a LinkedIn post showing an arcade game he coded in JMP with help from Claude. They may be writing a blog post on the topic, looking forward to that
gemini3 pro watched the video of this post.It claims that its effect must be better and faster