I think that in the near future a super object oriented language will emerge that will let us create computer apps instead of programming them directly. Systems analysts, not programmers, will be able to define a computer app with the objects at a very high level and LLMs will create a JAVA like blob that can be run on multiple systems. The advantage is that the objects will be completely bug free and the app size is only limited by the systems resources. Bottom line, most app development will be done by systems analyst. The AI object programmers, a specialty, will help tweak the underlying function of each object.

what llms reason best in might not be a language at all. more like a graph. graphs say more in less space than prose, and they stop the model from wandering off.

Because a library like cuda allows you to make code shorter?

Thinking twice, such a LLM would aim to create code that maximizes the power of the features it provides (rich api) and minimizes lines of code. So the LLM would aim to develop a language for reaching Kolmogorov complexity.