I want gpt to be able to behave within context, maybe optimise it to generate a simplification of data that it understands that can be reapplied to larger sets of data to change and alter the data, without needing to re read everything, redo everything or work in a linear push/pull request fashion or created a sort of memory buffer.
What is the limit to and how can I make an adjusted behavioural model that can be multi threaded in output and sort of work in a multi contextual awareness fashion, like multiple pages in a book?
Would this be doable as a plugin where do I find general info on creation of plugins and what I can and can’t do with the api or how I would get access or help to implement something that might not be api related?