User:Mullikine-shane/sandbox

Imaginary programming is a programming paradigm in which a computer program's behaviour exists in relativity to language models, such as OpenAI's GPT-3.

To make an analogy, imaginary programming is a type of programming, where similarly to pure-functional programming, the subset of code who's behaviour depends on the output of a language model (either pending or precomputed outputs) stands apart from the rest of the code that has no such association. What is deemed 'imaginary' code is code that uses or is waiting on output from a language model and its behaviour is so altered by it. What is deemed 'ordinary' code is code which is not 'imaginary'.

A distinction is made between grounded and non-grounded imaginary programming. 'Non-grounded imaginary programming' may also be called 'pure imaginary programming'. If an imaginary function has the language model as a parameter then it is considered 'grounded'. If an imaginary function relies on the output of functions that have in the past used a language model as a parameter, but does not contain a reference to specific language model used then it is considered 'non-grounded' or 'pure imaginary'. Non-grounded code is still code in a similar way to how pure functional code is considered code.

In contrast to imaginary programming, ordinary programming refers to the style of programming where functions have no dependence on the output of a language model and thus have no imaginary dimension.

'holographic programming' is a type of imaginary programming where the language models being used are trained on a dataset that is comprised of software. The imaginary code being written, therefore, may employ associations made between elements of the code during the training of the language model, and possibly how the application is used, to build applications. Such appliations may be based on knowledge of how a program has been used in the past and not contain any of the original source code.