Any code that is automatically generated is “merely the output of the program”. To say that therefore it is not generated seems strange to me. There is no doubt in my mind that clw and inc files are generated for you. You work at a higher level of abstraction and the lower level files are generated for you.
AI code assistants are also “just programs” but have a different format for the specification (sometimes natural language).
They still generate code.
The compiler has no idea whether the code it is compiling was written by hand by a human or generated by some other program working from some form of “instructions” or “specifications”.
It was only a question of time before someone did this, but its just made reverse engineering apps a whole lot quicker and easier which means if this output is joined up with an AI code assistant, will make it easier and quicker to port code into another language and platform.
AI code assistants will take plain English comments like what I expect this Ghidra script to output and turn those comments into code.
Programming is certainly getting interesting all of a sudden.
If this goes against Github’s CoPilot and Microsoft, then the Clarion Templates still look like the best code generating tool on this planet. Whilst some have cited Google scanning books and their lawsuit, Google didnt pass the books off as their own, unlike Github’s CoPilot and this could also be the first lawsuit to view AI’s as interactive databases, which brings data laws into play. A similar lawsuit is underway with artists and an AI that generates images in the style of different artists, which could also cement the view that AI’s are just interactive databases and the data belongs to someone else.