works on mdsdrv

[garrettjwilke]

No description provided.

[Clownacy]

Wow, this looks great!

[Clownacy]

Yesterday, I received an email from the author which dropped a bombshell: the code was generated by an AI called Cursor!

I suppose, in hindsight, there were a few clues: the comments are unusually verbose, the ifle logic is painfully naive, and it is very unusual for somebody to know one of my codebases well enough to make such extensive modifications.

Still, the fact that AI can pull-off something like this is shocking to me! I can't even begin to imagine how it figured out the assembler's text substitution subsystem, let alone well enough to use it to implement EQUR! I'm not even 30 yet, and I already feel like an old man that no longer understands technology.

---

I suppose that it has become trendy to debate whether projects should accept AI-generated code, and now not even I can avoid it! So, what do I think about merging this?

First things first: I'm not going to cry 'baww plagiarism'. It isn't plagiarism for one person to learn from another, so it isn't plagiarism when an AI learns from someone either. Not to mention, this code is very clearly modelled after my code, not someone else's.

Secondly, it seems to me that there is very little in the code that could be considered copyrightable: the code is extremely bare-bones, so there is very little room to write it any differently; if I had written the code myself, then it would have looked almost identical, because there is only one (good) way to write it in the first place.

Finally, there is the matter of quality-control: I have my doubts that the code works as intended, especially when it comes to EQUR functionality. However, these are the same kinds of doubts that I have of any pull request: my standards for programming are quite high, and I do not merge code without extensive scrutiny. None of my projects are collaborative: they are my software and they will not include code that does not meet my standards.

With all of that in mind, I plan to merge this pull request eventually, but only after I have tested it, corrected the issues that it has, and verified that it does things 'the right way'. Again, just as I would for any other pull request that I receive.