Re:All AI-generated code should be Open Source
by pcthrowaway on Monday November 16, 2026 @12:03AM
(#50514242)
I'll argue the case for permissive code not requiring users to redistribute changes.
Say you use some software to provide data you host to end-users. Say that software relies on some system package or architecture that isn't available on your server, so you make minor changes and rebuild it to run.
Now if you wanted to make a proper change to the software to accommodate different target architectures/packages, you'd do things very differently; you'd write some check and delegate to the provided apis based on that.
But you didn't do that, you did a find and replace.
This has no value to anyone else, and it also reveals internal information about your system that you may not want to reveal.
Say you use some software to provide data you host to end-users. Say that software relies on some system package or architecture that isn't available on your server, so you make minor changes and rebuild it to run.
Now if you wanted to make a proper change to the software to accommodate different target architectures/packages, you'd do things very differently; you'd write some check and delegate to the provided apis based on that.
But you didn't do that, you did a find and replace.
This has no value to anyone else, and it also reveals internal information about your system that you may not want to reveal.