Re: Ballot option: Allow AI-Assisted Contributions

On Wed, Aug 05, 2026 at 06:51:45PM -0500, Andrea Pappacoda wrote:
> 
> I think this wasn't the point at all. The issue may arise when the model
> reproduces someone else's code in my codebase. Of course it's not going to
> reproduce a newspaper in my C project.

Sure, but the argument that people have been making is that **all**
output from an LLM might be subject to copyright, and this was the
"open question".

A human being could accidentally reproduce code when programming code,
or they might cut and paste some sample code from Stack Exchange (say,
to send SMTP mail in Python), and then perhaps they might change the
variables from what was used in the Stack Exchange example, or the
implemention of Quicksort from the Sedgewick Algorithms book (which is
most definitely copyrighted).

A human being **could** do that --- but we don't say that we should
ban all human generated code because it *could* happen.

And yet, the argument being made in the GR is not that we should be
judging LLM output on a case by case basis, much like we do with
human-generated code --- but that all LLM generated code should be
banned.

						- Ted