Re: GR: Ban LLM contributions from Debian

On 2026-07-23 05:14, Theodore Tso wrote:
> On Wed, Jul 22, 2026 at 07:45:34PM -0500, Matthias Geiger wrote:
> Consider a patch that might do something like this:
> 
> 	ei->i_reserved_data_blocks += nr_resv;
> 	trace_ext4_da_reserve_space(inode, nr_resv);
> +	spin_unlock(&ei->i_block_reservation_lock);
> 
> 	return 0;       /* success */
> 
> To add a missing unlock on a spinlock.  Suppose this was generated by
> an LLM.  Do you *really* think this would actually be a copyright
> issue?
More examples:

  - what if the patch just deletes lines?

  - what if the implementation is entirely obvious? Ranging from adding
    a missing 'return;' to obvious programs such as FizzBuzz

  - int main() { printf("hello, world"); } is from a copyrighted work
    by K&R, how come everyone is still using it? (fair use)

  - Given some prompt: if a model were to produce the same output O
    with training input T either included or excluded, to what extent
    could O even be a derivative of T? And is that extent sufficient
    to be counted as "derivative" under copyright law?

Best,
Christian