Re: [RFCv3] Counter-Proposal -- Interpretation of DFSG on Artificial Intelligence (AI) Models
Aigars Mahinovs <[email protected]>
| Newsgroups | gmane.linux.debian.devel.vote |
|---|---|
| Message-ID | <CABpYwDW++3X1L-DkQZ5KmkVcLK==PAx9=DwKtmB2TM4x82rb_w@mail.gmail.com> |
On Sun, 11 May 2025 at 22:12, Thorsten Glaser <[email protected]> wrote: > > On Sun, 11 May 2025, Aigars Mahinovs wrote: > > >> If you JPEG-compress a photo of the original document then uncompress > >> it, it *is*. > > > >Please, restore a document from the output of "wc". Or from the output > > Can you even read? > Just because *one* software process produces a loosely compression, does not mean that ALL software processes are just lossy compression. I specifically listed processes that *are* actually used in AI training processes and that area also are very well known statistical text mining algorithms that produce a non-copyrightable result. And even lossy compression ....... does not *always* produce an actual derived work according to well established copyright case law, which I specifically qouted.