On 7/12/21 9:16 PM, marc wrote:
Hi, me again

So I am going to respond to multiple comments in one go:

I had a look at Julia Reda's post, and as far as I can
make out, she only focuses on the fact that individual
snippets are very short - but doesn't make any mention that
inserting *lots* snippets algorithmically is *all* that
copilot does...

Repermutating lots of snippets is not in itself copyright infringement, however- otherwise everything written in English would infringe on the copyright of dictionary publishers! That's not to say what copilot produces isn't a derivative work; you give a good argument to support the contrary. It just fails to follow straightforwardly from the fact that copilot produces "remixes" of existing code.

Another commenter said that the codebase used to generate
the model is just somehow the "input" and not actually *in*
model - but I am not sure the distinction is that clear.

You would be correct to wonder- if you're referring to my comment, I apologize for what was a warped and unclear explanation of a warped and unclear point. The intended argument was that copilot never executes or distributes compiled copies of the input (or the model derived from it) as *software* or part of any software, and therefore does not trigger the requirement of sharing under the (A/L/)GPL.

However, after further consideration, the requirement to include a license notice would pose a problem- and more generally, omitting or erring on the license notices of copilot's output could become a nightmare.

Attachment: OpenPGP_signature
Description: OpenPGP digital signature

_______________________________________________
Discussion mailing list
Discussion@lists.fsfe.org
https://lists.fsfe.org/mailman/listinfo/discussion

This mailing list is covered by the FSFE's Code of Conduct. All
participants are kindly asked to be excellent to each other:
https://fsfe.org/about/codeofconduct

Reply via email to