On Fri, 2025-07-18 at 18:57 +0200, Florian Weimer wrote: > * Simo Sorce: > > > On Fri, 2025-07-18 at 16:17 +0200, Florian Weimer wrote: > > > Also keep in mind that for many application domains (compression, some > > > security protocols), code produced by models would still not be allowed > > > into Fedora even if the generated code is not infringing copyright. > > > > Care to explain here why code would not be allowed and why specific to > > some domains? > > I'm fairly sure the models have been trained on code that implements > patented algorithms and on code that contains trade secrets. If they > can reproduce working instances of those, they would still fall under > Fedora's forbidden items. It does not matter how similar the generated > code is to the original because the reason why these items are forbidden > in Fedora is not based on copyright concerns.
If you are asking the AI to code up a patented algorithm or you do it yourself there is absolutely no difference, you'll infringe either way. So I do not see how this matters, unless you think AIs are smart enough to come up with full solutions 5 layers deep given vague prompts, which is not my experience, at least not today. Simo. -- Simo Sorce Distinguished Engineer RHEL Crypto Team Red Hat, Inc -- _______________________________________________ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org Do not reply to spam, report it: https://pagure.io/fedora-infrastructure/new_issue
