qrsBRWNMar 10
Redox OS
Redox OS has added a Certificate of Origin policy and a strict no-LLM policy to our CONTRIBUTING.md: https://gitlab.redox-os.org/redox-os/redox/-/blob/master/CONTRIBUTING.md
CONTRIBUTING.md · master · redox-os / redox · GitLab

Redox: A Rust Operating System

GitLab
Mar 9 at 4:37pmWeb
Show threadDavid Fleetwood - RG AdminMar 9

@redox This is quite good. I'm still not satisfied with your response about fascist contributors however. I'm not trying to make you look bad, but pointing at a CoC says nothing about how it will be implemented in practice. My ask is if you permit harmful people to contribute, you have repeatedly declined to answer that, and it's quite important.

Many people do not feel safe sharing a space with bigots and fascists. They should not have to wait to be harmed for action to be taken.

Show thread🌸 Ⓐnarkitty lumi-nhac  Mar 9
@redox cc @alexia
Show threadAlexia Mar 9
@lumi @redox heck yeah. I should try Redox at some point
Show thread🌸 Ⓐnarkitty lumi-nhac  Mar 9
@alexia @redox so happy redox is taking a stance against genai
Show threadTechokamiMar 9
@redox GOOD. I was looking for what their views were regarding genAI recently but found nothing. Glad to see they're taking the correct stance, like Haiku
Show threadTechokamiMar 9
@redox Now we just need to wait for the driver situation to improve with Redox, starting with, uh... *checks notes* the USB stack?
Show threadJason Bowen 🇺🇦Mar 10
@techokami @redox Contributions to accepted :)
Show threadTechokamiMar 10
@jbowen @redox I am neither a Rust programmer nor a systems programmer, so aside from reporting what the developers already know, all I could do is make it worse.
Show threadthe kamala harris of going to bed at a reasonable timeMar 10
@redox first of all this rules but second of all did you use the reflecbeat font for your R on purpose
Show threaddivVerentMar 10
Thanks for the idea to finally do something about it.

I'm gonna have a bit more fun by not explicitly banning but setting requirements that are impossible to fulfill by current and likely any future models - and that are directly a consequence from the project being under an open source license, to highlight that LLMs are inherently open source incompatible (unless one trains one's own local LLM on only one's own code, in which case I don't care, but it's also unlikely such models can even be useful).

## AI Policy

Contributions by a LLM or other AI entity still must follow the above [licensing] requirements, just like human contributions.

In particular this means:

- The model or entity that has been used must be named.

- It must not have been trained or prompted on works with incompatible licenses. Neither must the output for any other reason be derivative of works with incompatible licenses.

- All authors of works that went into the training or prompting, or that the submission is otherwise derivative of, must be named and the respective licenses be stated.

- Note that an excessive number of contributors likely makes reviewing accuracy of the licensing information infeasible and usually leads to rejection.

- The human submitter certifies they have fully understood their submission and can respond to comments or questions about it.
Show threadJohn MaxwellMar 10
@redox - Thank you.