> I just want the tools to be built on ethical datasets based on scrapping only Public Domain and CC-0 resources before using them (or the work of artists who consent).
Just a side note, but this seems like a movement of goalposts. Surely if a model is open weights with mega-attribution, CC-BY or CC-BY-SA is fine?
Not to mention Adobe Firefly already exists which would meet his definition of ethics.
Even the Creative Commons themselves have said something different:
> In the context of an enforceable right, the ability to opt out from such uses must be considered the legislative ceiling, as opt-in and consent-based approaches would lock away large swaths of the commons due to the excessive length and scope of copyright protection, as well as the fact that most works are not actively managed in any way.
Source: https://creativecommons.org/2023/10/07/making-ai-work-for-cr...
And let's not forget what free culture is about:
> The free-culture movement is a social movement that promotes the freedom to distribute and modify the creative works of others in the form of free content or open content without compensation to, or the consent of, the work's original creators
Source: https://en.wikipedia.org/wiki/Free-culture_movement
It's about giving power back to the users and consumers, not to artists or developers.
It's nice to see the project isn't ending because of this though, and only because he felt it was good to end the story on a high note.