On Sat, Nov 1, 2025 at 8:28 PM Charles Haynes <[email protected]> wrote:
It clearly goes further than that. If that were the goal then wouldn't a > clear copyright be sufficient? "all rights reserved, not to be used for > training AI without explicit written permission." > I cannot, of course, speak for the original author of this particular license. I am merely speculating about one particular (very large, IMO) use case. I further speculate that there might have been loopholes in the pre-existing licenses that this is intended to address. It goes much further than simply prohibiting training. "Even indirect use > of the software is forbidden. If, for example, a backend system were to > include such software, it would be forbidden for AI to make requests to > such a system." If one were to release a library under this license, and I > wanted to use that library in publishing my website, AI would be forbidden > to look at any content on my website. That's a much broader restriction > than simply not allowing training on the source of your library. > As I said earlier, one value of this approach is that it serves as an ideological line in the sand. One could imagine hybrid approaches (such as, e.g, when someone decides to release software under GPL as well as, say, BSD) which exist only because the ideological extreme exists to ensure the Overton Window doesn't shift too far to one side. Udhay
-- Silklist mailing list [email protected] https://mailman.panix.com/listinfo.cgi/silklist
