@musicmatze Hmm, I think prohibiting certain use cases (like analysis) is hard to get right. I.e. I think something like this would often lead you to restrict things you actually want to allow or allow things you really want to restrict.
Maybe a copyright reform is needed that defines how the copyright of input data affects the copyright of machine-generated outputs? Though that also opens a whole other can of worms ...
@musicmatze If we ever get to the point where machines can generate content on a human level, basically the whole concept of copyright breaks down. Just like the concept of capitalism will, if a large part of the labor force can be replaced by machines without simultaneously generating sufficient replacement jobs.
@musicmatze Given that we don't reach a plateau w.r.t AI research soon, I think it's going to be interesting to see how our legal systems/our societies will adapt to advances in AI that raise similar issues.
@musicmatze I have quite some code under GPL but also a lot under MIT. I think the most important part of these licenses is the "no warranty" part. The attribution stuff is nice to have. But really, what would I do against some company using my code in their proprietary binary-only commercial software? How would I even know they are using it? For them it's not legal, but well I actually expect this to happen.
@musicmatze I think it is more important to build something useful and am against more restrictive licenses. Also, I don't think any snippet copilot takes from my code is original to my software.
@musicmatze "Code analysis that is done remotely or reported remotely"