Add post: Community Considerations Around AI Engagement #265
Conversation
There was a problem hiding this comment.
Nice read 😃 Happy to get this in. An opinion piece but I think this is in scope and personally would like to see more such articles in our sphere.
The note I left is just a note, not meant to be seen as something to change or anything.
(Just waiting for 1 or 2 more green lights to merge.)
| But LLMs are unlikely to ever produce meaningful license updates, and by this stage they've read and assimilated most of our library codebases. | ||
| Patterns that were once copyrighted are now commonly duplicated and fully generic. | ||
|
|
||
| Personally, I am pulled in two directions: first, I care about people getting credit for the work they do. On the other hand, and also the reason we do not employ copyleft licenses, is that we really want our work to be used widely and impact as many lives as possible. Credit is important, especially for younger people starting their careers, and we'll have to continue thinking about how to give it justly. Practically speaking, however, I'm not convinced that licenses are an effective mechanism to enforce credit anymore. |
There was a problem hiding this comment.
I am still on the fence on that point personally. I think Gael wrote that somewhere, that we are making gifts when contributing OSS. We should have the right to say how we want our gifts to be used. So it's not just about credit for me, it's about my own right and freedom to decide how I want my productions to be used.
+100 though on the positive social impact and still why I like OSS. But I am definitely sharing less stuff now–besides things like grant work or else where there is an expectation. I have now the luxury to not need extra credit to keep some of my principles 😅
There was a problem hiding this comment.
Thanks, interesting point! Yes, again I think I feel philosophically drawn in one direction, but worried that practical mechanisms may have been exhausted. Open source licenses have been trampled on for many years—I remember some of those early GPL court cases for hardware boxes. But then we could at least look at binaries, source code, etc. and pretty easily see what was copied. With LLMs? 🤷 When you distill the entire corpus of libraries into a model, there's just no way they're going to even have mechanisms for providing attribution.
I'm not saying give up on attribution (I can think of a few ways in which Microsoft, Anthropic, and OpenAI can make amends ;)), just that licenses are likely not going to cut it.
|
I think this looks great @stefanv! I appreciate the personal elements - I think those will resonate with many readers. And the text is clear that this is a personal perspective not speaking on behalf of the entire community (although definitely starting a constructive conversation WITH the community!) |
|
Also, I'm OK to put this on my personal blog if it feels too personal a take; my reasoning was that this is a timely topic that we need to grapple with as a community, and that we haven't made much headway on during the summits. So wanted to prod it a bit to see what comes out. |
4 participants
pngquant.