|
|
|
|
|
by Rochus
22 days ago
|
|
> designed to maximize source availability Right. It depends on what you mean by "use"; GPL maximizes use in the sense that it prevents anyone from taking the code proprietary and thereby restricting future users' access. But it doesn't touch my actual point, which is that GPL explicitly permits commercial use, broad distribution, and also LLM training (none of which are restricted by the license). The source availability requirement is the condition, not a restriction on who can use the code. > why would anyone pay if they can just use the source Red Hat, Qt, and countless others have built commercial businesses on GPL code. So apparently there is a business and people willing to pay even if the source code is available. But that was not my point anyway. |
|
But I can still understand the Kefir author; if you previously defaulted to GPL (over MIT/BSD) mainly because you wanted to foster an open-source software development culture, then the emergence of LLMs might well be a turning point were publishing your project makes no more sense to you; thanks to LLMs, publishing your open-source project might do more for commercial closed source actors (via better trained LLMs used by the developers they employ) than for open-source developers (or open source culture overall).
Instead of potentially creating a valuable GPL project that pushes other users towards GPL/open source, you might end up making your project a sort of "commodity" easily available to all closed-source developers for a moderate cost in tokens...