In the 1980’s, there were no freely available C compilers (I guess Borland was inexpensive, but not free). The GNU C Compiler was the first such. The project focused on being bootstrappable, reproducibility from source with a fixed dependency chain.
Corporates took advantage of GCC, and continue to, but have also been happy that LLVM came along with its permissive licensing. LLVM was a free lunch for GPU deployment, letting corporates customize it to compile shaders without having to release the sources.
In the LLM era of mostly closed weights, closed and pirated datasets and inaccessible compute requirements, what would, or does, an open alternative look like? Or should something else be done?
No, I’m not saying that. How do you even get to that from what I’ve written?
That is the implication if open licenses can’t deter corporate leeching while positioning for collaboration with volunteers. Unless there is an ideal license unknown to most projects that avoids the tragedy of the commons.
The tragedy of the commons doesn’t apply directly to software like that. If you have GPL code and people use it according to the license and you don’t get paid, that’s not the tragedy of the commons. GPL code is a donation. That is the expected outcome. You are free to choose whatever license, like FUTO has, for instance, a weird mix of source available, open source, and proprietary.
Tragedy of the commons, in open source, would be someone modifying your code and distributing it on a closed source license. They took from the commons, without returning anything. But you have to take developer expectations into account. If someone does not expect this to happen and chooses the MIT license, well, they’re being naïve at best. But you could say the tragedy is in effect. If they are happy with that happening (as the case of Tannenbaum and Intel), then no tragedy happening.
Actually, it is about commercial value obtained from using GPL’d S/W (I don’t know why any volunteer would have wanted to donate S/W to corporates!). But some of the GPL’y projects atleast had self-interested funding of developers (by, e.g. Redhat). OSS’y S/W has had much less of that.
If you look at steps like the OIN, they tried to share IP amongst corporates rather than costs, so corporates do play along when they have to. Nothing like that happened for GPL’y or OSS’y S/W. LLM training is probably a licensing violation, and also the biggest possible leeching.
Sure, a “normie” using GPL’d S/W is on point, because volunteers donated their skills for that purpose. That part of the commons worked.