Notices by csolisr (csolisr@hub.azkware.net)

Arguably, the scene is split mostly in three:

• The pro-AI side, that's already building entire vibe-coded applications, and doesn't mind much about the copyright and environmental qualms of LLM training;
• The neutral side, that doesn't mind as long as there are some safeguards, and will leave the copyright question to the future because LLMs seem so useful and productive today, that opposing them directly seems like reactionary Luddism;
• And the anti-AI side, opposed to be associated with even the neutral side and, unfortunately to themselves, in the cultural minority, forced to use their limited resources to fork the last untainted versions, settle for a restricted subset of the technologies of today, and be fundamentally unable to interact meaningfully with the rest of the community for ethical reasons.

I don't expect the anti-AI side to be sustainable in the long run, but I expect it to be persistent, same as the FSF purists that refuse to even mention proprietary software. To an extent, there is already some overlap between both groups: shunning the majority of the community, requiring a high degree of purity and divestment for exclusively ethical reasons, and either attempting to rebuild the wheel the hard way in order to keep up with the rest of the world, or just learning to go without wheels instead. In fact, projects like Hyperbola already have taken both stances simultaneously (along with rejecting the ubiquitous systemd architecture), so it's not entirely a hypothetical.