Who is trying to do that, please? All suspensions and other restrictions that have been carried out have been actual suspensions, not “shadowbans”. The users suspended get a notification about what has happened, and we usually post on the forum explaining what’s been done and why. This is the complete opposite of shadowbanning.
Again, the vote was about using LLMs for translation. Not using LLMs to create “contributions” broadly. That is banned and is going to stay banned even if we do explicitly allow LLM based translations.
We (those opposed to LLMs) have explained this repeatedly on the other thread. Perhaps you disagree with our reasons, but we do in fact have them. My biggest ones have nothing to do with ethical concerns about who makes the technology (though sure, that is a factor too, as it is with everything.)
Most of the community are not contributing code to Haiku. The number of people who have contributed code, at least, and who are still in favor of LLMs is much smaller. There are some of those who are in favor of LLMs in general but actually against it for Haiku itself.
The community at large’s opinions are noted. But they’re not the ones who determine the actual technical details of how Haiku is developed. They may determine what features we prioritize implementing, or what bugs are more important to fix, or any number of other things like that. But those are outcomes, not the means by which the outcomes are accomplished. So demands that the development team not just achieve some outcome, but achieve it in a specific way, seem out of place to me.
If you’re talking about @nipos, his posts have been flagged and made to be edited in threads on this subject. He’s not the only one of the anti-LLM camp that the mods have done that to, either. So that’s not true.
Part of the problem here is that both sides are very heated. We can pick out plenty of emotionally charged language from both sides over the past two days that under normal circumstances probably would have been pushed back on by the mods were it happening in isolation. But it’s not, and people are already mad enough, so that means we’re actually moderating less than we would under normal circumstances. Because under normal circumstances, if one person flies off the handle, people barely blink if the mods step in to do something about it. Whereas now, if we were to try and restrict people from saying emotionally charged language like “inquisition” on the one hand and “spam everything with their shiny new toy” on the other, both sides will just get even more mad at us.
Personally I have had so many emotionally charged arguments (in real life too, not just online) that I’ve gotten used to things like that being thrown around, and I try to take the emotion into account but stick to replying to the substantiative claims. I’m rarely offended by such language, it’s people’s way of expressing the strong emotions they’re feeling. But other forum members are certainly not the same here and will take offense at things that are meant to offend, so I have to take that into account too.
