Since when is there an obligation to reply to a question ?
While Andrea’s initial post was a bit over the top, I don’t think he deserves this: he isn’t insulting anyone and he’s only stating his opinions.
He has violated the rules severall times in this thread, hence my Initial warning. (Including insulting parts of the community, even if sneakily)
That could already have been a ban, but I chose not to do that. Output from LLMs is banned on the forums, hence the question. If it would have simply been a “yes this is LLM” or a “some parts are” that would be a basis of discussion, but it is not “only stating his opinions”.
1 Like
Well, strictly speaking, the rule is not quite so broad as that. From the FAQ:
It’s specifically in the context of “research”, and it says “not helpful”. I think using an LLM to generate posts that aren’t for research and not disclosing it is also “unhelpful”, to say the least; but if Andrea answers “yes”, it won’t result in him getting a permanent ban or even an immediate suspension. It’d be the basis for a discussion; and depending on how that discussion went (though admittedly given how this one’s gone so far, I don’t have high hopes about it) we would probably decide what change (if any) should be made to the forum rules. And then after that anyone contravening the rules would get sanctioned, yes.
1 Like
At a minimum, when the question has to do directly with one’s behavior on the forum and comes from the moderators in an official capacity.
2 Likes
I think the opposite. It is a symbolic gesture, yes. But it is one that allows me to show that software can be built without LLMs, and LLMs are not necessary, and whatever Big Tech is doing is just a waste of energy with little to no results.
5 Likes
I don’t understand this ban decision.
There is no rule base for it, and it will only make the discussion even more heated…
If you want to change the rules about the use of LLM in forum posts, please first start a discussion about that before you start banning people that may disagree with you. Otherwise it’s just a power grab…
4 Likes
The suspension isn’t for using LLMs, but for refusing to answer a simple question about whether or not they’re being used. (And it wasn’t just me asking the question, nephele was also.) Had he answered the question directly, even if he’d said “yes”, I would not have suspended him.
1 Like
This whole “tutorial” has derailed pretty badly.
Can we at least temporarily lock the thread to cool things down?
5 Likes
As this topic amply proves, the forum rule against doing politics is basically impossible to uphold, even if several of us didn’t disagree with it outright. The reason is simple: everything people do together (and to each other) is politics. That includes a long-running project like Haiku. Just earlier today I was reading about the sad state of another open source project whose members failed to account for that.
Arguably this topic has derailed long ago and no longer fits the category it’s in, but it was provocative from the start anyway, and used the initial subject more as a pretext.
As I like to say: if you don’t do politics, politics will do you. From behind. With a police baton. (Apologies for being blunt.) Take the entire conversation as a sign that we need to work things out.
2 Likes
This look like a rule you just made up to justify a ban.
3 Likes
You don’t; nothing like this can be perfectly reliable.
But you start with self-reporting; that’s enough. Maybe detections get better, maybe they don’t.
From my perspective, containing LLM code is and should be a stigma; from yours it is not. That’s fine. Unless someone is ashamed of including LLM code, they should have no problem flagging their projects as such.
As for where to draw the line, containing even one line of code produced by an LLM is enough that a project should be flagged as “containing LLM-derived code”, mainly because that’s where I’d want to start ignoring it. 
A system that allowed for many different levels of preferences would have at least a few levels:
- contains LLM code
- significant %age of LLM code
- LLM-architected
For example.
2 Likes
I only registered today to check out the project, but what I found is a small, exclusive community full of barriers. Maybe focusing on evolving the project and accepting all contributions would be a better way to speed up this extremely slow process
I agree. Since the original tutorial intent has been achieved, let’s lock the thread insead of banning people for arbitrary reasons.
1 Like
There are frequently newcomers in the Haiku community. Some decide to stick around, some don’t, but most don’t seem to find us “exclusive”.
I don’t think this is true, at least the way you mean it. But barriers aren’t inherently bad. Dikes keep out the sea, farm fences keep livestock in or wolves out, and so on. The question isn’t “is this a barrier?” but “why is this barrier here, and does it make sense that it’s here?”
Including contributions that plagiarize projects under incompatible licenses, such as the GPL? Obviously we can’t do that. And if we can’t, then as a consequence we have to consider whether LLMs are a potential GPL violation (among other things.)
4 Likes
At what point in my post did I say it was about money? It is about energy consumption and environmental impact. An that does matter to me. If it doesn’t matter to you that’s your choice.
This thread was about using AI in regard to development of Haiku applications. And we are discussing the pros and cons of it. I would consider that very much on topic. If you label that political and left wing, that’s your problem.
2 Likes
You caught us on a bad day. But don’t worry it’ll all blow over.
3 Likes
Haiku project is too small to cause any measurable environment impact from using LLM agents. So Haiku is not relevant here. You should ask someone bigger like Google or Microsoft, which LLM agents usage may cause environmental impact.
1 Like
If you think it does not matter then there is no need to discuss it for you, your “propaganda” accusations are clearly missplaced.
2 Likes
That is not an argument at all. There’s always someone bigger with more impact that you can use as a reason not to do anything. We all have extremly little to practically no influence over what Google, Microsoft and other large companies are doing. But we can discuss how we want the Haiku project to position itself regarding important topics of the time, such as the use of AI. Which is exactly what we are doing here.
4 Likes
This thread moves too fast for me comment on all the things that I could…
That’s the same “reasoning” that smaller countries don’t need to do anything until the countries with large emmissions do. Never mind that some of the big polluters are actually massively investing in renewables, for example (see China).
It’s a typical debating strategy, it seems: First deny anything is wrong until that position becomes untenable, then deny it’s human’s fault until that position becomes untenable, then say “OK, you first”, and finally: it’s too late anyway.
In any case, the Haiku project may be a very small player, but it can show other projects - big and small - that we’re standing with them against the tide of AI. Together we may just save some of our computing culture.
9 Likes