thank you for adding that clarification, but now I have to rephrase it by saying Haiku is needed in other place beside in own OS, such example being Hyclone to allow Linux slop to use haiku apps and technology or OSNews: "First bits of a Haiku compatibility layer for NetBSD" This that apparently allowed BSD to run haiku applications
This should make your guts twist instead: https://www.researchgate.net/publication/403073048_The_data_heat_island_effect_quantifying_the_impact_of_AI_data_centers_in_a_warming_world
We estimate that the land surface temperature increases by 2°C on average after the start of operations of an AI data centre, inducing local microclimate zones, which we call the data heat island effect. We assess the impact on the communities, quantifying that more than 340 million people could be affected by this temperature increase.
2 Likes
Only this ? Bitcoin, oil industry, plastic, war… ??
1 Like
Yes, all those are big problems too, but none of them are directly relevant to Haiku. Generative AI is. That’s the one we can do something about, by sending it where it belongs.
2 Likes
You could limit a ban of LLMs to Haiku & Haiku-related software on this forum. Unfortunatly some big foss-projects that Haiku (and a lot of software in general) depend on have decided that allowing slop code would be a good idea. I don’t hink we can do much against that but we can at least shape our ecosystem.
I personally did not think this would be good and instead suggested banning slop, which would be better in my opinion as it would stop a flood of slop.
In there local temporal subcontexts these arguments were important. Also slop machines just feed on tons of resources which has a big environmental and ecological impact.
Moderation is quite taxing and if you have an (possibly endless) source of slop that you have to read through, you cannot ensure a even remotely civilized and good discussion. And I’d generally consider unmoderated forums dangerous.
Iirc, most points from that list were already pointed out on this thread and the “tutorial” thread and I don’t think that LLM slop and packaging are comparable either. (Of course you could have said that in a softer tone, but I think nipos did a good sum up of some of the negative aspects of LLMs).
Maybe because a lot of people don’t want to have the software ecosystem flooded with slop and thus state their opinion. Of course this generates a lot of noise. This noise now might just save us a lot more noise on a later date.
Slop also on a larger scale clogs up information channels and makes misinformation harder to spot. Misinformation that can cause devastating effects and in the worst cases spur of hatred under which people will suffer and possibly die.
I think I already said this previously, but a lot of funding for LLMs comes from those that also support a certain administration that wants to see the likes of me dead. And LLMs happen to be a very good tool in flooding the zone with bs which very much helps said administration.
Probably, but I don’t see how using something that is directly damaging to us (If we have the choice to avoid it) would be sensible.
Many LLM scrapers ignore robots.txt and logins are not suited for open services. I had my cgit instance already hammered by some scrapers and that instance is tiny. The way I use it though, neither anbis nor a login would work well.
Besides having a slight taste of whataboutism, I don’t think anyone is seriously using Haiku to professionally mine bitcoins without hardware acceleration
3 Likes
Did you read my original post? Maybe it got buried in the flood of posts here. I proposed a way to discuss and formulate an AI policy so we can get along without any thread being sidelined by the same discussion over and over again.
But it seems that little sentence about package management is all that people are interested in. It doesnt have any significane to what I was trying to get across. I could have left it out. But I included it because I’m human and not a robot. Humans sometimes tend to include non necessary information when they are talking. So that’s what I did. I didn’t anticipate that it would completely deflect from what I was trying to say.
3 Likes
The hypocrisy is deep in this one.
No one seriously thinks that you are going to stop global warming by banning any mentions of LLM-coded projects from the haiku forum. Not even you.
No one seriously thinks that we are going to stop DDOS scraping attacks by banning any mentions of LLM-coded projects from the haiku forum. Not even you.
This is a purely symbolic gesture that will have no effect on the wider world whatsoever. And that’s fine. Symbolic gestures are also important. But then they have to be consistent. Don’t stand on priniciple one moment and then say “well, there’s no real practical effect” the next.
Practical effect is the pro-LLM camp’s point of departure, not yours.
2 Likes
No, your original post very much is “whataboutism”. You are correct that a bitcoin miner should probably be not in haikuports… sooooo maybe make a post on the haikuports mailing list? It is not related to LLMs. You can’t when someone goes “There is a big problem with A” go “yeah but you personally did not solve B, so clearly A should not be solved either”…
This is not true. The is not a symbolic gesture, and will have tangible impacts on this forum and surounding community. I don’t see how you can see this is symbolic or a gesture.
But hey, you know, most moderation actions here don’t change geopolicical developments.
3 Likes
I also don’t think the world will be any less cruel and uncaring if I strive to be kind and helpful. Of course one person’s little efforts aren’t going to change anything by themselves. Or even one little community’s efforts. So… who’s supposed to take the first step?
1 Like
It became symbolic when your side started bringing in global warming and scraping attacks. Which you are NOT going to solve by this action. The argument is no longer “This will be bad for Haiku”, because when you refuse to look at the actual application, you can’t actually know that, so you are reduced to just using the term “slop” as much as possible and associating it to vast concerns that you know perfectly well are beyond anything you can solve with this.
Look, just admit that you really, really don’t like AI. You hate LLMs. You don’t like what they are doing to the computing world and you don’t want to be reminded of that reality when you come here to your beloved forum. I can respect that. But the rationalizations are getting whackier by the day.
3 Likes
So instead of engaging with people and helping to wean them off their reliance on LLMs, showing them a better way, you support banning any mention of their work on the forum, essentially telling them to stay away from Haiku.
That’s a definition of “kind and helpful” I am unfamiliar with.
2 Likes
My side? Sorry, that is just a baseless simplification. Clearly these aspects are important, but I never claimed they were the main justification for this.
What’s next? are you going to have a little rally for the “AI lovers” Side? or maybe we can get back to the topic at hand, and continue debating with arguments because, so far the “other side” has not really given any good justification of why LLM content should be allowed on the forum.
2 Likes
It’s not the place of people here, or off the Haiku forum, to tell people they are wrong for using this, clearly most users here think this is reprehensible. But that doesn’t mean we should create a hostile place for these people. We are also not missionaries whos job it is to show people the “light”. If someone is using an LLM, has not programmed before, and now wants to learn doing it without an LLM they can just, ask, you know? There is no need for us to go evangelize in each forum topic. (Like some people in the AI crowd are doing currently, case in point the claude topic)
How would you feel if every time you posted something with an LLM and you get (albeit justified) hostile responses about your use of these technologies? I don’t expect everyone to know the impact of this stuff.
In any case. This topic, especially for the use of “slop” coding seems incredibly devisive. So simply not having this topic on the forum seems to be the best option for our collective peace of mind.
2 Likes
Kind is not the same as nice. I tried being nice, and look what it got me.
So far, it’s been the AI people dissing Haiku developers and the real results their efforts brought.
It’s been the AI people insisting that we carve out a space for them, too. Why should that be needed if LLMs contributed something of value to everyone here?
It’s been the AI people engaging in whataboutism and false equivalences. Dismissing everyone else’s concerns out of hand. Asking for a “compromise” that would only benefit them.
I trust our moderators and will support their decision either way, but hoo boy.
3 Likes
Nobody seriously thinks that if they vote or don’t vote in an election it will make a difference. But they vote anyway because they hope that thousands of other people will vote the same way, and that WILL make a difference.
Nobody seriously thinks that if don’t they litter the highway with sweet wrappers as they drive along that it will make a noticeable difference. But it DOES make a difference when lots of people do the right thing.
And if a good chunk of people refuse to support AI then again, it WILL make a difference.
4 Likes
I will absolutely defer to your experience on whether it’s practical.
And @PulkoMandy doesn’t think it’s appropriate for Haiku Inc. to go that far anyway.
I suspect that’s the majority opinion among the developers, for various reasons?
Actually, it was the anti-LLM side that proposed corralling that stuff into a specific category. 
I’m as opposed to generative “AI” as anyone here and avoid it as strictly as I can, with is why I suggested this as a solution.
Personally, I hate everything that contributes to global warming. It’s why I gave up driving and why I use a heat pump for heating. Wherever possible I buy used instead of new, which is why I am typing this on a 13 year old Thinkpad.
It’s not always possible to buy used, of course. Used sandpaper isn’t very effective, and not being a dung beetle I prefer not to eat used food. I suspect we have that in common, at least.
1 Like
Dissing is a overstatement, but there has been a large amount of beef between Ai and the dev especially with haiku server being attacked or the Claude incident which already didn’t give Ai user a good rep
As far from the Dev warning people over using Ai for some of the most minimal reason or them having a large amount of biases already toward AI.
As far as it trying to make a middle ground for there to be a category for the Ai user and developers of Ai to co exist is a good idea in theory but alway gonna have a bad execution as show right now as both the Developer of haiku and the Ai user of haiku have both contributed and help haiku as a whole
Your problem is that you want division… I don’t feel like I’m part of the AI supporters. It’s something that suits those who want to delete something from the forum.