As a forum moderator, I try to read everything here. This indeed takes a lot of time and I would rather spend it doing something else. Currently with so much debate about AI it also takes a lot of energy.
This also raises another question: who is going to read and moderate that new LLM category? I expect there would be a LOT of things posted there, since slop coding is so fast, you can have a new unmaintainable app up and running every few days. I don’t want to suffer through reading all of that. I think several of the other active moderators feel the same about it. So, the side effect may be an unmoderated category. On the topic that is the most heated currently. That doesn’t seem wise.
I also agree with Nipos that ti would establish that LLM generated/assisted code is welcome here. For me, that is very much not the case. So, I would rather vote against the category, and for a stricter ban on LLM generated things. I will also note that this is not an attack on freedom: write any app you want, do whatever you want with Haiku. Do not ask this forum to provide a platform for your advertisement. If people are so strongly unhappy about that, they can start another forum in some other place. With so much people voting for this category just so they can hide it, I would say this seems appropriate.
There is ongoing discussion in the moderator team to decide what should and shouldn’t be allowed. As x512 says, LLM-assisted code is present in a lot of software projects now, and we have to take that into account, and be consistent about whatever decision is taken (ranging from the extremes of “any LLM slop is welcome” to “no LLM use at all, let’s keep using old Mesa versions from before the LLM apocalypse”).
How convenient to handle the arguments against LLMs and discussion about the forum moderation process as “hallucinating”. After not having even bothered to read the full thread?
Please, please can you clearly state which parts of your posts are from a chat bot and which are your own words? How is anybody supposed to attribute sentiments to you or the chat bot otherwise? I mean I’d rather only read your own words if possible…
The community extends much further outside the reach of one single forum (fortunately).
I have given my personal opinion, which has not any more value than that, juste one voice between many.
If the forum overall decides to be welcoming to LLMs generated apps, I handle take the consequences of that and not expose myself to the forum. I am starting to think it would be a good thing for everyone involved.
Anyone reading this should assume you’re responding to another topic. No one here wants to make decisions for developers, and the topic isn’t whether or not to use AI.
The proposal “AI-Assisted” category is intended to defuse pointless discussions you’re constantly pushing forward, so it’s not you who’s sick, but us.
To reframe a subset of this “dumpster fire”, I’ve started another thread called Wrap it up! Native code written in…Dart!. It proposes a particular category of application development that actually can benefit from AI or other pattern-match algorithms. Wrapper languages from mainstream origin.
If contributions posted on the forum have to respect contributions rules to the OS itself then perhaps the forum should be about the OS only. That would be sad but it’s probably the price to pay. It won’t really solve the problem that will be moved to another place and the risk is to see another Bebits story. I mean things will happen there until they’re lost.
This is also true, to a degree. “welcome” may be overstating it, since it’d be restricted to a single category of the forum, but still welcome to some degree.
I actually agree with the full ban side. The options, in order of my preference, are:
Haiku Inc. issues a ban on generative AI-produced material on official Haiku resources; this would effectively ban it on the forum
the forum issues a full ban
the forum restricts it to a single category
the forum continues as it is
I don’t think the last is tenable, and as I’m not an insider I don’t know how likely the earlier ones on the list are.
I’m just pushing for a “we should at least do this” option.
I want to re-iterate what I said on the discussion about the same issue in haikuports, for the pro-LLM folks: should the decision on whether to allow a space for LLM-generated material here be to ban it outright, I will help y’all get some resources set up to establish a pro-LLM Haiku community.
(I won’t host it or pay for anything, on account of I actually disagree with you, but I’ll help you get set up at GitHub/Codeburg/whatever. It’s worth some effort to keep the idea that there’s a Haiku user community, even it some of our spaces can’t be shared over this issue.)
The forum doesn’t have to be about the OS only,but about the OS plus those third-party apps that follow the same rules as the OS itself.
I don’t think the rules need to be that strict,but it would most likely be better than the situation we have now.
I was already thinking about starting a non-official Haiku forum with a strict ban on all LLM content.
I don’t think this is a great solution as the biggest part of the community will most likely not move over,so the new forum would be a rather silent place.
I’ll keep watching the situation,but if the slop becomes so prominent that it kills off the rest of my motivation to contribute here and no countermeasures are taken,I’ll probably work on that.
My dedicated servers are mostly idle and domains don’t cost a lot,so why not?
That is kinda the flip-side of this decision: either we find a way to share this space (almost certainly by isolating into mute-able categories) or one side or the other is going to decide that this forum is too unpleasant to continue using it.
Guys, seriously. What is going on here? I’ve never seen such hostility and cynicism in this community before. Not even when package management was introduced.
The Haiku community has managed to develop a working operating system, a collection of native apps, a porting infrastructure, a relatively friendly community (until now). Is the division over AI going to break all this? I hope not.
One thing is for sure. AI is going to stay, in one form or another, whether we like it or not. That means we have to deal with it. Why not do this in a clear, constructive and transparent way?
As far as I see it we have to make decisions on the following things regarding to AI and agree on who makes the decicions:
Haiku (the OS itself and the bundled applications)
Haikuports
The forum and the mailing lists
3rd party applications ( different from the other points because everyone can make an app any way they want. But we have to decide whether AI created 3rd party apps can be advertised on the forum or distributed via Haikuports)
I assume the group of people that decides for Haiku itself will be the people with commit access and at Haikuports it will be the members. For the forums it seems less clear.
Come on, let’s get together and work to establish clear AI policies so we don’t have to discuss them in every thread in the future. If there is anything I can do to help let me know.
I think it’d be best to wait for the moderators to finish the discussion @PulkoMandy mentioned at this point; they know the options and should have a decent idea of user sentiment by now.
That sentence about the package management was just there as an example because it was one of the bigger controversies since I follow Haiku. Nothing more, nothing less. In no way did I compare it to AI.
Did you read the rest of my post? We’re not getting anywhere with rage posts like yours. I actually agree with a lot of things you say, what I definitely don’t agree with is the way you deal with other people’s opinions. You have to accept that other people have different opinions, even if you can’t understand them. You are not alone in this project. We have to find ways to get along with each other in situations where we don’t agree on something.
