Welcome to the forum @Carlosjauregui918! Or out of lurk mode, at least.
Thank you for your contribution, it was beautifully put. I appreciate your re-centering this thread in our most important guiding principle with respect to this issue.
1 Like
I hope I’m not alone in saying that I don’t expect moderators to be mandated to do anything. And requiring a tag would be just that. I don’t think it fair to ask moderators to police this.
The best way to handle this is with culture. Have people proudly post in Showcase that they made this thing and it’s handmade. Have people participate in real discussions. Ignore the slop posts. If they don’t get attention, people will go elsewhere for the accolades they are hoping for. The best way to discourage the (incompatible) behavior we don’t want is to ignore it. If we have repeat offenses/someone malicously not complying, then moderation would be waranted. It’s the individuals that make up the community, and we have to make it the place we want it to be by controlling what we each individually do as part of the community.
I’m not saying don’t have a policy. I’m not saying we shouldn’t update the policy. The current policy, especially around treating other patrons with respect, goes a long way and is working well so far. However, I also recognize that when we flag a post because the policy is vague or unclear, it puts pressure on and takes time from the moderators.
As far as tagging goes, I think the best solution from what we’ve discussed is to encourage self-tagging particularly in showcase. Not enforcing it, or requiring it, but making it an expectation. We as the community can support this by asking (politely!) on any posts that don’t tag for the author to update it with the appropriate tags.
7 Likes
In this thread, has anybody called another user here fascist? The owners of the AI systems were mentioned several times, and I think there are some legitimate questions to discuss there, but your response calls this out as if it is directed at fellow users and I believe you are mistaken.
6 Likes
It has been, we’ve had to moderate exactly that in the past. We’ll do it again if we have to.
I must have missed that. Even in this thread I wouldn’t think that’s okay, not unless a really specific charge is being made.
That said, there are a number of fascists in Silicon Valley these days and calling them out in this thread seems perfectly appropriate. I must have conflated the two.
1 Like
Sure, get it out of your system. Normally that’s completely off topic and we don’t want it here.
1 Like
There is an argument to be made that LLM are built by tech billionaire that destroy the world though. In the zig zen, it is stated that together we defend the users. If we truly want to defend all of the users, we need to defend minors on the other side of the globe working in mines, or data worker being exploited to annotate data. They are certainly using (our) software too. Or we must amend the zig zen and defend only some of the users. Or remove this from the zig zen ?
Fascism is relevant to the discussion of whether we want to pretend the output from a billionaire-owned machine is equivalent to a human creation when some of those billionaires are openly promoting fascist agendas.
4 Likes
I also understood it this way and not meant against users. Otherwise, I wouldn’t have agrred upon the post.
However, if most Big Tech CEO’s are fascists is up for discussion. At least, imho they’re all capitalists in the baddest sense. But to discuss this in detail is off topic here.
What is not is the impact on Zig(git). I appreciate Zig, its maintainers and also the community for a clear stance against big corporates habits (Codeberg-move etc). That includes LLMs becausr its brought to us by those companies. And also the claim that LLMs are unavoidable from now on seems to be created and injected into public discussion by those Big Tech jerks. And I woild like to discuss this further in a fully off-topic thread.
Beside that, for code showcases etc we have to look on it realistic regarding a volunteer maintained community like Ziggit. Thus, I would support a “handmade” tag which is not obligatory but highliy encouraged to use.
Being about twice Andrew’s age, I probably have half as many hours remaining on this marble, and that makes me even more suspicious of wasting those hours on things I do not conciously decide I want to spent my time on.
I stopped watching television in the mid-90s. Even though there are many funny/interesting/moving series, movies, and documentaries, I don’t want my usuable lifespan reduced by a few minutes every hour by having the interesting stuff interupted by, life wasting, advertisements. ( And the percentage ad time has probably increased since then).
If I would start posting Dutch translations (using Google translate) of posts here on ziggit, I would probably be told that everyone who wants or needs to have those translations, can do that for themselves. If I want to know how an LLM would solve a Zig related problem, I will just fill in the prompts for that myself. No need to read about that on Ziggit
I prefer my Ziggit ( i.e. the ziggit I want to spent time on, I am not tyring to imply I own Ziggit) without LLMs altogether, and if that is not possible there should be a clear marker identifying any thread that allows LLM posts (even if those have yet to be made on the thread), and some system to mark those as read/skip.
I don’t have much faith in such compromises, as compromises make everyone involved unhappy.
As an aside; Something that would be nice with or without LLM generated posts, is being able to make it possible to “follow” people on Ziggit and then be notified of those threads where the followed person has contributed.
8 Likes
It looks like this instance of Discourse (the forum software that powers Ziggit) does not have this feature. (I think it might be a little disconcerting, personally, to have my forum participation generate notifications for others.)
That being said, you can approximate some of that functionality by bookmarking the Ziggit profile page of anybody you would like to follow in this way and perusing their new activity at your leisure.
1 Like
I think this is a good idea. I don’t really care about agents or language models as a topic, so I could filter them out. However, I would be interested in updates about Ghostty, so just because a project uses AI doesn’t mean that a post about it belongs in the AI topic.
One question: would this be a tag or category or topic? I assume you were using topic in a general sense and not in the discourse sense. Making it a tag makes the most sense to me.
I think that adding tags describing how language models were used would be good. They should be used, but if someone forgets they can just be reminded. I don’t see how adding such tags could be counterproductive.
In my opinion a binary system will lead to unnecessary conflict around categorization. Even the handmade tag would be ambiguous. I think tags like “zero language model usage”, “language model was consulted”, “contains language model output” would be better. If you think of any more, please post them.
For people who want absolutely nothing to do with projects using language models in any way, they can look at the “zero language model usage” topic.
If everything in the project was done by humans, but language models were asked questions and consulted when working on the project then the “language model was consulted” tag should be used.
If the project contains any output from a model whether it be the readme, documentation, a test, or other code, then the “contains language model output” tag should be used.
These labels are neutral and informative. People who use language models won’t and shouldn’t feel shamed (just my opinion) when they add these tags to their post. I think most people will opt in to using them and won’t try to pass off language model code as their own.
I think this is very important. Nobody in the community should waste their own time insulting others. When you say these things you harm yourself too.
3 Likes
We have had a policy against “AI participation”, posting chatbot prose in comments, for a long time. That won’t change.
1 Like
I meant category actually. I don’t think discussions in the intersection of agents and Zig programming should be in Brainstorming or Explain, primarily because they don’t really fit. Some people don’t want to see them at all, and I don’t want to see them in there.
Categories can be muted as a whole from user preferences, so as a primary firewall between those who don’t care to be reminded and those who wish to discuss, they work pretty well.
As a general comment about tags: all anyone has to do is type in the tag they want, and the post will have that tag. So all of that can start now, on a voluntary basis, I think that’s a good choice.
Moderators are also able to add tags as part of our janitorial toolkit. I see that as fairly low stakes, like correcting a spelling error in a topic header, or moving a topic to the right category. Tagging posts isn’t something I’ve done often, but I have done it.
1 Like
I think it’s been said somewhere before, but I’m perfectly happy if people post in the language that they are comfortable in.
I use language models in a similar way and agree 100%. I guess I am on the free tier, but I can’t imagine the difference is that big. With traditional search engines I started getting poorer quality results and now my search term skills are pretty badly atrophied. I feel kind of stuck. I should post on forums more, but it’s uncomfortable for me.
It sounds like you also want to avoid spending time on projects whose people “only” consulted language models. Is that correct or am I misinterpreting?
I hope you will stay and that a system which you find adequate for avoiding brain rot will be put in place.
I would like to know your opinion on the tags that I brought up in my previous post here.
Mine for example haha.
This sounds reasonable to me, but tags can help to filter out posts easily.
I don’t want Andrew or anyone else to leave the community whether or not the problem is real.
I had to add directions to not do this, but it doesn’t work generally. It just changes the phrasing slightly: “That’s a sharp observation!” I should really use online fora like this one more.
This person gets it.
Thank you, mods.
I’m not sure how much harder it is to add 3 tags instead of one, but I think having disclosure be through tags would make this easier to automatically filter.
I would be absolutely mortified if my crappy beginner code was labelled slop.
Me too.
I partially agree, but I think that the tags should be neutral. Otherwise people will be incentivized to lie for clout.
You have my condolences. I hope this isn’t agitating, but what is your opinion of Ecosia AI?
I hope you will stay. Your voice in this community is important to me at least.
Even if some people make drive-by posts that don’t meet these criteria, I hope that people will stay for the subset of the community that shares their values.
Very well put.
I think this is a good policy.
Oh, so a tag could be any arbitrary string? That’s neat.
I hope including all the responses in one post like this isn’t a bad thing.
1 Like
Tags are not in limited supply! They come into existence the moment one types them in, and the filter displays popular tags which match.
I hope we’re all seeing the same thing, which is that using tags is broadly considered courteous, no one seems to mind and many would appreciate it. How us mods might end up providing some official guidance on that is not something we’ve finished discussing.
3 Likes
Just a short followup… @mnemnion I love your idea of the “handmade” label, which I think solves most of the issues identified here. The false positive rate will be low. The inverse (an “ai” label) doesn’t work. They keep trying it on reddit, but newbies or low-effort posters don’t even know they should be using it. As llm usage becomes more prevalent it just fails harder.
I do think we should bias toward making ziggit nice for @andrewrk though. This is an important community for zig, right? These spaces need to be protected.
5 Likes
why can’t we have the both then ? people who are genuine and don’t feel uncomfortable disclosing AI usage because they can stand behind their code, but still want to respect that others might not want to interact with it at all, can put the AI usage tag.
People who have built something entirely without the involvement of AI can also showcase their project, and mention that to signal others, that it’s a project they can feel comfortable commenting or looking at.
And on top of that it’s a nice, polite conversation starter when someone new post something and it has neither of those tags on ?
3 Likes
I’m not sure if this is right or not, but I can certainly imagine that most usage would be in the “Showcase” category. So would it make sense to split that into “Showcase (handmade)” and “Showcase (llm assisted)”? That might make it more intuitive for new users.
2 Likes
I was talking about the owners class, and I didn’t mean to imply I was calling users, people here, fascists. I was not and am not.
I believe you should care about my political convictions (in this instance) because they are the core basis to my reasoning on LLMs, which is what this thread is about. If you disagree, then so be it, but to me it’s an incredibly important, the most important, detail.
I have resorted to name-calling in the past on a couple LLM-adjacent posts, and I apologize. I took a step away for a while and came back. It was a rude and unproductive way to engage, which is why I tried to take a positive approach this time, focusing on what I believe and how I feel, rather than telling people what to do and name-calling.
10 Likes