Based on recent comments this feels like a discussion we should have. So…topic, basically.
I’m not looking to be chief noisemaker on this, but I stand by what I wrote in !privacy and what’s in my post history.
https://lemmy.ml/post/48724623/26190950
Let’s have at; do we want a [AI] and [NOT AI] tag. Why or why not?
A mandatory [AI] tag? Sure.
A [NOT AI] tag? No, that’s the default. Why normalise AI bullshit even further?
But mandating [NOT AI] means that people have to go out of their way to declare their work is AI-free. It requires active lying rather than lying by omission—I think there are a non-zero number of people who would be inclined to omit an AI tag but would not want to go as far as explicitly lying about their work being AI-free.
Agreed. “Failed to disclose” isn’t condemned as harshly as “Blatantly lied”, even if it should be. So obfuscating a project’s AI usage may be seen as less risky than being upfront about it.
A responsibly transparent project should advertise itself as AI-free if it truly is.
Then failure to disclose can be condemned, for instance by temporary bans
What does it mean for a project to deserve the [AI] tag? This matters, because you may have a lot of projects where a developer may think “no” and someone else thinks “yes”. Some examples from my day job:
- Developer used AI to understand part of the codebase and suggest ways to accomplish goal. Developer incorporated that suggestion, though using their own knowledge deviated from AI’s suggestion in parts. Developer wrote the code themselves. Is this project [AI] or [NOT AI]?
- Developer used AI to review existing (human-written) code for quality and security purposes. AI noticed some issues and proposed fixes. Developer reviewed and accepted them. Is this project [AI]?
- Developer knew they wanted to implement a feature, and while implementing it there was a boilerplate function. Developer asked AI to write this function, manually reviewed it, confirmed it worked, and added it to the codebase. Is this project [AI]?
In these examples the developer carefully reviews the AI’s output, which I think distinguishes it from vibe-coded slop, which at least is what I want to ignore.
It’s also worth noting that an open-source project may receive and incorporate a well-written contribution where the human developer used AI carefully like this. Unless they disclosed that they used AI, it may be unknowable to the project maintainers whether their project is [AI] or not, depending on how you define it. What tag should these projects use?
Sir, this is Lemmy. If you use AI in any way, you are clearly in league with the devil and deserve to burn.
I agree with all your points, BTW.
I posted this discussion because I wanted to explore both guard rails AND nuance around that sort of work flow, particularly for our new mod (and in light of several other scattered convos).
A lot of the diffuse FuckAI Lemmy crowd have poor understanding of code workflow. “AI bad” knee jerks so hard it’s going to dislocate something.
I’ve tried to argue this point, because roughly… ooh…100% of code gen touches AI something. So, do we tag everything?
What people really want is a [SLOP] tag, which is both lazy / not doing your own due diligence and impossible to implement.
In hindsight, I think the pragmatic approach is ultimately the workable (albeit blunted) one. Have the ai tag. It flattens everything but if stops brigading and slop, that’s the least amount of moderation work.
I appreciate you posting btw.
Sir, this is Lemmy. If you use AI in any way, you are clearly in league with the devil and deserve to burn.
Bahahahahahaha!
@festus @selfhosted Excellent examples. What the tag [AI] conveys is not what you really need to know, which is the quality of the code (component/unit), unit testing, and so forth. I assume there is some acceptance testing done at the project level. The human who submits the code must understand that flaws in their code is their responsibility, just as those who contribute/maintain the project are responsible at the system level. It is both an objective and reputational process. Does it really matter what tools are used if the work product passes the test, verification and validation criteria? Sloppy code is not unique to AI tools.
No tag for not AI.
Only AI tags needed, which helps remind people that slop should be warned against. We don’t need to warn for slop free apps
But wouldn’t that be far more useful? Many people seem to be looking for projects who don’t ever touch AI. Devs who use a [No AI] tag show that they follow the same agenda and most likely will not change their opinion on the next release.
No, because the absence of the tag indicates it’s free of AI
Or that the author didn’t know about the rule?
That’s why we have mods
Or…we could make it easier for them and make the use of the tag a conscious decision indicating something instead of relying on the mod’s voluntary work to correct for an implied meaning.
We could make it easier for mods by reporting articles to the mod.
Even better: We could mandate every post to put [Moderate This] at the end if they don’t follow the rules.
Either ban vibe coded projects entirely or ban vibe coded projects that have less than a year of history. If allowing “mature” vibe coded projects, require the tag.
Spaces like this become so much worse when “i made this last week look at the shiny ui 🎉🎉🎉🚀🚀🚀🚀🚀🚀🚀” projects that will never ever see any form of maintenance are allowed.
This is a rule that could actually be implemented and would help with the slop vs not-slop judgement call
I think this is a major over generalisation that misses the main point of why people don’t want AI projects. The real questions are:
- Is this slop?
- Does a human understand all of this code?
- Did a human design this deliberately or is it completely derivative and uninspired?
- Will a human take responsibility for bugs that come up?
- Did a human write the docs?
- Will this be maintained or just a weekend project with no substance?
- Does this actually serve a purpose?
Idk how to address these things really. I could see the AI tag going both ways, but I do think it’s painting with too broad a brush.
I’m in the process of launching a new business as a single dev, and I want to be upfront about my use of AI, but I notice as soon as I mention any LLM, most people assume what I’ve built is “vibe coded” without even looking at the applications themselves.
I spent almost two months just setting up the devops side of things before I even considered publishing. Feedback buttons in the apps automatically open issues on my Forgejo, push mirrors to GitHub and Codeberg, and I do weekly progress reports internally (I stopped posting to Lemmy after people felt spammed and now I just post on the site blog)
I’m just not sure how to make it easy to tell that I’m actually putting my heart and soul into building software that should help people.
If curious: https://circuitforge.tech/
I looked at your site. Realistically I think we’re talking about two different things here. There are tools which were programmed with the assistance of AI and then there are AI powered tools. Your stuff is clearly the latter.
I think ai powered tools are a lot harder to sell people on right now. The ai powered area is where I see by far the most shovelware/slop.
To be blunt, when I see a new ai powered tool I assume several things:
- This is low effort slop
- A tech bro vibe coded this 100%
- No one cares about this
- It will either make a ton of money or be abandoned next week
- If it’s popular, anthropic will buy/clone it and everyone will use their version instead
I know this isn’t true of every ai powered app but it’s true of 99% of them right now. I really have no idea how you could convince people yours isnt slop.
One idea might be to stop using the term “AI”. It’s a buzzword with strong connotations. Some people hear it and think “gold rush”, some people think “slop” or “data center”.
Personally I would be a lot more likely to take a project seriously if they used the term “LLM” or “machine learning” to describe what powers the product.
Also, I don’t see anything that looks like obvious AI art to me but DONT USE AI ART. AI art is already terrible on its own but when I see AI art mixed with AI text telling me about an AI powered app I’m 1000% done giving my attention.
For sure on the art. I’m a graphic artist by education and I’m already working with another full-time artist on the application logos. When I make enough I’m planning to hire a proper web developer for the frontend/web design as that’s onc area I do lean pretty heavy on the LLM’s for assistance.
Funny you should mention the AI vs LLM labeling as I much prefer to call them LLM’s or “models”, but I’m trying to keep it accessible for non-technical people. I think I’ll go back through and rename though.
Also most of the tools are deterministic, with the LLM’s filling in gaps where there needs to be some amount of probabilistic interaction, like figuring out what you could cook given only the ingredients in your pantry, or rewriting a resume 26 million ways to satisfy the ATS filters.
If you don’t want to use those features, nothing forcing you. They’re useful for tracking and organizing without the LLM at all.
Yeah… Normies for sure don’t use the term “LLM”. I don’t think it would scare them though. They’re used to being confused 💀
When I see “AI” I think “tech bro hype” whereas “LLM” makes me think the developer is more likely to have a realistic view of the technology.
I guess this depends a lot on your target demo. But but even just cutting down on usage of the exact term “AI” would help I think. Basically anything is better.
Much participated on the advice!
What I am curious about is why this should be a negative for anyone, devs who want to use AI get an easy way to filter out the people who will kick back against it, the people who will kick back against it get a quieter existence, Lemmy should be happy.
I keep seeing how having to categorise will provide a perverse incentive to not disclose and I guess I don’t understand why that would be the case.
It’s not like they are tricking people into buying these free programs, it’s not like they are soliciting contributions from other devs (they have an AI for that), and its not like there is some sort of score being kept (besides earning some sort of credibility on Github as a pro-AI developer through that star thing I guess).
So what would be the motivation to try to trick the community into embracing these sorts of projects? Open and enthusiastic disclosure and a community push to simply move on if you find that style of development distasteful would work better for everyone.
I have walked away from using a project that was developed with AI and I didn’t feel the need to slam the developer for it, I just moved on. They didn’t betray my trust because they don’t owe me anything, and I didn’t unfairly judge their work because I don’t owe them anything. Everyone’s a winner.
But that’s just my humble opinion.
Just ban AI already
This is a better plan.
I think the reality is that people will downvote the AI posts, and that will incentivise people to not disclose it or outright lie. The other thing that came to mind here is the fact that I’ve been trying to set my RSS reader up to not show me anything if AI is mentioned. It turns out that I haven’t been able to do that because it couldn’t discern “AI” from “fair”, “pair”, “air” etc. but the sentiment was there because I’m sick of hearing about it, and I imagine a lot of people are. This could cause readers or whatever else is configured to block AI content to block the non-ai content too, just because it’s mentioned. Additionally it does bring AI to the forefront, which doesn’t help with that AI fatigue.
Ironically…ai is probably an excellent tool to prefetch your content, perform sentiment analysis and then sanitise the content to your liking.
😂 yes indeed!
Me: I’m sick of hearing about AI. It’s very existence brings me down. It’s bad for the environment, bad for code, bad for mental illness, bad for humans.
AI: Sorry to interrupt, but have you considered using AI to help with this matter?
Legitimately, I agree it would be a good use, but I will be sticking with non-AI solutions regardless.
Hey, even Che Guvera wore a Rolex. :)
PS: I think you’re circling something though - people are objecting to the idea of what they think AI is, based on emotional appeal. It’s a category error.
As in - you know damn well that the correct tool for sentiment analysis is an AI but you’d rather avoid using it because … whatever.
It’s that “because whatever” I’m pointing at, because right now it’s unexamined and at best scores you a pyrrhic victory. Sentence transformers, rankers and re-rankers are AI, the right tool for the job and you won’t use them because… Ai bad.
(You in the general sense, not you you).
“Small,” non-generative, BERT-like models, would probably be more appropriate.
Anthropic does sentiment analysis with regex in Claude Code though ¯\(ツ)/¯
Because use of AI is bad for me. YMMV.
I’m midly curious as to how.
How would (say) you using Qwen 3.5 4B be bad for you (specifically) in this case?
Qwen’s open weights, already trained, runs locally on your rig, doesn’t leak PII to the cloud and does the job.
Surely if AI discussion in feeds is causing grief, anything that removes that for you (your stated intention) is “good” for you?
I find that whenever I have used AI to find a solution, I have forfeited learning, so I avoid it completely so that I can learn about the thing that AI would have otherwise shortcut for me. Regarding this specific example, while it may or may not be the case that AI is the best tool for sentiment analysis, it’s not the best or most efficient tool for keyword filtering, and keyword filtering is all I’d really need. In the event that I cared enough about the filtering to find or build a solution, it wouldn’t be with AI for this reason, but also because of the aforementioned reasons. Even if they’re not applicable to this very scenario, I’d rather not involve myself in the technology at all if I can help it.
Can we talk shop? I don’t want to come across as badgering you if you’re happy to put a pin in it, but I think this is a bad take. Like, if you’re going to try and solve this with Regex soup and spite, it’s going to hurt.
What happens if Codex, Claude, Cursor, LLM or phrases like “machine learning”, “generative ai” etc are mentioned?
Or when someone wants to have a discussion like this?
\bAI\b will miss almost all of it.
Note: I have no stake in you using or not using AI, an I am not trying to convert you to the church of Latter day Aiology. I am simply trying to [
]avoid doing real work[] chat.
Hmm, can I RegEx this?
[\s-]AI[-,\.\s]This is assuming it’s not at the start of the article.
EDIT: Thinking about it 2 more seconds, this might actually be more precise:
[\W_]AI[\W_]Doing more, like
\WAIwould filter words like “ailment”. Haven’t found a word matchingAI\Wyet, but I’m careful atm.It won’t work. You need a text classifier to do sentiment analysis, because “ai” is a concept, not just “ai”. TinyBERT or MiniLM I reckon could do it or if you really want to cut off your nose to spite your face, code the equivalent in python from scratch.
Say what you want about M$, but TinyBERT / MiniLM are awesome.
Smart play would be for the RSS reader to have that as optional plug in module, IMHO.
You know, now you mention it, I haven’t tested to see if the filter functionality of my reader will accept a regular expression. I’ll give it a go later, thanks!
Having the tags? Sure.
Making them mandatory? Only if we have 1.- an actual process to determine whether a tag is incorrectly applied (up to a respectable level of confidence) and 2.- an adequate, *enforceable+ punishment for infringers.
I, like other respondents, don’t care if AI is used, I only care if AI was trusted.
AI is a tool to enhance a workflow, and as long as a skilled human is reviewing it and fixing it, fine.
We would be better off defining a programmer’s project vs an ametuer hour vibe coded monstrosity, but that won’t ever really happen.
No, because it’s about the what, and with or without AI is the how.
We don’t have disclosures “built on a Linux/Windows/macOS machine” or “built using IntelliJ/Eclipse” so why is it important what tool was used to do something?
Some people have serious ethical and quality concerns about AI usage in code in a way that’s just irrelevant to the OS and IDE used to code it.
I understand that people have concerns, but those concerns are only relevant if they come from a maintainer or repo owner point of view. There is an entire spectrum of how AI is used in code, and it’s not a simple yes or no thing. I am for example completely against vibe coding as it’s just a risk and liability in the long run. However, to use AI to brainstorm, get suggestions, discuss architecture, learn with examples, and assist basically like someone else is sitting next to you while you code yourself… that is something completely different, and results in a completely different outcome.
In the end, it depends on can the person at the steering wheel take full ownership and accountability of the code they produced, with or without AI.
I understand that people have concerns, but those concerns are only relevant if they come from a maintainer or repo owner point of view.
This just isn’t true. Whether I’m a developer or not, I have to deal with the security issues that come with running the code. I have to deal with the bugs that come from it. I have to decide if I’m willing to support practices I may consider unethical used to produce the software, especially if I’m considering donating to the project. I don’t need to be a topic expert to understand that AI code is prone to bugs and security vulnerabilities, nor do I need to be one to consider the massive ecological damage and copyright violation required to train the plagiarism machine.
However, to use AI to brainstorm, get suggestions, discuss architecture, learn with examples, and assist basically like someone else is sitting next to you while you code yourself… that is something completely different, and results in a completely different outcome.
Generally, people take the most issue with using it to actually generate code in any capacity. There are purists who might insist you not even touch AI, but I think most draw the line at including code or graphics not written by a human.
In the end, it depends on can the person at the steering wheel take full ownership and accountability of the code they produced, with or without AI.
If we presume ethically neutral tools, sure, but the massive damage to ecosystems and towns that comes with training, using, and powering AI are seen by many as outweighing the utility they’re able to provide.
You don’t blame the knife when people get stabbed, so why blame AI for code it generated on the developer’s behalf? AI is a tool. Nothing more.
For the other stuff - I can’t take people seriously, because most people pick and choose when it’s convenient to be moral and ethical. Same people who don’t like AI pollution are using cars and plastics in their daily lives and would absolutely flip out if someone would ask them to give it up for the environment. AI is just the current “trend” to hate and be moral about because not many people are depending on it yet in ther day to day life.
I hate it when the default state is turned into the negative. Every time I have to specify “unsweet tea” I feel the sands of my lifeforce slipping away.
Absolutely.
Please do! It’s always my first question when reading new projects.
I would love to see an [AI] tag, so I can easily hide it. Coupled with temporary bans in case of missing disclosure it would really sanitize the community
