lobsters score: 162 · tags: meta, vibecoding
I vibecoded a
vibecodingtag ->llmstag greasmonkey script a while back which made my life better.https://lobste.rs/s/68n22g/lobsters_rename_vibecoding_llms
I’m glad to see the tag be closed https://lobste.rs/s/h0eate/updates_site_guidelines_ai_llm_tags
score 1 · canonical
I wish we had “vibecoding” and “vibecoded” as separate tags. One for posts and discussio about genai, one for things made with it.
I for one think a lot of vibecoded software is kinda neat. It’s good to know what it is so I don’t give it the keys to my house, but i still want to see it even if I don’t want to read another article about vibecoding rn.
score 1 · canonical
deleted by creator
my experience on lobsters has been way better since i filtered the tag. it happens to fit my taste. i don’t think everyone loses.
score 2 · canonical
Applying the vibecoding tag here is only going to create a “bubble”, and seems like a terrible idea for building a cohesive community.
score 1 · canonical
The tag today has completely devolved into a generic “i don’t like AI” trigger warning tag. It’s placed on stories where LLMs are used in any form whatsoever. I have accepted it at this point, that this is what the community is using it for, but I have actual doubts that the majority of the community wants it that way.
For a start, if it was just a way to not enter conversations at all and filter it out, it might even work. Except almost every single post that involves that tag, also has a meta discussion in the comments about AI, instead of actually talking about the topic at hand.
score 3 · canonical
I’ve been lurking for almost a decade, after leaving Reddit due to inane cultural dog-piling. Online communities (and especially aggregator communities) always degrade to warring over binary identifiers and loyalty tests; Maybe the medium is just not expressive enough to preserve real human connection. Maybe the medium is too good at convincing us our personal opinions are the popular and righteous ones…
In the time since I joined this site I’ve started my own business and integrated into my local community, made lots of friends, and spend maybe 5 minutes a day surfing aggregators. Probably because I just have the bookmark in my browser toolbar and click it out of habit.
I think the era of online communities is dead, reading through the comments on this post just affirm that. And it wasn’t Gen AI that did it, it’s these “communities” inability to build rewarding friendships that extend past ‘X thing is Bad, and You are Bad for liking it’
I’m gonna delete my bookmark for this site, and deactivate my account. I hope I inspire others to as well. It’s never been a better time to make friends in meat space that truly care about you, and understand you
score 1 · canonical
I’m sorry you feel this way but I disagree with your assertion here. While many people here have the view, sometimes very strongly, that gen AI is bad - for society, for the environment, for individuals, for code quality, whatever - I’m really not seeing a lot of examples of “you are bad for liking it.” It’s a difficult line to walk but I think most discussion of an otherwise very fraught topic stays respectful.
Online communities aren’t dead, and I don’t want them to die. This is one of the remaining good ones. It’s not perfect but I think the suggestions in this thread are entirely sensible and will help.
score 1 · canonical
I think my general issue with the tag is that it acts as a magnet for terrible commentary, and brings the level of discourse down across the board. People keep rehashing the same arguments, both for and against AI, and ultimately contribute nothing to the topic at hand. There is a non-trivial contingent of users who have decided that anything even tangentially related to AI is unworthy of discussion here - it’s not about the content quality for them, it’s purely about waging a war against AI/AI usage, and punishing those they disagree with.
I’m completely on board with throwing out low-effort, soulless, look-at-what-I-just-generated garbage - but there is high-value, interesting conversation to be had about things that are related to AI, or involve some degree of AI assistance, and it is nonsense to label them as the same thing. If the majority of the community here is more interested in disregarding anything even tangentially related to AI, regardless of quality, then it is what it is - we all have the choice of going elsewhere - but I do believe the community loses out with that attitude.
score 6 · canonical
Do you think a more neutral or positively worded tag would act less like “a magnet for terrible commentary”?
score 1 · canonical
Perhaps - I think more precise labels would be helpful in some respects. The vibecoding tag has its place for content that really meets the definition, i.e. stuff that is largely or completely AI-generated with little human review or alteration. Work that is AI-assisted isn’t the same thing to me, but likely still deserves an “ai-assisted” tag for those that don’t want to see content that involves any AI at all. I also think AI as a topic deserves its own tag, to distinguish meta discussion about AI from AI-derived work.
Ultimately though, it depends on the attitude and culture of the community at large - if those tags attract the same type of low-quality commentary, then it didn’t matter that we made the tags more precise, the problem persists. That maybe means it’s more a moderation issue - but I don’t want to burden the mods overly much either. The community itself has to self-police a bit to encourage good contributions, and discourage poor behavior
score 1 · canonical
From the perspective of someone who filters both
aiandvibecoding, I’m happy with the filtering I am getting. Thevibecodingtag is being applied in a way I don’t object to.The problem is that AI people hate the
vibecodingtag being applied to articles they like because the word “vibecoding” is derogatory. Filo left lobste.rs because of this.My suggestion is to rename the tags to be emotionally neutral. My suggestion:
ai⇒ai-theoryvibecoding⇒ai-usage
score 1 · canonical
The problem is that AI people hate the vibecoding tag being applied to articles they like because the word “vibecoding” is derogatory. Filo left lobste.rs because of this.
I think part of the problem is that labeling anyone who uses LLMs in any capacity “AI people” is the kind of divisive language that contributes to this site-wide toxicity.
score 2 · canonical
I think part of the problem is that labeling anyone who uses LLMs in any capacity “AI people”
You don’t even need to use LLMs in any capacity, it is enough to not hate LLMs.
Anyone who does not hate LLMs enough (to support the slapping of an obviously derogatory brand on submissions left and right) is “AI people”.
score 2 · canonical
I could get on board with that with the caveat that “ai-usage” be restricted to “GenAI as applied to software engineering”.
score 2 · canonical
What I find problematic is that it’s widely used as a big implicit downvote. Most of my recent submissions about browser/web security got the tag applied by someone else. “lots of people won’t see you anymore, because you touched the devil’s machine” is kinda weird.
score 7 · canonical
Some people want to continue developing like they have for the past decade or two, without constantly seeing LLMs everywhere.
So they filter the vibecoding tag as a way to keep a pre-AI style view of lobsters, in a certain sense a “what would the lobsters front page look like if LLMs had never been invented?”
Why do you see content warnings (which is what this tag is primarily used for) as an implicit downvote?
If users who didn’t want to see your post were forced to see it, that’d surely create even more debate in the comments below these posts.
score 2 · canonical
I’d vote, if there is a vote, for splitting it. I don’t filter but I do consider tags (and hidden count) when deciding whether or not to read the article. I took the “vibecoding” tag to mean that the post was generated slop, boosterism, or argumentation. That is not how it’s being used and I’ve stopped paying attention to it. Will additional tags fix that? I don’t know.
I’ll ask (myself) a set of questions: Are LLM’s tools? Yes. Is it possible to use LLM’s productively? Yes, I believe one can. Are the current LLM’s ethically trained and offered by trustworthy people? No to both. Is it ethical to use an unethically trained LLM? I don’t believe so. Is arguing these Q&A’s in every thread productive? Absolutely not. Is rehashing these points the most interesting part of the conversation about an article? No, not for me, thanks.
If I were creating a tag hierarchy, I’d separate it out entirely. These aren’t catchy but it should make the point:
I. AI A. Research B. Meta B1. General (likely flamefest...) B2. Politics & Law C. Application C1. AI-Using C2. VibecodingThen one could choose their level of filtering. If they don’t want anything related in some way to AI then filter the top level. If they want to skip the meta-discussions, filter ‘B’.
score 1 · canonical
I like the tag and kind of how it is used currently. What I think that causes a bit of a problem is its name “vibecoding”, because of the negative connotations for many, so it doesn’t really work great for LLM, LLM adjacent or LLM related.
I would love a better tag, used how it is used currently, and perhaps filtering the tag considering other tags on the post. For example, I wouldn’t want to filter that tag if it is together with “meta”. Not sure how feasible that is though.
At the end, it seems that this keeps coming back as a meta post. Is this one adding anything new from last time? Not really.
score 3 · canonical
My tastes say that I should filter “vibecoding” out. But since it means so many things to so many people, decent posts end up being tagged with it. And so, I don’t filter this tag out, and here I am–engaged in yet another community discussion about this tag that means nothing to everyone as a whole, but something to every individual.
In my opinion, there is no salvaging this tag. The tag is itself a pejorative for something that reasonable people have different appetites.
Archive it, split it, kill it–whatever that means–but there’s clearly no consensus on what the tag means, or should mean, and continues causing unnecessary division in one of the internet’s best communities. This isn’t a complicated problem; nobody is invested in this tag. Its absence will be forgotten a lot faster than the acrimonious discussions that pop up around it.
score 7 · canonical
The current state is a problem, and you can’t make an improvement without a suggested change, but in past discussions no suggested change has resulted in action.
Maybe we’d find it easier to agree on a desired state, or some details thereof, than a particular implementation? For example, I think:
- Tags are for filtering out content.
- Bans are for filtering out people.
- Tags are not for filtering out people. If submitters feel pushed out of the community due to other users tagging their posts, and a particular tag is an outlier in this effect, the tag isn’t functioning properly.
score 2 · canonical
I agree. I think the tag is basically useless for people who want to use the site primarily for what the site says it’s about:
Lobsters is more of a garden party than a debate club. We’re learning things we didn’t know to be curious about and sharing what we’ve made.
I read posts here because I’m curious to learn more about computing, especially compilers and programming languages.
Daniel Lemire is someone I’ve learned a ton from, especially his work on SIMDjson and related performance optimization projects. He has started using AI in his optimization work, but has not stopped writing thoughtful pieces about computing. Another example is Mitchell Hashimoto and his posts about Ghostty internals. If I filter the “vibecoding” tag, then I will miss out on learning anything about computing that people like them post, so filtering that tag would defeat the stated purpose of Lobsters for me: curiosity about computing.
Conversely, I’m not curious about “AI takes” - whether they are boosting, lamenting, predicting future impact, or anything else. There are a million websites where you can get AI takes, and I come here for computing. I understand that people want to post and upvote AI takes here, but I would greatly appreciate a tag that let me filter those out without losing high-quality computing content from people like Daniel Lemire and Mitchell Hashimoto.
score 9 · canonical