I find it very frustrating that this is only tagged "vibecodding", and because it has already be "user corrected" (or in this case, maybe user sabotaged), it can't be changed again.
on top of that, is the purpose of the 'vibecoding' tag to indicate that something was vibecoded? or that something is about vibecoding?
I might be interpreting it wrong, but on the tags page it is listed under the "tools" category: https://lobste.rs/tags. Other examples of tools are "vim" or "databases".
I don't suppose anything coded using vim should automatically get the vim tag, nor should everything that needs a database always get a database tag. My understanding is that the "vibecoding" tag is meant for blog posts about vibecoding, or perhaps something like a coding agent.
If the tag is supposed to mean "AI was used here" for the people who want to filter away AI generated content the tag is appropriate.
If the tag is supposed to mean "tools/advice for better vibe coding" for people who are interested in getting better at it, the tag is not appropriate.
My reading is that this is the second tag and that it is wrong here.
As mentioned, we don't have a tag for "vim" for things which are written with vim but people want the first use-case so a tag seems not unreasonable. Especially given that using AI to prototype things is easy but it's not clear yet - at least to me - what it will mean for maintenance. In other words, if it only took you an afternoon to build this, maybe I'm not interested in it yet.
exactly. if there was such a flag, it should probably be a flag to clearly indicate low human effort. Like "slop". But it could be argued that such a flag would not even be useful, because it's not like slop should even make it to the front page, regardless of what tags it has.
I mean the entire upvote mechanism (and hiding/flagging) is designed to prioritize content with real value. Tagging such things would not improve the pre-existing mechanism.
If the tag is supposed to mean "AI was used here" for the people who want to filter away AI generated content the tag is appropriate.
I agree in sentiment, but is it appropriate here?
But from a sibling comment, the project does not look vibecoded and only 1.4% of the commits show a symptom of being AI-generated. Saying that this tool is AI-generated does not seem reasonable. And if we are going to filter projects/articles where AI might have been involved in any amount, well, I'm not sure what will be left without the tag at this point... Does asking an agent "please review this PR and give me feedback / propose fixes if any" make a project instantly worthy of the vibecoded tag?
Oh, 100$ and I have no good solution. That's why I gave the example of AI use somehow implying "low effort" because you can now do things which would have required a lot more effort before with AI.
I could see making an exception for "oh we tried it once a few months ago and then stopped", but looking at the commit history it appears to have been ongoing continuously since october.
Same. I didn't know about prek at all and assumed "vibecoded slop" based on the tag. But I checked the link anyway because I'm forced to use pre-commit at work and don't like it. So then, as I looked at the repo... I didn't get the sense of vibecoding, and as I kept digging, I really didn't see any indications that this was slop.
So at first I got surprised by the tag.
And then I got pretty frustrated by the vibecoding label. It has nothing to do with vibecoding, and it's not slop. It's a genuine project.
If anyone thought that this deserved vibecoding because AI might have contributed some commits (as another comment in this thread shows, like 19 commits out of all), I must +1 the disagreement as others have pointed out for the many reasons that have already been raised. We don't tag stories developed with vim as "vim". We shouldn't have slop in the frontpage at all, no matter if it's hand-created or AI-assisted.
It has nothing to do with vibecoding, and it's not slop. It's a genuine project.
At this point, it feels like vibecoding is being used as trolling because I would not have seen this link at all with that tag filtered out. There are several posts I was interested in this week that were later re-tagged and I would have never been exposed to them.
perhaps it would help if the very first element in the readme weren't an obviously, almost offensively, AI-generated image.
I know it's a non-sequitur, but people have a limited attention budget, which is currently being abused, and AI slop images are a very good indicator that a page is going to be a waste of time. If you want a human audience, the least you can do is post human content.
I don’t see an issue issue here. Creator spend some time with AI to create a nice looking image. I think it’s totally OK even shows what he chose to do that.
I've migrated a bunch of repositories from pre-commit to prek, it was easy and painless, just running the hooks is instant. For those that don't want to be hampered by commit hooks, you can just run all the hooks in CI anyways. You can see what the creators of uv, ruff, etc are doing on their own repo. With hook stages, you can also keep fast hooks for pre-commit and slower ones for manual or pre-push stages.
I'm not involved with this project but I tried it and had a really good experience. I was surprised to see that it was not posted here yet.
I really like pre-commit but unfortunately it never reached its true potential, mostly because it never reached the popularity it deserved.
I don't really think it needed a Rust rewrite for any technical reason... It's not like you couldn't make it just as performant in Python. But perhaps adding a native command is more palatable to non-Python teams. Or at least, more popular for whatever reason.
prek seems to be a painless drop-in replacement, and it already seems like it's gaining more widespread adoption even outside of the Python ecosystem. So I'm really excited to see where this greater popularity will take the concept.
I like what you’re doing here but I wonder if there’s an open issue on the lobsters repo and if it would make sense to do it automatically. In the end, the submission form is already listing them when you want to hit send.
krtab | a day ago
I find it very frustrating that this is only tagged "vibecodding", and because it has already be "user corrected" (or in this case, maybe user sabotaged), it can't be changed again.
[OP] kantord | a day ago
I tagged it as rust originally, not sure who retagged it as vibecoding and why
edit: and btw as far as I can tell, it has pretty much nothing to do with vibecoding
hjvt | 22 hours ago
dbremner | 23 hours ago
19 of 1,335 commits (about 1.4%) were attributed to Claude or Copilot. Apparently that's enough to qualify as "vibecoded" to some people here.
[OP] kantord | 23 hours ago
on top of that, is the purpose of the 'vibecoding' tag to indicate that something was vibecoded? or that something is about vibecoding?
I might be interpreting it wrong, but on the tags page it is listed under the "tools" category: https://lobste.rs/tags. Other examples of tools are "vim" or "databases".
I don't suppose anything coded using vim should automatically get the vim tag, nor should everything that needs a database always get a database tag. My understanding is that the "vibecoding" tag is meant for blog posts about vibecoding, or perhaps something like a coding agent.
Edit: fixed a typo
apromixately | 12 hours ago
If the tag is supposed to mean "AI was used here" for the people who want to filter away AI generated content the tag is appropriate.
If the tag is supposed to mean "tools/advice for better vibe coding" for people who are interested in getting better at it, the tag is not appropriate.
My reading is that this is the second tag and that it is wrong here.
As mentioned, we don't have a tag for "vim" for things which are written with vim but people want the first use-case so a tag seems not unreasonable. Especially given that using AI to prototype things is easy but it's not clear yet - at least to me - what it will mean for maintenance. In other words, if it only took you an afternoon to build this, maybe I'm not interested in it yet.
bkhl | 6 hours ago
If you want to contribute to that discussion, that's probably better continued in https://lobste.rs/s/rkjpob/proposal_add_ai_generated_as_flag_reason
[OP] kantord | 12 hours ago
exactly. if there was such a flag, it should probably be a flag to clearly indicate low human effort. Like "slop". But it could be argued that such a flag would not even be useful, because it's not like slop should even make it to the front page, regardless of what tags it has.
I mean the entire upvote mechanism (and hiding/flagging) is designed to prioritize content with real value. Tagging such things would not improve the pre-existing mechanism.
jmmv | 6 hours ago
I agree in sentiment, but is it appropriate here?
But from a sibling comment, the project does not look vibecoded and only 1.4% of the commits show a symptom of being AI-generated. Saying that this tool is AI-generated does not seem reasonable. And if we are going to filter projects/articles where AI might have been involved in any amount, well, I'm not sure what will be left without the tag at this point... Does asking an agent "please review this PR and give me feedback / propose fixes if any" make a project instantly worthy of the vibecoded tag?
apromixately | 5 hours ago
Oh, 100$ and I have no good solution. That's why I gave the example of AI use somehow implying "low effort" because you can now do things which would have required a lot more effort before with AI.
natkr | 3 hours ago
It's very easy to not use the plagiarism machine.
I could see making an exception for "oh we tried it once a few months ago and then stopped", but looking at the commit history it appears to have been ongoing continuously since october.
jmmv | 6 hours ago
Same. I didn't know about
prekat all and assumed "vibecoded slop" based on the tag. But I checked the link anyway because I'm forced to use pre-commit at work and don't like it. So then, as I looked at the repo... I didn't get the sense of vibecoding, and as I kept digging, I really didn't see any indications that this was slop.So at first I got surprised by the tag.
And then I got pretty frustrated by the vibecoding label. It has nothing to do with vibecoding, and it's not slop. It's a genuine project.
If anyone thought that this deserved vibecoding because AI might have contributed some commits (as another comment in this thread shows, like 19 commits out of all), I must +1 the disagreement as others have pointed out for the many reasons that have already been raised. We don't tag stories developed with vim as "vim". We shouldn't have slop in the frontpage at all, no matter if it's hand-created or AI-assisted.
Halkcyon | 5 hours ago
At this point, it feels like
vibecodingis being used as trolling because I would not have seen this link at all with that tag filtered out. There are several posts I was interested in this week that were later re-tagged and I would have never been exposed to them.[OP] kantord | 5 hours ago
I have the same suspicion
natkr | 3 hours ago
Vim doesn't make your decisions for you.
Agreed, but keeping the slop in its own filterable corner still seems less objectionable than letting it take over the whole thing.
albino | 21 hours ago
perhaps it would help if the very first element in the readme weren't an obviously, almost offensively, AI-generated image.
I know it's a non-sequitur, but people have a limited attention budget, which is currently being abused, and AI slop images are a very good indicator that a page is going to be a waste of time. If you want a human audience, the least you can do is post human content.
antonmedv | 4 hours ago
I don’t see an issue issue here. Creator spend some time with AI to create a nice looking image. I think it’s totally OK even shows what he chose to do that.
gerikson | a day ago
Mods can change any tags, DM one of them and explain the issue.
natfu | a day ago
I've migrated a bunch of repositories from
pre-committoprek, it was easy and painless, just running the hooks is instant. For those that don't want to be hampered by commit hooks, you can just run all the hooks in CI anyways. You can see what the creators of uv, ruff, etc are doing on their own repo. With hook stages, you can also keep fast hooks for pre-commit and slower ones for manual or pre-push stages.Tenzer | a day ago
It can also be a lot simpler than that example, for instance just two steps: https://github.com/Tenzer/alga/blob/f3d85f62d055825044b60d78122ff0e0cd29683b/.github/workflows/checks.yml#L38
Tenzer | a day ago
I've replaced pre-commit with prek everywhere and it's working a treat!
[OP] kantord | a day ago
I'm not involved with this project but I tried it and had a really good experience. I was surprised to see that it was not posted here yet.
I really like pre-commit but unfortunately it never reached its true potential, mostly because it never reached the popularity it deserved.
I don't really think it needed a Rust rewrite for any technical reason... It's not like you couldn't make it just as performant in Python. But perhaps adding a native command is more palatable to non-Python teams. Or at least, more popular for whatever reason.
prek seems to be a painless drop-in replacement, and it already seems like it's gaining more widespread adoption even outside of the Python ecosystem. So I'm really excited to see where this greater popularity will take the concept.
fanf | a day ago
previously (6 comments)
freddyb | a day ago
I like what you’re doing here but I wonder if there’s an open issue on the lobsters repo and if it would make sense to do it automatically. In the end, the submission form is already listing them when you want to hit send.
fanf | a day ago
My “previously” links aren’t completely brainless. In this instance I remembered there have been discussions recently and did a quick keyword search.
freddyb | a day ago
I didn't mean to imply your comment was bad in any kind of way. It just seems I didn't understand it correctly. My apologies to you :-)
[OP] kantord | a day ago
for the record the submission form did not list this, but I suppose that is just because it's technically a completely separate website
freddyb | a day ago
Yeah I couldn’t find an issue and will file one when back at a larger screen
sknebel | a day ago
I think it only shows if its the same site, and then it already shows up under the comments too, see e.g. here https://lobste.rs/s/3fevyk/fast_servers_interesting_pattern
pgray | 16 hours ago
We switched to prek a while back and it's so fast