good! Now the next challenge will be flagging & executing these proposals.
If you come up with an automated system that’s not going to provide such high accuracy, then if you rely on manpower alone that’s also prone to fail - I think it’s a gray area; What’s next, a combination of both with human asked to approve on the flagged repos? perhaps
You can get high accuracy if you don’t care about false negatives and only ban the most egregious shit.
I think simply never using a project with emoji in the readme is sufficient to capture 90% of slop without removing much useful.
That being said, there is no way to determine AI generated from not, consistently, and over the long term. I worry you’ll have people making accusations and there’s no practical way to disprove them. You’ll also have people working harder to hide AI generated code (maybe that’s good overall if they are actually editing and revising the slop code?).
But if someone uses AI to build a rough draft, then rewrites 80% for clarity, correctness, etc. is that worthy of banning their account? What if someone runs their repo through AI to check for bugs and vulnerabilities?what proportion if content becomes “predominantly”? What about someone using it for translation of the text parts of their repo?
I just think there are very hard problems this policy creates, which take away from other resources to manage.
I guess if you can find the worst offenders you can ease server load, like if they make pushes around the clock too fast, that’s one good signal.
Like if you have 2000 pushes in a day (pushes, not commits) and it’s clearly 24/7, that’s probably an autoslopper.
If someone is not doing that, who cares, the “agents directing agents” folks are the ones that cause the most problems.
Hm, I’m a heavy pusher. Tho, I don’t believe I’ve gotten past 200 a day
Git is actually very efficient. You need to be very inefficient with your workflows to cause strain on a git server.
The fact that you can just host a git server for free and most people don’t need to ever worry about limits is a wonder.
This is partially why I chose a paid host with resource limits. I have one repo with a large benchmark library that I wouldn’t feel okay putting on a free service, for example.
There’s been studies about the shape and structure of text. Humans go on tangents AIs don’t
Models are already trained so it’s harder and harder to tell. They don’t put emojis in code anymore, they don’t even comment it. The new ones stopped doing the massive comment blocks that was so telling or signing their name somewhere at the top. It’s going to be comoletely impossible to enforce in just a few months.
I use sometimes AI for some research, so I asked an AI about vibe coding
Simple question, do you trust an app coded by an AI from Google, Facebook, Musk and cia., even if it works?
Yes, you can use an AI to code some routines, helping in your work, but way different to create and release an app with a simple prompt.
100% it will eventually be a problem, specifically the knowledge loss at the bottom. I just don’t think banning generated code will work in two months.
Look at where image gen is at, and it’s way harder to make a real looking photo than to make something that looks like it’s human coded. It’s not feasible if you look at where the tech is going and how fast.
Eventually? It’s already.
Site doesn’t load
?? AlternativeTo and the article load perfectly and fast. Check your connections.
Check your browser. Try again in a text browser or security hardened browser. It doesn’t load.
Using a browser which can’t open a community driven, well known, normal and secure Webpage, isn’t security but masochism.
Text browsers (Lynx, Links, etc.) are usefull for some tasks but for sure not for browsing in daily use, same with a excess of security and unneeded blockings in a browser, which break most websites. Better check unknown websites (Blacklight, WebbKoll, Mozilla Observatory, etc.), monitor the netactivity (Portmaster, PiHole) and block data which really can filter your identity, not the tecnical data which are the same for millions of other users and are needed to render properly a page.
Based
I think this is stupid.
Why let someone else tell you what you can and cannot do? This is going to turn into some bullshit witch-hunting and bullying over who’s using AI instead of helping devs migrate off of the rapidly enshittifying microslop GitHub and championing a new generation of code hosting.
As much as I want to like Codeberg, and I do like forgejo, I find stuff like this a waste of resources to bother with. But then, this is partly why I didn’t migrate to codeberg in the first place, I don’t think I get along with the community at all.
Have you actually read their blog post about it? They’re pretty reasonable about the enforcement of this.
I had read the OP and “I Regret Migrating to Codeberg” post, I have now read Codeberg’s blog post.
I don’t see how it avoids the issue of “cry-bully” type behaviour that I mentioned, where it’s basically unprovable you aren’t heavily using LLMs, and I don’t see how this really helps with their loss of trust argument (I think this goes the opposite way).
The big takeaway I get from the blog post is if you have “community” then you can do whatever.
As I said earlier, I am not the intended Codeberg audience, which I already knew based on evaluating them and disagreeing on philosophy and use cases.
Link to that post:
The decision of Codeberg is due to its user votes
I’m still allowed to disagree with it, and as I said, I decided on a different host because my views differ from the codeberg community (pre this AI decision, and I wouldn’t call myself a vibe coder).
That being said, it sounds like not everyone is happy with how the vote was done (see the downvoted post about someone leaving codeberg).
Sure. Why don’t you host Slopberg?
I use a paid host that allows commercial works.
Codeberg was always incompatible with the projects I was working on, and that’s totally fine with me.
it seems like just a placeholder, I cant see sign up or anything.
I think it’s satirical anyways, I was able to see a login page when I clicked explore
Good, stay away!
