Goodbye GitHub ... ?

doul

doul

Hello,

Goodbye GitHub: that’s what I would have liked to say...

Codeberg is my original repository; I added GitHub to use Codex Cloud, which I no longer use, relying instead solely on the Codex CLI. I don’t like the GitHub interface: it’s slow to load, the Markdown rendering is unattractive, and it creates the extra burden of maintaining two separate repositories. While GitHub offers high visibility, it’s clear that, after two years, it hasn’t provided me with any real additional benefit. Furthermore, my increased confidence in Codex means that maintaining a second repository to isolate coding agent is no longer justified. For remote backups, I use Proton Drive to store my code and Git files.

Things have evolved: I’ve become more proficient with Git, I have greater trust in Codex, and I have an additional backup on my work computer. I plan to keep my triple-branch structure -> `codex/<feature>` -> `test` -> `main` on Codeberg, while putting GitHub on hold for the time being. The only potential drawback with Codeberg is their stance on AI; they discourage large volumes of AI-generated code. While my own work doesn’t fall into that category, their criteria aren’t strictly quantitative, and they could potentially reject any AI-generated code outright.

What do you think of? How do you handle it on your end? Thanks.
 
I understand minimizing traffic to the site by Ai agents.
Not sure what other reasons there would be for the limitations of Ai.

Github is resembling LinkedIn in it's use, i.e. flexing for HR.
Only a matter of time it has some horrendous CVE like GitLab.
Proton drive is smart move - different stack and jurisdiction.
 
doul said:
Hello,

Goodbye GitHub: that’s what I would have liked to say...

Codeberg is my original repository; I added GitHub to use Codex Cloud, which I no longer use, relying instead solely on the Codex CLI. I don’t like the GitHub interface: it’s slow to load, the Markdown rendering is unattractive, and it creates the extra burden of maintaining two separate repositories. While GitHub offers high visibility, it’s clear that, after two years, it hasn’t provided me with any real additional benefit. Furthermore, my increased confidence in Codex means that maintaining a second repository to isolate coding agent is no longer justified. For remote backups, I use Proton Drive to store my code and Git files.

Things have evolved: I’ve become more proficient with Git, I have greater trust in Codex, and I have an additional backup on my work computer. I plan to keep my triple-branch structure -> `codex/<feature>` -> `test` -> `main` on Codeberg, while putting GitHub on hold for the time being. The only potential drawback with Codeberg is their stance on AI; they discourage large volumes of AI-generated code. While my own work doesn’t fall into that category, their criteria aren’t strictly quantitative, and they could potentially reject any AI-generated code outright.

What do you think of? How do you handle it on your end? Thanks.
Click to expand...
So basically instead of using Github resources with AI stuff that Github love all about, you prefer to use Codeberg resources which are limited and does not really want to have AI generated project in it to put your AI generated code ?

Seems kinda weird of a workflow.
 
DanielBowman said:
Not sure what other reasons there would be for the limitations of Ai.
Click to expand...

There are other reasons : Protecting our FLOSS commons from LLMs

monwarez said:
So basically instead of using Github resources with AI stuff that Github love all about, you prefer to use Codeberg resources which are limited and does not really want to have AI generated project in it to put your AI generated code ?

Seems kinda weird of a workflow.
Click to expand...

My project is not AI generated, no CI/CD. If Codeberg doesn't want any AI-generated code, they could just say so directly, but that’s not the case.
 
doul said:
There are other reasons : Protecting our FLOSS commons from LLMs



My project is not AI generated, no CI/CD. If Codeberg doesn't want any AI-generated code, they could just say so directly, but that’s not the case.
Click to expand...
Aren't using codex cli imply being somewhat in those category listed by the following section:

However, we also need to be honest about certain use cases that might no longer be welcome on Codeberg. If you see yourself on this list, you don't need to move right away, but there might be other places that better fit your needs:
  • Projects that are created by LLM "agents" in autonomous ways
  • Projects written and maintained with heavy use of LLMs
  • Projects where the amount of resources (e.g. storage, CI/CD) is significantly larger than what the involved amount of people could have created by hand
  • Projects heavily tied to the LLM ecosystem, e.g. LLM-written tools to ease LLM usage
  • Users sending LLM contributions in violation of project's custom policies+
 
monwarez said:
  • Projects that are created by LLM "agents" in autonomous ways
  • Projects written and maintained with heavy use of LLMs
  • Projects where the amount of resources (e.g. storage, CI/CD) is significantly larger than what the involved amount of people could have created by hand
  • Projects heavily tied to the LLM ecosystem, e.g. LLM-written tools to ease LLM usage
  • Users sending LLM contributions in violation of project's custom policies+
Click to expand...
I can answer "no" to these 5 questions, so I don't have to leave right away. That’s a big help, I must say.

It seems to me that the foundation of FOSS is sharing and my code is shared, even though it was created with the help of AI. Ultimately, we always come back to the same problem: I could easily hide the fact that the code originated from an AI (breaking down large commits, caching config files, no AI names). So, what do we do then? Where is the line, is 20% of AI-generated code is acceptable, but 25% isn't? Or is it better to lie and claim there’s no AI-generated code at all?
 
doul said:
I can answer "no" to these 5 questions, so I don't have to leave right away. That’s a big help, I must say.

It seems to me that the foundation of FOSS is sharing and my code is shared, even though it was created with the help of AI. Ultimately, we always come back to the same problem: I could easily hide the fact that the code originated from an AI (breaking down large commits, caching config files, no AI names). So, what do we do then? Where is the line, is 20% of AI-generated code is acceptable, but 25% isn't? Or is it better to lie and claim there’s no AI-generated code at all?
Click to expand...
It seems that you didn't fully read the actual link that you gave us here.
And you didn't read all my message, since you said that since you can no to these 5 questions you don't have to leave right away.
My previous quote literally said: If you answer yes you don't have to move right away, but this may not be welcome on Codeberg in the future.

Well while we can all agree that by sharing your code and allowing it to be modified and shared you are in the foundation of FOSS .
However not everything is about the code itself, there are other factor like the ecological impact of the usage and encouraging of LLM tooling.
There is the economical aspect that make owning your own computer more and more expensive.

So no Codeberg won't do the witch hunt to know how much AI was in your project, that would be a waste of resources for them.

But by using LLM tools, people encourage the big AI company to continue to buy more resources which will have the direct effect of increasing the price of anything related to a computer.
And it is the same for paid developer, they should when possible (like on country where worker protection actually exist) push back against theirs employer against LLM, unless of course if they embrace the LLM ways, and then won't complain about being able to buy any computer.
Because it cannot be both ways, you can't use LLM tools and then complains that you can't buy ram.
 
Cthulhux said:
I don’t think that Git is the best choice for most people in the first place, so I avoid it’s whole ecosystem.
Click to expand...

Most people use it because everybody else does. No other reason. Just bandwagon jumping.
FreeBSD uses it for the same reason but slightly more legitimate. The claim is that it helps when trying to get other developers to work on FreeBSD because they're used to the tool.
There are other tools that are easier to use and do a better job but are in the minority now and I fear may fade away.
 
The issue isn't about Git itself; it's about GitHub. I spent time learning Git, it does what I need it to do, and I don't see any reason to switch to something else.
 
monwarez I understand what you mean, but let’s not get into a political debate, what you’re saying could apply to anything, and it doesn’t change the fact that Codeberg lacks a clear stance on AI, using vague terms like "heavy use."

I’m just looking for a remote repository to store some code in and gain a bit of visibility 😔
 
You can always selfhost gitea, forejo or other options. The main reason to be on one of the shared sites is either for visibility or to share, and you can always send the private repo to a public one.

I generally just use git on my local filesystem, but I'll likely switch to gitea in a jail if I start coding regularly on my laptop.
 
ai is stupid.

why even use a GUI (like gitlab or gitea or...)? You're already using the shell to interact with your VCS. My git server is headless (you should do that too).

git, fossil, hg, whatever is basically the same for a one-person-project so, what does that have to do with anything (they are all decentralized VCS--the "interface" is a minor)? there is also, knowing git will allow you to switch to got.
 
You must log in or register to reply here.
Back
Top