Live data from Hacker News

Saying Goodbye to GitHub

ersei.net

31–40 of 450 posts

Re: Saying Goodbye to GitHub

#31

Why hasn't someone just changed the GPL license already: "If you train an AI on this code, you must release the source code and generated neural net of that AI as open source" or something to that effect. It won't stop it, but it will slow it down, and it seems like the right T&Cs to put on training against GPL code because it gives an advantage to open source AIs, however minor.

Aren't they claiming that it's fair use? IANAL, but wouldn't that make the licence irrelevant if training AI/ML models was found to be fair use? And if not, it's a licence violation anyway?

IANAL either, but the license still applies to the end-user (the person who trained the AI) so it would seem like it would add at least 1 non-trivial license violation for that user?

Edit: I googled "fair use copyright US" and have now decided that US copyright law is stupid.

Re: Saying Goodbye to GitHub

#32
post #22
post #5

Q: What about the social features of GitHub? I enjoy the discovery and star features. There just isn't any community like it. Where can I get that? It would be great if there was something that had the discovery & social aspects but was agnostic to the actual site hosting the code. Does something like that exist?

Forgejo [1] is working on integrating activitypub [2] which will let instances federate with each other. [1] https://forgejo.org/ [2] https://codeberg.org/forgejo/forgejo/issues/59

+1 for Forgejo

Re: Saying Goodbye to GitHub

#33
> The code that was regurgitated by the model is marketed as "AI generated" and available for use for any project you want. Including proprietary ones. It's laundering open-source code. All of the decades of knowledge and uncountable hours of work is being, well, stolen. There is nothing being given back.

Leaving GitHub wont change that, OpenAI is training its models on every bit of code they can have, sourcehut, codeberg etc. If its public, they will train on it.

Also from my experience of trying to leave GitHub, you just end up having a couple of projects on your alternative platform, and everything else on GitHub. You are still active on GitHub, probably even more than your new alternative.

And if you want to build a community, you will quickly find out that the majority want to stick to GitHub, and leaving it can kill your projects chances of getting contributions.

Personally if the courts decide its fair use, that's it, I'm going back, its the best got platform out there, gitlab doesn't even compare in free features. However I have been eyeing Gitea and Gitea Actions, with it Codeberg could become a realistic choice for me.

To end it, here is a Hot take, I really hate Sourcehut.

it hard to use, the ui is .. Not great and trying to browse issues or latest commits is a nightmare.

Every time a project uses it, its a pain to deal with.

Re: Saying Goodbye to GitHub

#35
post #16

Earlier quoted context omitted.

They are really not going to care about what you put into your license file, they are just going to claim that the use of GitHub binds you to their terms of service and that this supersedes your own license. Good luck fighting that.

ToS doesn't supercede copyright though, does it? The real rub will be the first court precedent on whether GPTs infringe on source data IP. Could see it going either way: fundamentally transformative or not.

> ToS doesn't supercede copyright though, does it?

It does unless you can afford to sue Microsoft.

Re: Saying Goodbye to GitHub

#36

While I agree with the general observation of the article, the tone and wording feels needlessly harsh and subjective. Sentences like „I could go to GitLab, but they have their own fair share of issues and a disgusting UI/UX.“ really diminish the article and by association, also the author.

The author is a college freshman, and that seems pretty par for the course for someone in their late teens. It's not really an "article" targeted for HN critical review, just a blog post on a personal site.

This is what I noticed too.

We have a kid with no experience in the real world telling us what they think. Why should we care? I actually don't

Re: Saying Goodbye to GitHub

#37

I tend to disregard articles that default to the "Stochastic Parrot" argument. These tools are useful now, I don't personally care about achieving actual intelligence. I want additional utility for myself and other humans, which these provide now, at scale. By a lot of measures many humans perform at just about the same level, including confidently making up bullshit. This post reads like one of the "Goodbye X online…

It’s pretty ironic that the argument that LLMs are stochastic parrots actually sounds like the actual parrot.

I think we’re now way past that now with LLMs now quickly taking on the role of a general reasoning engine.

Re: Saying Goodbye to GitHub

#38
You don't give up GitHub as long as you're employed in many companies that have accepted it and are looking for information on pretty much most open source software these days. The only thing is to just host your own projects elsewhere. I did this a long time ago because I was unhappy over GitHub's interface. I am realizing further that all I need is a web page and a link on my personal web site. If I really wanted to publish a project, I'd probably find other avenues. I think SourceHut does look nice if I cared enough.

In some cases, "GitHub" has become a required skill in some jobs now (not git, mind you, but _GitHub_). I personally fumbled a little bit when I was asked that - what is it? The CI/CD stuff specifically, or the "using git" part which is the skill? :) I was a bit befuddled.

Re: Saying Goodbye to GitHub

#39

> It's a stochastic parrot Stopped reading right there. The author clearly has no clue of how neural networks work or what makes them tick.

Do you have any resources you'd recommend to form a better understanding of how NNs tick? I'd like to get a better intuitive grasp on whats going on - I've mostly just been responding to that with "Well, if stochastic parrotism can do all this..."

Re: Saying Goodbye to GitHub

#40
post #15

Unfortunately moving to Gitlab or Sourcehut doesn't really help, because the underlying model (GPT-x) is trained on the entire internet, so that includes all scrape-able websites. The only way for your data not to be used in GPT (and therefore Copilot) is to not to put it on any website or make it very difficult to access, like encrypting it.

> Unfortunately moving to Gitlab or Sourcehut doesn't really help, because the underlying model (GPT-x) is trained on the entire internet, so that includes all scrape-able websites. The only way for your data not to be used in GPT (and therefore Copilot) is to not to put it on any website or make it very difficult to access, like encrypting it.

Having the entire git history decorates specific chunks (at least entire commits) with context by the commit message. So you may not only process the entire repo at one specific state in time, but the entire history in at this point in time. There is valuable knowledge while making sense of it; But this is not accessible to us. It relies in the knowledge base of one company (or two).

Post reply on HN