Live data from Hacker News

Give up GitHub: The time has come

sfconservancy.org

21–30 of 585 posts

Re: Give up GitHub: The time has come

#21
post #10

Earlier quoted context omitted.

Are you saying they use GitHub (via the web editor?) but have never heard of Git? or don't know how Git works? or just don't know the Git CLI? or something else?

I think many a junior just never initialized a repository, and 'git clone' is what they've used to "get a repository". So I would forgive them for thinking that a remote is needed for having a repository on their drive.

IMO the issue here is just that people think of git as it was something like SVN where you need to have some kind of external server to host project. Which isn't surprising as most of the time (when you cooperate with others) you do exactly that: push changes to one server. Idea of decentralization and independent copies of repositories may not be easy to grasp for newbe, because it's really technical concept.

Re: Give up GitHub: The time has come

#22

> For its part, Git was designed specifically to make software development distributed without a centralized site. Just yesterday I had to explain the basic premise/history of Git to a young intern. I had asked him if he was using Git to manage his little pet project the company gave him to play with. “No”, he replied, he didn’t know what the company’s policy was to posting code in public on GitHub. As I explained to…

Github has carefully nurtured this culture of centralization. It's a marvel of our times.

Re: Give up GitHub: The time has come

#24

Earlier quoted context omitted.

What is it with Americans and their dismissal of other people's concerns as always being "politics". It totally does break licencing laws, just because Americans think they can change the terms of an agreement and relicense your code because you didn't remove it doesn't make that legal.

I haven't looked into this extensively or tried Copilot myself, so I might be completely wrong about this. But from what I understand, the code Copilot generates is generally different enough from the source data that it shouldn't be an issue. In a sense, Copilot reading lots of code on Github to train a code-writing model is analogous to a human reading lots of code on GitHub and learning from some of the design pat…

But we never had humans who could read so much code and "learn" so much. :p Our old system's rules and logical conclusions, taken up a level, now make us uncomfortable.

Re: Give up GitHub: The time has come

#25

Quoted post unavailable.

What is it with Americans and their dismissal of other people's concerns as always being "politics". It totally does break licencing laws, just because Americans think they can change the terms of an agreement and relicense your code because you didn't remove it doesn't make that legal.

https://youtu.be/sCvw_eAT9aM?t=166

Re: Give up GitHub: The time has come

#26
post #9

This GitHub Copilot thing is pretty polarizing. According to one of the expert papers they sollicited, it's unlikely MSFT/GitHub is breaking any copyright laws. What they're doing does feel icky, and they could have mitigated a few concerns by a) making the inventory of the full training set public and b) at least attempt to attribute if there is a direct copy (which by their own admission happens about 0.1% of the t…

>at least attempt to attribute if there is a direct copy (which by their own admission happens about 0.1% of the time). IIRC they just block that suggestion if they detect it's a copy now

More specifically it's an option, the VSCode Copilot extension asks you to configure it and takes you to a GitHub page where you can choose what happens if a suggestion is a direct copy.

Re: Give up GitHub: The time has come

#27

> For its part, Git was designed specifically to make software development distributed without a centralized site. Just yesterday I had to explain the basic premise/history of Git to a young intern. I had asked him if he was using Git to manage his little pet project the company gave him to play with. “No”, he replied, he didn’t know what the company’s policy was to posting code in public on GitHub. As I explained to…

That scenario plays itself out time and time again. 1. Base technology arrives and is adopted en masse 2. Some entity wraps it in an easier-to-use interface 3. After some time, that interface becomes the de facto standard 4. Developers who started developing after #2 and especially #3 don't understand how the underlying technology works or even that the wrapper is just a wrapper I'm reminded of how many Juniors I tra…

I find this so weird, I always feel like I have to at least a bird's eye view of how bits and bytes are flowing through the system. Not that a lot of boxes I don't just label as "magic", but at least I need to know the box exists

Re: Give up GitHub: The time has come

#28
post #16

Earlier quoted context omitted.

What is so baffling about explaining git to a novice? I'm more baffled that you are baffled about this. Git wasn't even invented yet when I was in school and I did not learn about SVN until my first programming job in 2001 where a senior explained it to me.

I think the parent's point was that the intern knew about github and assumed git was github.

I used to be confused by this. It's because "git" is in the name and GitHub is the first thing you encounter, so I was like, "oh, is git some command line extension of GitHub?" rather than the reverse. :D

Re: Give up GitHub: The time has come

#29

Earlier quoted context omitted.

What is it with Americans and their dismissal of other people's concerns as always being "politics". It totally does break licencing laws, just because Americans think they can change the terms of an agreement and relicense your code because you didn't remove it doesn't make that legal.

I haven't looked into this extensively or tried Copilot myself, so I might be completely wrong about this. But from what I understand, the code Copilot generates is generally different enough from the source data that it shouldn't be an issue. In a sense, Copilot reading lots of code on Github to train a code-writing model is analogous to a human reading lots of code on GitHub and learning from some of the design pat…

code output is unrelated to the issue.

the issue is if GitHub hosts your code with an open source license that does not allow for-profit reuse without concomitant sharing, like the GPL, they will incorporate that code into their product in violation of the license, and claim the license doesn’t apply because the code isn’t code but merely text.

they built their business on precise and detailed articulations of consent and its boundaries, but disregarded all of that, post-acquisition, because Microsoft has enough lawyers that they think they can get away with it.

so, for the subcommunities and creators who put their work on GitHub in the context of these very specific and fine-grained articulations of consent, this may be theft and is certainly betrayal.

Re: Give up GitHub: The time has come

#30
post #18

Does Github give users a toggle to opt-out of being in the copilot training dataset? I feel like that would be a decent compromise for folks not down with their code being used in copilot.

a nice, easy compromise they could have proposed and implemented quite some time ago.
Post reply on HN