Live data from Hacker News

GitHub scraped your code. And they plan to charge you

twitter.com

141–150 of 212 posts

Re: GitHub scraped your code. And they plan to charge you

#141
post #28

Earlier quoted context omitted.

It's not about it being free, it's about GitHub taking something you licensed with conditions (ie. attribution or keep this copyright notice and license file, etc.) and blatantly ignoring your license because they know you probably can't afford to sue them for copyright infringement. Open Source doesn't mean you can reproduce and copy the code freely, licenses exist for a reason. Also: of course people care about Fac…

Your license means nothing if you can't defend it. If you paid for the service they wouldn't dare using your code regardless of the license.

> Your license means nothing if you can't defend it.

It's okay to commit crimes if the victims are poor enough? (And yet what you've said is still true.)

Re: GitHub scraped your code. And they plan to charge you

#142
post #132
post #115

I have a genuine question about this whole thing with Copilot: A similar product, TabNine, has been around for years. It does essentially the exact same thing as Copilot, it’s trained on essentially the same dataset, and it gets mentioned in just about every thread on here that talks about AI code generation. (It’s a really cool product btw and I’ve been using and loving it for years). According to their website they…

The feature of TabNine that uses the "public" dataset is optional. It can also provide completions only based on local code. That optionality is important. Also, tabnine has a smaller scope; you type "var " and it suggests a variable name and possibly the rest of the line, like autocomplete has been doing for decades. Perfectly normal. My understanding of copilot is that you can type "// here's a high-level descripti…

> It can also provide completions only based on local code. That optionality is important.

I don’t see how? The question is about the ethics of building such a tool, not whether anyone is forced to use it.

Re: GitHub scraped your code. And they plan to charge you

#143

Is this analogous to gtp-3 reading every sentence ever written without attribution to all of mankind?

A little. But all works output by GPT-3 are provided in “source form” to everyone who uses them – whereas lots of the output of Co-Pilot (trained on copyleft code, among other things) is going into proprietary software projects.

(Also, GPT-3 wasn't trained on nearly as much writing as that. Even if you ignore lost writing, GPT-3 was trained on a small subset of the 'net.)

Re: GitHub scraped your code. And they plan to charge you

#144
post #116
post #53

Earlier quoted context omitted.

If an individual hypothetically painstakingly searched through GitHub to see how others wrote an API call and copy-pasted, almost no one would have a problem with that even if they didn't attribute every little code snippet. But some are bringing out the pitchforks because ML can basically do that painstaking search (yes, I know it's not literally a search) so efficiently that it's actually (maybe) useful as a tool.…

Scale, intent and impact do actually matter in copyright. Copying a single API can be ok, and copying a million of them might not be OK. There's a reason for the word "fair" in "fair use". If a whole project was copied verbatim and the license violated I think everyone would agree that was wrong. So then is copying the same quantity of code across 1000 projects wrong? Is setting up a process and a system that does th…

Doesn't "fair use" limitations go out the window if the license is permissive enough to allow commercial use of the source code? As I've said in other comments, I still think this was the wrong way for GitHub to do this-- opt-in would have been much better-- but outside of licenses with commercial restrictions I'm not sure there's a license violation here, no more than scraping hundreds of public domain books to create paid reprints would be a problem, or using the same to create a corpus for NLP machine learning.

Re: GitHub scraped your code. And they plan to charge you

#145
post #53

Earlier quoted context omitted.

If an individual hypothetically painstakingly searched through GitHub to see how others wrote an API call and copy-pasted, almost no one would have a problem with that even if they didn't attribute every little code snippet. But some are bringing out the pitchforks because ML can basically do that painstaking search (yes, I know it's not literally a search) so efficiently that it's actually (maybe) useful as a tool.…

If large company violated my open source license, I'd certainly have a problem with it. Use my code all you want, but follow the basic rules. There's entire websites dedicated to GPL violations; people do care.

If the license allows commercial use, what would the GPL violation be in this case?

Re: GitHub scraped your code. And they plan to charge you

#146
post #15

If you open-sourced code and allowed it to be used for commercial purposes, I don't see the point of being pissy about Github using it, I'm saying this as someone who's written quite a lot of MIT code. (And charging for a product which adds value to your developer experience and needs money to be run is not a bad thing)

> If you open-sourced code and allowed it to be used for commercial purposes Uploading it to Github does not transfer ownership or imply allowances for any use. If you upload it without a license it is a copyright violation to copy the code. Even with an MIT license it is a copyright violation to copy the code without attribution. > I don't see the point of being pissy about Github using it, I'm saying this as someon…

> Even with an MIT license it is a copyright violation to copy the code without attribution

Probably a copyright violation. There are surely circumstances in which copying a small portion would either fall under fair use, or for other reasons not constitute a violation. The question then is whether or not codepilot is causing a violation. I don't think it's as clear cut as most commenters are making out.

All in all though, it's probably going to take a few court cases to figure out. In the mean time, I'd expect most companies to steer clear of codepilot.

Re: GitHub scraped your code. And they plan to charge you

#147
post #130

So.. I can see that this ML model is generating some code exactly same as the original dataset, which definiately a problem. A defect model, sure. Beside that, I cannot understand why the overall idea, using open-source project to train a ML model that generates code would ever be a problem. We human beings are learning as the model, we read others code, books, articles, design patterns... and it becomes part of us.…

> A defect model, sure

More like a defect approach, behavior like that is well known(1) to be basically guaranteed to happen with GPT-3 and similar approaches.

(1): By people involved in the respective science categories (Representation Learning/Deep Learning, NLP, etc.).

Re: GitHub scraped your code. And they plan to charge you

#148
post #130

So.. I can see that this ML model is generating some code exactly same as the original dataset, which definiately a problem. A defect model, sure. Beside that, I cannot understand why the overall idea, using open-source project to train a ML model that generates code would ever be a problem. We human beings are learning as the model, we read others code, books, articles, design patterns... and it becomes part of us.…

The difference is that it's a judgement call when to include attribution, whom to attribute with how much, and overall whether something is too close to be counted as a copyright or other license violation or not. Intelligent humans sometimes, or even often times, have a hard time doing this judgement call. An artificial intelligence would too, and a somewhat simple ML model (no offense) certainly does.

I'm really waiting for this to blow up from the open source license angle. Freely combining code with different license is a hellish undertaking on its own. But already just re-using some, say, GPL code, even staying under the same license, but without proper attribution, is Forbidden with capital F.

Re: GitHub scraped your code. And they plan to charge you

#149

Earlier quoted context omitted.

> It it's actually open source without commercial restrictions though, I don't see an ethical difference in using the code directly or for an meta analysis driving ML for enhanced code completion. Most Open Source code comes with a requirement to carry over license note, which Copilot does not do. Additionally, ethics dictate you attribute the source when copying directly, something the Copilot also doesn't do.

Are they actually using the source to make a derivative/fork though? If reusing the code in another codebase then definitely attribution would be required. But using it as a dataset seems a bit different-- a grey area. Though I would still agree that the right thing to do would be to have an attribution area, even if it was thousands of entries long. Whether technically required by the license or not, the spirit of t…

And if course GPL would attach where applicable .....

Re: GitHub scraped your code. And they plan to charge you

#150

Can't you host code on GitHub that is not "free" for commercial use? If GitHub scraped these projects then it's a problem. Otherwise, Into honestly trying to have a conversation on this to understand the objections because I haven't made up my mind but struggle to see the problem. So pease consider the following: if the code was not encumbered by restrictions I don't see an obvious problem with this. Using code or da…

The difference appears when copyrighted material is repeated verbatim. And because it's obvious Github has no control over how much copyrighted material is being repeated verbatim. And that copyrighted material is intended to be used by commercial companies who copyright their own material and don't want to have their copyright challenged.

Under US rules at least, everything automatically falls under copyright. But if the license allows full use even for commercial purposes, verbatim repetition is no different in this context than if I included it in my own new piece of software. (Of course attributions and distribution of the code & modifications would frequently be requires... with Pilot it's... potentially a gray areas on whether code lumps together with countless other software is actually modified in the traditional sense of the term, but attribution should still be an obvious requirement. And where the other license requirements are strictly required or not, I still think it would be the right thing to do for GitHub to honor the spirit of those strictures, especially considering their entire business model is based on a majority of users trusting their code to them.

If they're not going to act accordingly, there's no reason someone couldn't roll their own GitLab instance, or a competitor with more respect enter the marketplace.

Post reply on HN