Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

All public GitHub code was used in training Copilot

twitter.com

Post reply on HN
Print view View original

671–680 of 734 posts

Re: All public GitHub code was used in training Copilot

#671

Post by TuringNYC »

user5994461 wrote: post #480
TuringNYC wrote: post #165

Earlier quoted context omitted.

IANAL but the serious answer -- i think -- is that you always use things at your own risk, even purchased tools, and are protected via indemnity agreements. If there is no indemnity agreement (is is the case here), you assume the risk. That said, if enough people are bitten by this, i'm not sure what happens -- does anyone know of a relevant case. One somewhat relevant case that caused mass pain was the SCO Linux Dis…

If you're thinking about the liability waiver found in many licenses and contracts and EULA and other, they are often void, depends on the jurisdiction. The official answer from Github that they take all input on purpose doesn't play in their favor.

I'm speaking specifically about the indemnity agreement that you get as part of a purchased license. It is the opposite of the liability waiver -- it is saying that the software publisher will take on responsibility in certain cases and with certain limits.

For example, if I purchase certain corporate Linux licenses, i'm protected against being sued if something in the distribution ends up having misappropriated code.

Check out the SCO Linux Dispute for how bad things can get for corporations: https://en.wikipedia.org/wiki/SCO%E2%80%93Linux_disputes

Re: All public GitHub code was used in training Copilot

#672

Post by kulikalov »

fleddr wrote: post #551

To me, the particular use case and whether it is fair use or not, is of minor interest. A far more pressing matter is at hand: AI centralization and monopolization. Take Google as an example, running Google Photos for free for several years. And now that this has sucked in a trillion photos, the AI job is done, and they likely have the best image recognition AI in existence. Which is of course still peanuts compared…

This seem to be inevitable. An individual doesn't have horizontal scalability, you know... So, unless we'll have some kind of brain extension capabilities, there is no other choice but to build such technologies collectively.

Also, I think you are overdramatizing this. Governments used to be omnipresent (maybe still are), in a different way, more threatening to individuals and probably as threatening to societies as "everything companies" could be.

Re: All public GitHub code was used in training Copilot

#673

Post by redeux »

sdevonoes wrote: post #669
fleddr wrote: post #551

To me, the particular use case and whether it is fair use or not, is of minor interest. A far more pressing matter is at hand: AI centralization and monopolization. Take Google as an example, running Google Photos for free for several years. And now that this has sucked in a trillion photos, the AI job is done, and they likely have the best image recognition AI in existence. Which is of course still peanuts compared…

We can decide to stop using some (or most of) Google services. It's hard, but it's not that they are pointing us with a gun in order to use their services, right? Sure, for the cases when one cannot escape Google, use it; but for the rest of scenarios? It's all about tradeoffs: Can I live without YouTube? Can I live with DuckDuckGo (Google Search is "better" but I don't mind)? etc.

You, a person knowledgeable in this field, may choose to stop using Google services, but that won’t have any societal impact if you can’t also convince the “average” user to do the same.

Re: All public GitHub code was used in training Copilot

#674

Post by strogonoff »

habitue wrote: post #353

I am really confused by HN's response to copilot. It seems like before the twitter thread on it went viral, the only people who cared about programmers copying (verbatim!) short snippets of code like this would be lawyers and executives. Suddenly everyone is coming out of the woodworks as copyright maximalists? I know HN loves a good "well actually" and Microsoft is always suspect, but let's leave the idea of code la…

As a mature, skilled engineer, you wouldn’t mind sharing your knowledge—but you’d really prefer to do this on your own terms.

First, you might choose to distribute your code under a copyleft license to advance the OSS ecosystem. Second, the older you get, the more experience you accumulate, paradoxically the harder it is for you to find a job or advance your career in this industry—so, to maintain at least some source of motivation for tech companies to hire you, you may choose to make some of the source available, but reserve all the rights to it.

You’re fine making the source of your tool or library open for anyone to pass through the lens of their own consciousness and learn from it, but not to use as is for own benefit.

Now with GitHub Copilot suddenly you see the results of your labour you’ve previously made (under the above assumptions) public being passed through some black box, magically stripped from your license’s protections, and used to provide ready-made solutions to everyone from kids cheating at college tests to well-paid senior engineers simply lacking your expertise.

I hope it’s easy to spot how engineer’s interests in the above example are not necessarily aligned with GitHub’s, how this may be perceived as an unfair move disadvantaging veteran rank-and-file software engineers while benefitting corporate elites and investors, and subsequently has the potential to disincentivize source code sharing and deal a blow to OSS ecosystem as a whole.

Re: All public GitHub code was used in training Copilot

#675

Post by jefftk »

rubyist5eva wrote: post #432
jefftk wrote: post #290

Earlier quoted context omitted.

That doesn't work: your suggestion applies at too late a stage in the flowchart. It looks like: 1. Do you need a license to use materials for training, or to use the output model? 2. If so, does the code's license allow this? GitHub is claiming 'no' for #1, that they do not need any sort of license to the training materials. This is reasonably standard in ML; it's also how GPT-3 etc were trained. Now, whether a court…

If the answer is 'no' for #1 than the GPL might as well not exist because now we can just launder it through co-pilot and close it off, a rather distorted interpretation of "fair use" if you ask me. "Dear copilot, I'm writing a Unix-like operating system...."

I don't think that's right; I wrote a response above: https://news.ycombinator.com/item?id=27779155

Re: All public GitHub code was used in training Copilot

#676

Post by abiro »

fleddr wrote: post #551

To me, the particular use case and whether it is fair use or not, is of minor interest. A far more pressing matter is at hand: AI centralization and monopolization. Take Google as an example, running Google Photos for free for several years. And now that this has sucked in a trillion photos, the AI job is done, and they likely have the best image recognition AI in existence. Which is of course still peanuts compared…

> If we predict that ultimately AI will change virtually every aspect of society, these companies will become omnipresent, "everything companies". God companies.

What we currently call AI is very from AGI, and it's not clear that sitting on piles of proprietary data gives an edge towards AGI. If the goal is human-level intelligence, that has been demonstrably achieved with the far lesser resources of the public school system. :)

Current DL systems need huge amount of data, because they are very primitive: they work with immediate associations, so they require seeing data very similar to all possible inputs to generalize well.

As we develop more sophisticated systems, I expect that the leverage from data will tip over to engineering finesse, and nothing is better at fostering great engineering than the permissionless tinkering environment of open source.

Re: All public GitHub code was used in training Copilot

#677

Post by visarga »

abiro wrote: post #676
fleddr wrote: post #551

To me, the particular use case and whether it is fair use or not, is of minor interest. A far more pressing matter is at hand: AI centralization and monopolization. Take Google as an example, running Google Photos for free for several years. And now that this has sucked in a trillion photos, the AI job is done, and they likely have the best image recognition AI in existence. Which is of course still peanuts compared…

> If we predict that ultimately AI will change virtually every aspect of society, these companies will become omnipresent, "everything companies". God companies. What we currently call AI is very from AGI, and it's not clear that sitting on piles of proprietary data gives an edge towards AGI. If the goal is human-level intelligence, that has been demonstrably achieved with the far lesser resources of the public schoo…

> If the goal is human level intelligence, that has been demonstrably achieved with the far lesser resources of the public school system.

Seems unlikely human education costs less than AI education in total.

Re: All public GitHub code was used in training Copilot

#678

Post by visarga »

fleddr wrote: post #551

To me, the particular use case and whether it is fair use or not, is of minor interest. A far more pressing matter is at hand: AI centralization and monopolization. Take Google as an example, running Google Photos for free for several years. And now that this has sucked in a trillion photos, the AI job is done, and they likely have the best image recognition AI in existence. Which is of course still peanuts compared…

They are not hoarding the latest results, except for a few cases where the general public is a year behind their secret sauce. Take a look at the huge zoo of planetary-scale models that are published by the big companies and universities (HuggingFace, https://modelzoo.co/, ...)

The problem with the huge models like GPT-3 is that they are too expensive even to run by regular people, not train.

Re: All public GitHub code was used in training Copilot

#679

Post by SirensOfTitan »

abiro wrote: post #676
fleddr wrote: post #551

To me, the particular use case and whether it is fair use or not, is of minor interest. A far more pressing matter is at hand: AI centralization and monopolization. Take Google as an example, running Google Photos for free for several years. And now that this has sucked in a trillion photos, the AI job is done, and they likely have the best image recognition AI in existence. Which is of course still peanuts compared…

> If we predict that ultimately AI will change virtually every aspect of society, these companies will become omnipresent, "everything companies". God companies. What we currently call AI is very from AGI, and it's not clear that sitting on piles of proprietary data gives an edge towards AGI. If the goal is human-level intelligence, that has been demonstrably achieved with the far lesser resources of the public schoo…

> If the goal is human-level intelligence, that has been demonstrably achieved with the far lesser resources of the public school system.

Pretending that the scientifically managed public school system, that attempts to manufacture uniform educated humans on a conveyer belt, is responsible for human education is fairly ridiculous.

Children have a remarkable capacity to learn, and do so automatically through free play and exploration until public education wrings that curiosity out of them and turns education into a job.

Humans get educated despite the public education system, not because of it.

Re: All public GitHub code was used in training Copilot

#680

Post by devetec »

eloisius wrote: post #664
devetec wrote: post #622

Earlier quoted context omitted.

It would be legal! But it wouldn't "reconstruct" Microsoft software. The way Copilot works is just that, a copilot. It's not the pilot. It's your own fault for what you do with it, it's just giving you some help along the way.

So long as the "copilot" is a black box that no one can inspect, how is it substantially different than me creating a website with a link to download a licensing-stripped version of Microsoft office, but it only gives you a verbatim copy 1/10 times you try it?

There's an entire academic paper detailing exactly how it works. https://arxiv.org/abs/2107.03374
Post reply on HN