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

GitHub Copi­lot inves­ti­ga­tion

githubcopilotinvestigation.com

Post reply on HN
Print view View original

641–650 of 1001 posts

Re: GitHub Copi­lot inves­ti­ga­tion

#641

Post by 123_abc »

woah wrote: post #491

It would be sad if someone succeeded in shutting down CoPilot for this kind of copyright stuff. It is genuinely useful. I don't care that it reproduces copyrighted content. The only way you can get it to do that is to bait it with the function names of functions that have already been copy and pasted thousands of times onto GitHub without proper licenses. Luckily, someone will probably come out with a "renegade" vers…

>The only way you can get it to do that is to bait it with the function names of functions

I get that using function names is an obvious way to get copilot to generate contested code, but has someone tried to get copilot to generate contested code in a way that users might sincerely be using copilot for productivity ? How do you get to the claim it is the "only way" ?

Re: GitHub Copi­lot inves­ti­ga­tion

#642

Post by Vetch »

iepathos wrote: post #554
schoen wrote: post #486

Here are a few thoughts I haven't formulated before: It seems clear enough to me that training AIs on copyrighted works is typically or commonly a fair use under existing law, because the AIs can and commonly do learn non-copyrightable elements and aspects of those works. It's very obvious from enormous numbers of examples that current AI systems are capable of learning much more abstract features of human culture (g…

Humans do violate copyright if they use copyrighted passages directly in their work and pass it off as their own without any attribution, which is what copilot has been show to sometimes do, though not always. Copilot will sometimes offer chunks of code that can be found verbatim in open source code bases and passes it off to users without attribution. I agree it is ok to learn from copyrighted work and reproduce new…

> which is what copilot has been show to sometimes do

In those cases it seems that humans are already copying code without also propagating licenses appropriately. LLMs are more likely to memorize things which occur a lot (and I'd bet rare things that are representative of some conceptual axis).

The main examples presented so far, Davis and Carmack, have the property of having been copied a lot. The generative model is only surfacing an existing pattern of ignoring attribution. Sort of like the code-gen version of generating bigotry if appropriately prompted.

I'll also note that this pattern of retrievable memorizing of copyrighted and sensitive material is present in GPT-3 too and not just for code. As the situation is equivalent, a lawsuit should address the concerns of non-programmers too.

Re: GitHub Copi­lot inves­ti­ga­tion

#643

Post by Siira »

nonbirithm wrote: post #571
mkr-hn wrote: post #529

A sizable, possibly plurality cohort of fully adult tech people is young enough to not know about United States v. Microsoft Corp. This would explain a lot of comments I see on this topic. If you don't know Microsoft's history, a lot of what more informed people are worried about seems overblown. Copilot was Microsoft's first test of people's trust after the GitHub acquisition. It's going very, very, very poorly. The…

> Too many people are focused on what's legal. It's fine to think of, but law is the last stop before the breakdown of society. Microsoft skipped society and went straight to sparking an inevitable test of and possible reshaping of copyright law. Maybe it's illuminating of a trait of human nature. On the stable diffusion webui repo many people have stated that they would continue to use the code even if it were stole…

When someone presents some information publicly, copying it is a natural right. Copyright can only be legitimized as the state sponsoring short-term monopolies in select areas to subsidize industries that benefit the public (from receiving the said subsidies). Obviously, a technology that let’s the public creative the desired output with little cost is the sort of thing that can eliminate the need for such subsidization in the first place.

Re: GitHub Copi­lot inves­ti­ga­tion

#644

Post by welder »

This reminds me of pirating music. Lawyers tried futilely to stop it, but if something is technically possible people will find a way to keep doing it. Maybe you set some legal precedent on fair use with AI, but it won't prevent the real world usages if there's a benefit to the technology.

Re: GitHub Copi­lot inves­ti­ga­tion

#646

Post by Siira »

adastra22 wrote: post #516
maoeurk wrote: post #490

Earlier quoted context omitted.

> There are two issues -- (1) feeding copyrighted material in to an AI model, and (2) getting copyrighted material out. > The latter is obviously a violation of copyright, full stop. It's not obvious to me that (2) is a violation of copyright. Unlike patents, copyright violation is not as simple to prove. My understanding is that, at least in the US, independent creation is a valid defense against copyright infringem…

Get Stable Diffusion to output Micky Mouse and see how far you can use that commercially without Disney stomping down on you hard . Outputting copyrighted material is a violation of copyright, period. Whether that violation is enforceable depends on your means though.

And why is Micky Mouse not in the public domain as of 2022? There lies in the root of all these questions. The system is not designed to benefit people, but rent-seeking.

Re: GitHub Copi­lot inves­ti­ga­tion

#647

Post by ShamelessC »

smaudet wrote: post #613
nopenopenopeno wrote: post #606

Earlier quoted context omitted.

AI copyright drama is my favorite gossip these days because it can’t be reconciled until we accept that intelligence is created and held by societies, not individuals. Recent AI is a new way to exercise that intelligence, but it presents a major conflict with capitalism.

Not a conflict with capitalism - it's another trajedy of the commons - the robber barons of old stole the owned commons land and started said 'capitalism'. Capitalism is still very alive, and will continue to be. It's in conflict with the general welfare of the people...

I assume they meant “capitalism as it is currently implemented”. In any case, not like capitalism is suggested as being under threat - just that an economic model based on competitive markets will have a lot of issues with fairly allocating resources to all individuals, instead accumulating most of it to a few industry leaders.

Something will have to change.

Re: GitHub Copi­lot inves­ti­ga­tion

#649

Post by victor9000 »

kmeisthax wrote: post #568

If Copilot itself is infringing then so is GPT-3, DALL-E 2, NovelAI, and Stable Diffusion. There's no legal argument that would solely target one application of this technology, and you can't build generative AI using current ML tools without relying on a very large corpus of public data. All AI is built on free-riding[0]. While there is no US case law that explicitly says "training AI is fair use", the Second Circui…

Personally, I draw the line at corporations profiting from the derived works. But if companies want charge for tools that make these models easier to interact with, then that seems pretty reasonable.

Re: GitHub Copi­lot inves­ti­ga­tion

#650

Post by Siira »

janef0421 wrote: post #196
19h wrote: post #52

I'd be rather saddened if Copilot was shut down or neutered because of a few vocal few protesting against it. It's been a massive productivity improvement to our senior devs, and I got so used to it that it's an annoyance when Copilot doesn't respond.

The people protesting aren't a "vocal few"; We're the people who made copilot possible. We are frustrated that our work is being used to profit a massive corporation without any compensation and in a way that we at best did not intend to allow and at worst is in direct violation of the terms we set.

Perhaps a good compromise is to make it opt-out, if it’s not already. Though even this is just pandering to the developer’s ego. AI writing code is a massive boost in giving users power and thus freedom. Of course, we need to make AI itself FOSS, but I doubt a legal case could be made for that. A more productive path is to clone the model like SD did with Dall-e.
Post reply on HN