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

GitHub Copilot, with “public code” blocked, emits my copyrighted code

twitter.com

Post reply on HN
Print view View original

691–700 of 806 posts

Re: GitHub Copilot, with “public code” blocked, emits my copyrighted code

#691

Post by Dave3of5 »

CapsAdmin wrote: post #156
yjftsjthsd-h wrote: post #134

Earlier quoted context omitted.

That's an entirely plausible explanation, but it doesn't mean that Microsoft has any less of a legal nightmare on their hands.

I'm not really sure what I think about this. How responsible should Microsoft be for someone's badly licensed code on their platform? If they somehow had the ability to ban projects using stolen snippets of code, I don't think I'd dare to host my hobby projects there. If you can't trust that the code in a project is compatible with the license of the project then the only option I see is that copilot cannot exist. I…

> How responsible should Microsoft be for someone's badly licensed code on their platform?

That's actually a very real problem that mega money has been spent on. The same legal problem appears on sites like YouTube around fair use and copyright. In terms of fair use that doesn't apply here see:

https://softwareengineering.stackexchange.com/questions/1217...

Regardless platforms are partially responsible for the content that their users upload into them. Most try to absolve themselves of this responsibility with their terms of service but legally that's just not possible.

Personally I'm an advocate for fair use but I'm also an advocate for strong copyright laws and their enforcement. In the short time the internet has been available to most people in the world there is a habit of stealing others work and claiming it as your own. Quite often this is for some financial gain.

Re: GitHub Copilot, with “public code” blocked, emits my copyrighted code

#692

Post by rtkwe »

blendergeek wrote: post #426
rtkwe wrote: post #274

Earlier quoted context omitted.

> Left: “Girl with a Pearl Earring, by Johannes Vermeer” by Stable Diffusion Right: Girl with a Pearl Earring by Johannes Vermeer Even that if done by a person as far as I understand it would not constitute a copyright infringement. It's a separate work mimicking Vermeer's original. The closest real world equivalent I can think of is probably the Obama Hope case by AP vs Shepard Fairy but that settled out of court so…

Have you been following the Andy Warhol Prince drawing case? It is current at the SCOTUS so we should see a ruling for the USA sometime in the next year or so. https://en.m.wikipedia.org/wiki/Andy_Warhol_Foundation_for_t...

No hadn't heard of it, I don't follow copyright law extremely closely it tends to make me annoyed. On it's face reading the case summaries and looking at the two pictures, it feels like the act of manually repainting and the color choices should be enough to render it a transformative work. It's one of the fundamental problems with trying to apply copyright to anything other than precise copies, art remixes and recombines all the time, it's fundamental to the process.

Re: GitHub Copilot, with “public code” blocked, emits my copyrighted code

#694

Post by bayindirh »

stale2002 wrote: post #686
bayindirh wrote: post #682

Earlier quoted context omitted.

You're completely right. My premise is not extending the (court tested, honored) license I attach to my code. What I say with the GPL license is clear: If you derive anything from this code base, you're agreeing and obliged to carry this license to the target code base (The logical unit in this case is a function in most cases). So the case is clear. AI is a derivation engine. What you obtain is a derivation of my GP…

> you're agreeing and obliged to carry People are not agreeing though. They are not agreeing, because there is a perfectly reasonable ethical and legal principle called fair use, which society has determined allows people to engage in limited use of other people's IP, no matter what the license says. > Carry the license, or don't use Or, instead of that, people could reasonably use fair use, and ignore the license, a…

I will agree to disagree on your overly broad definition of fair-use which consists of ingesting a whole code base and using its significant parts for another code base with or without derivation while disregarding the attached license to its whole and/or parts.

Thanks for the discussion, and have a nice day.

I may not further comment on this thread from this point.

Re: GitHub Copilot, with “public code” blocked, emits my copyrighted code

#695

Post by sireat »

acoustics wrote: post #6

I’ve noticed that people tend to disapprove of AI trained on their profession’s data, but are usually indifferent or positive about other applications of AI. For example, I know artists who are vehemently against DALL-E, Stable Diffusion, etc. and regard it as stealing, but they view Copilot and GPT-3 as merely useful tools. I also know software devs who are extremely excited about AI art and GPT-3 but are outraged b…

I must be in the minority of programmers that I really really like Copilot but am indifferent about Stable Diffusion/Dall-E/midJourney.

Copilot on Python makes me x5 more productive. I used Copilot in Beta for a year and continue paying for it now.

For example: I can make a command line data wrangling script for a novel data set in a few minutes with a few prompts with full complement of extras (proper argparse parameters with sane defaults, ready to import etc etc). # reasonable comments included for free as well

Before copilot I could do the same in about 20-30minutes but my code would be a mess with little commenting. I would spend 30-60 minutes just looking up docs for various libraries.

Now without Copilot, if all I was doing was writing data wrangling scripts 4 hours a day I could approach this Copilot like productivity for a single task.

However with Copilot I can switch problem domains very quickly and remain productive.

Interestingly, on something like CSS or Javascript - Copilot is helping only slightly, maybe because my local training set is insufficient and my web-dev prompts are too generic.

So I think AI can be fantastic force multiplier in a skillset that you already are reasonably familiarity. I can handle the 5-10% wtf Python code that Copilot produces.

I do not particularly like copyrights and do wish Copilot had been trained on private Microsoft code as well.

Re: GitHub Copilot, with “public code” blocked, emits my copyrighted code

#696

Post by inkeddeveloper »

elcomet wrote: post #689
grecy wrote: post #606

Earlier quoted context omitted.

An equally simple answer is that copilot is pulling code (or at least analyzing) from repositories that are not public.

I think that's very unlikely, they said and repeated that they are not using private code. People catching them lying on this would be very bad for GitHub.

Yet here we are.

Re: GitHub Copilot, with “public code” blocked, emits my copyrighted code

#697

Post by rtkwe »

kevingadd wrote: post #229
jzb wrote: post #214

Earlier quoted context omitted.

I haven’t seen any side by sides that seem like a lift. Any examples? I don’t see Midjourney (et al) as remixes, myself. More like “inspired by.”

Not safe for work, but one example I saw going around: https://twitter.com/ebkim00/status/1579485164442648577 Not sure if this was fed the original image as an input or not. Also seen a couple cases where people explicitly trained a network to imitate an artist's work, like the deceased Kim Jung Gi.

That's clearly lifting style, pose and general location but in each of those there are changes. Even for the original art we could find tons of examples of very similar poses and backgrounds because anime girl in a bathing suit on a beach background isn't that original of an image at the concept level. That pose also is a pretty well worn.

This is the problem of applying the idea of ownership to ideas and expression like art. Art in particular is a very remix and recombination driven field.

Re: GitHub Copilot, with “public code” blocked, emits my copyrighted code

#698

Post by cowtools »

isworenot2 wrote: post #657
LtWorf wrote: post #571

Earlier quoted context omitted.

Humans have creativity. The first C developers wrote C code despite lacking a training set of C code. AI can't do that. It needs C code to write C code. See the difference here?

The training set for C was algol and a bunch of other languages. AI could be used to create languages based on design criteria and constraints like C was, but it does bring up the question of why one of the constraints should be character encodings from human languages if the final generated language would never be used by humans... I mainly think it's funny watching all of these Rand'ian objectivists reusing ever ex…

The problem here is that copilot explots a loophole that allows it to produce derivative works without license. Copilot is not sophisticated enough to structure source code generally- it is overtrained. What is an overtrained neural network but memcpy?

the problem isn't even that this technology will eventually replace programmers: the problem is that it produces parts of the training set VERBATIM, sans copyright.

No, I am pretty optimistic that we will quickly come to a solution when we start using this to void all microsoft/github copyright.

Re: GitHub Copilot, with “public code” blocked, emits my copyrighted code

#700

Post by tablespoon »

throwaway675309 wrote: post #168
tablespoon wrote: post #124

Earlier quoted context omitted.

> You can slow this, you can't stop it whatsoever. It's about as ultimately futile as an effort as trying to stop piracy. ... But STOPPING the use of these tools? Go ahead and try, won't happen. So? No one needs to stop it totally . The world isn't black and white, pushing it to the fringes is almost certainly a sufficient success. Outlawing murder hasn't stopped murder, but no one's given up on enforcing those laws…

You'll never be able to push it to the fringes because there will never be a legal universal agreement even from country to country on where to draw the line. And as computers get more powerful and the models get more efficient it'll become easier and easier to self host and run them on your own dime. There are already one click installers for generative models such as stable diffusion that run on modest hardware fro…

> You'll never be able to push it to the fringes because there will never be a legal universal agreement even from country to country on where to draw the line.

Huh? "Legal universal agreement" has never been required to push something to the fringes in a particular country.

If (in the US) these models were declared to be copyright infringement, or the users were required to pay license feeds to the creators of the data that was used to build the models, they will vanish from the public sphere. GitHub/Microsoft's legal department will pull Copilot down immediately, and development will effectively cease. No US company will sponsor development, and no company will allow in-house use. It will be dead.

Some dude might still run the model in his bedroom in his spare time on his own hardware, but that's what irrelevance looks like.

> And as computers get more powerful and the models get more efficient it'll become easier and easier to self host and run them on your own dime. There are already one click installers for generative models such as stable diffusion that run on modest hardware from a few years back.

If that's the only way you can run something, because it's illegal, you're describing a fringe technology right there.

Post reply on HN