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

We've filed a lawsuit against GitHub Copilot

githubcopilotlitigation.com

Post reply on HN
Print view View original

491–500 of 824 posts

Re: We've filed a lawsuit against GitHub Copilot

#491

Post by devmor »

njharman wrote: post #409
Cort3z wrote: post #270

I’m not a lawyer, but here is why I believe a class action lawsuit is correct; “AI” is just fancy speak for “complex math program”. If I make a program that’s simply given an arbitrary input then, thought math operations, outputs Microsoft copyright code, am I in the clear just because it’s “AI”? I think they would sue the heck out of me if I did that, and I believe the opposite should be true as well. I’m sure my ow…

Say you read a bunch of code, say over years of developer career. What you write is influenced by all that. Will include similar patterns, similar code and identical snippets, knowingly or not. How large does snippet have to be before it's copyright? "x"? "x==1"? "if x==1\n print('x is one')"? [obviously, replace with actual common code like if not found return 404]. Do you want to be vulnerable to copyright litigati…

This is a logical fallacy. A human is not an algorithm. We do not have to extend rights regarding novel invention to an algorithm to protect them for people.

Re: We've filed a lawsuit against GitHub Copilot

#492

Post by devmor »

benlivengood wrote: post #434
Cort3z wrote: post #270

I’m not a lawyer, but here is why I believe a class action lawsuit is correct; “AI” is just fancy speak for “complex math program”. If I make a program that’s simply given an arbitrary input then, thought math operations, outputs Microsoft copyright code, am I in the clear just because it’s “AI”? I think they would sue the heck out of me if I did that, and I believe the opposite should be true as well. I’m sure my ow…

Humans are just compression with extra steps by that logic. There's a fairly simple technical fix for codex/copilot anyway; stick a search engine on the back end and index the training data and don't output things found in the search engine.

I dont think that would work very well because there are not infinite ways to succinctly solve most programming problems. In fact the majority of solutions will look exactly the same.

The real solution is very, very simple. Only use opt-in training data. Don't acquire codebases from people who didn't agree to it.

Re: We've filed a lawsuit against GitHub Copilot

#493

Post by pelagicAustral »

blackbrokkoli wrote: post #469

I am sorry for not bringing any kind of legal perspective here, but: *Jesus Christ*, I hope I live long enough to see copyright die. Here we are at the cusp of a new paradigm of commanding computers to do stuff for us, right at the beginning of the first AI development which actually impresses me. And we are fucking bickering about how we were cheated out of $0.00034 because our repo from 2015 might have been used fo…

I stand behind you on this one. I just hope they fail spectacularly at their attempt to hinder innovation.

Re: We've filed a lawsuit against GitHub Copilot

#494

Post by throwaway675309 »

blackbrokkoli wrote: post #469

I am sorry for not bringing any kind of legal perspective here, but: *Jesus Christ*, I hope I live long enough to see copyright die. Here we are at the cusp of a new paradigm of commanding computers to do stuff for us, right at the beginning of the first AI development which actually impresses me. And we are fucking bickering about how we were cheated out of $0.00034 because our repo from 2015 might have been used fo…

Agreed. My only consolation is that as the technology improves and it becomes easier to train these types of models on modest hardware at home, the detractors of this technology have already lost, but rather than a mercy killing they prefer to bleed out slowly.

Re: We've filed a lawsuit against GitHub Copilot

#495

Post by devmor »

naikrovek wrote: post #387
eropple wrote: post #376

Earlier quoted context omitted.

Host that code. Serve that code to other users. It does not grant the right to create derivative works of that code outside the purview of the code's license. That would be a non-starter in practice; see every repository with GPL code not written by the repository creator. Whether the results of these programs is somehow Not A Derivative Work is the question at hand here, not "sharing". I think (and I hope) that the…

copilot isn't creating derivative works: copilot users are. the human at the keyboard is responsible for what goes into the source code being written. to aid copilot users here, they are creating tools to give users more info about the code they are seeing: https://github.blog/2022-11-01-preview-referencing-public-co...

Your argument is essentially the same as the argument that the pirate bay didn't infringe copyright, it only facilitated infringement.

And we all saw how well that went legally.

Re: We've filed a lawsuit against GitHub Copilot

#496

Post by rolenthedeep »

heavyset_go wrote: post #172
cheriot wrote: post #85

Earlier quoted context omitted.

> that then reproduces near replica's of my code Copying a few lines is not the same as copying the whole thing. Sharing quotes from a book is not copyright infringement.

If your intent is to create a competing product for profit, chances are that won't be found as fair use, given that determining fair use depends on intent and how the content is used. Using clips from a movie in a movie review is probably fair use. Using clips from a movie in knock-off of that movie for profit? Probably not fair use if it's not a parody. Copilot is not like a movie reviewer using clips to review a mo…

Interesting analogy.

Consider every repo on github to be a movie. Copilot is taking individual frames out of every movie on github and composting them into a new film.

I think most of us would agree that individually, each frame is copyrighted. But what if you take one frame from a million different movies and put them in an order that produces a new coherent movie?

The core question we need to settle in court is: does the new movie become its own copyrightable work, or is it plagiarism?

Re: We've filed a lawsuit against GitHub Copilot

#497

Post by chlorion »

bpicolo wrote: post #318
bogwog wrote: post #199

Earlier quoted context omitted.

This is what I hope comes out of the lawsuit. If a company wants to sell an AI model, they need to own all of the training data. It can't be "fair use" to take other peoples' works at zero cost, and use it to build a commercial product without compensation. And maybe models trained on public data should be in the public domain, so that AI research can happen without requiring massive investments to obtain the trainin…

> It can't be "fair use" to take other peoples' works at zero cost, and use it to build a commercial product without compensation. You just described open source software. That's the whole heart of this lawsuit, and equally Copilot. It was trained on OSS which is explicitly licensed for free use.

FOSS license does not mean "do whatever you want". The GPL requires all derived work to also be licensed under a GPL compatible license for example.

Re: We've filed a lawsuit against GitHub Copilot

#498

Post by dmix »

bogwog wrote: post #221
ChrisLTD wrote: post #206

Earlier quoted context omitted.

Multiple cursors wont do what the parent comment is talking about without a lot more work.

Why? You can copy and paste the entire section, and use multiple cursors to add in the brackets. going from a: 'one', to a['one'], just requires you to add two brackets and remove the colon. With multiple cursors you can do that exact same operation for all lines in a few keystrokes.

In my use cases I’ve long moved on from writing the original hash. Having it autocomplete without having to open a file and tab back/forth (or find then copy/paste a block to the other file to temporary work on it) etc.

But what's lost in my over simplified example is the contetxt is usually way more involved. I'm usually passing those as arguments to some function or other unique syntax situation that a glorified find and replace can solve. It's all about doing it in the times you would never think even bother writing a custom command because typing is faster given the unique syntactical context... The only thing faster then is autocomplete.

I'm not actually recreating a new hash with the convienient same format.

Re: We've filed a lawsuit against GitHub Copilot

#499

Post by spiralpolitik »

Swizec wrote: post #26
dmitrygr wrote: post #7

Earlier quoted context omitted.

They demonstrate generated code being identical to some training code.

How many ways are there to write many of the basic algorithms we all use though? Can I copyright "({ item }) => {item.label} "? Because I sure have seen that exact code written, from scratch, in many many places. I guess my question boils down to "What is the smallest copyrightable unit of code?" . Because I'm certain suing a novelist for copyright infringement on a character that says "Hi, how are you?" would be con…

Apple Books slaps an attribution notice on the end if you copy 4 or more words from a book. The Verve got sued by The Rolling Stones for a 4 second sample on 'Bittersweet Symphony'. Post 'Blurred Lines' you can now be sued for copying "the feel" of a song.

Really what it comes down to is do you have enough resources to convince a judge or jury that X is a copy of Y? Doesn't really matter the size of X.

Re: We've filed a lawsuit against GitHub Copilot

#500

Post by mplewis »

Aeolun wrote: post #427
Cort3z wrote: post #364

Earlier quoted context omitted.

People would likely not share any code if they could not trust that their work would be respected, and attributed. So yes, I believe it to be fundamental to open source.

Maybe researchers that are used to hunting for publications and attributions. If I’m sharing my code publicly, it’s because I want it to be _used_.

I use a license which requires attribution. You do not speak for me.
Post reply on HN