Should GitHub be sued for training Copilot on GPL code?
fosspost.org
fosspost.org
On a tangential note, I always find the discussions surrounding FOSS licenses and copyright rather amusing in a sad way. There's a certain kind of entitlement a lot of people feel towards FOSS that they certainly do not express towards proprietary software and I imagine this a great source of the resentment and burn-out FOSS maintainers feel.
IANAL, but isnt the concept of "derived data" pretty standard? You dont need to copy data for it to be infringing. I've tackled derived data clauses regularly when negotiating data contracts at work and there is always verbiage and discussion around it (e.g., are we allowed to publish an average of the purchased data)
Provided you have rights to access a body of statistics, then copyright has nothing to say -- save overreaching national caselaw (!) -- on your derivation of mathematical, technical, or scientific data from that work.
But a contractual clause, in general, doesn't care about copyright; of you've contracted not to derive data from a work then that's orthogonal to copyright.
IANA(IP)L, this is my opinion and unrelated to my employment.
Of course it isn't the same as a human programmer doing anything. It's a complex piece of software, which we happen to misuse the term "AI" to describe, but it is not intelligent.
That is an extremely disingenuous take. It produces novel output, so is not merely an “index” in any sense of the word.
The illusion of creativity is similar to that of technology. Sufficient advanced technology is indistinguishable from magic, and sufficient advanced math is indistinguishable from intelligence. The relation between AI and math is the same as the relation between magic and technology.
Infringement isn't about how the infringing system works, it's about the product of that work.
Exactly this. It makes zero difference that you produced your infringing work with the help of a program that happens to be extremely complex and marketed as "AI".
It is not the same as a human, no, but it's not hard to choose a definition of the word "intelligent" that can accurately describe something that can be done by a program.
When a human walks around a puddle, are they demonstrating intelligence? When a horse avoids stepping in a hole, is the horse intelligent? When a robotic vacuum avoids a stairway, is it intelligent? When a self-driving car avoids a bollard, is that intelligent?
Whether there's a being inside the device that believes it experiences consciousness or not, the same outcome happens. A Searle's Chinese Room that produces copies of Chinese IP, a trained monkey that does so, or a human that does the same thing, the outcome is very similar.
It's just a tool used by the developer; the onus is on the developer to ensure they don't infringe the licenses of the source code they incorporate in their software. Since Copilot makes it impossible to know where it's barfing code up from and what license that code is under, a developer who cares about not getting sued probably needs to avoid using Copilot.
The tool itself lacks agency, it did what it was programmed to do.
If you took the tool's suggestions and proceeded to published a derivative work, you may have infringed.
This really doesn't feel any different from P2P filesharing services. Rightsholders have targeted tool publishers in the past, because they are the largest single target and not anonymous; but ultimately the infringement is performed by the end user.
If he were able to remember a large enough piece of copyrighted code, and reused it, then it still wouldn't be fair use, even if he changed a variable name here or there, or the license message.
But I think the conclusion most people are settling on is that it's definitely infringing.
You know, pretty much all of programming can be summed up as "tedious labor elimination," and this tool directs that same labor elimination at the work of programming itself (I no longer have to constantly google syntax idiosyncrasies etc.), and NOW coders are pissed? I don't get it. Eat your own dog food, people, because this is what it looks and tastes like.
As to the copyright infringement or licensing-violation claims, I have yet to see it autocomplete an entire algorithm correctly, or one copied verbatim from somewhere, although that could be mitigated. You still have to pay attention (kind of like Tesla autopilot), it's not going to eliminate your job.
We're upset because it's quite literally infringing on intellectual property. Infringing on intellectual property that's been set aside for the exclusive use of the commons.
If it infringes everyone equally and everyone equally benefits from the infringement, has a net wrong actually occurred? (which of course begs the "do the ends justify the means" question...)
I don't see how this is any different a form of "infringement" than me copying and pasting snippets of other peoples' code, and then modifying it to suit my particular context, without specific attribution, except that the latter is a much more laborious and time-consuming process than copilot autocomplete, and programming is all about tedium elimination
It’s not done equally though. Copyleft code is extremely likely to be on GitHub somewhere, while internal proprietary code is often not. Copilot will thus have been trained more on the former than the latter.
> I don't see how this is any different a form of "infringement" than me copying and pasting snippets of other peoples' code, and then modifying it to suit my particular context, without specific attribution
It’s no different, but that is also copyright infringement.
so basically all of Stackoverflow is copyright infringement and has been for decades? Find me the programmer who has never either 1) copied and pasted directly from the internet, or 2) taken an idea found on the internet and massaged it for their own purposes. I mean... this is basically why programming is so lucrative IMHO. Everyone is piggybacking off of everyone else's work (at least in open source)
If the code is really sufficiently trivial (and I’d guess that most code samples you’ll find on StackOverflow are) you may have a fair use argument in the US. Generally speaking though (and especially for anything nontrivial) you need to respect the license. CC BY-SA 4.0 is one-way compatible with GPLv3, though, so that helps if you’re including it in a GPLv3 codebase: https://creativecommons.org/2015/10/08/cc-by-sa-4-0-now-one-...
https://en.wikipedia.org/wiki/Threshold_of_originality
Then even if it is copyrightable, under some circumstances your use of it may be considered fair use anyway:
https://en.wikipedia.org/wiki/Fair_use#U.S._fair_use_factors
Or potentially de minimis:
https://blogs.library.unt.edu/copyright/2017/09/05/the-de-mi...
But when in doubt, ask for permission or ask a lawyer.
A typical case mentioned in the training is that code from StackOverflow is (probably) licenses under CC-BY-SA 4.0 and as such it can never be copied inside their proprietary-licensed code base.
If Microsoft included their own source code when training copilot then at least they would be intellectually honest, but they don't. They only consider GPL and other free and open source code to be up for grabs.
copyright was established to increase the innovation and creative will of the arts and sciences, what could increase that creative force more than an AI assistant who has seen every creative work ever made?
Except that is not what is happening here. The problem is that AI is being used to take code, which was provided to the commons under the explicit condition that anything built with it is also released under the same terms, is now being fed to a magic mystery machine to produce code that can supposedly legally be witheld from the comments. The only code that this affects is the one that was already shared - you won't see Microsoft feeding Windows and Office source code into Copilot anytime soon.
The argument would be along the lines of: you as the user are the one who asked the eidetic programmer (nice terminology, @bencollier49) to produce code for your project; all we did is make the programmer available to you.
Does GitHub own the code generated by GitHub Copilot?
GitHub Copilot is a tool, like a compiler or a pen. GitHub does not own the suggestions GitHub Copilot generates. The code you write with GitHub Copilot’s help belongs to you, and you are responsible for it. We recommend that you carefully test, review, and vet the code before pushing it to production, as you would with any code you write that incorporates material you did not independently originate.
Does GitHub Copilot recite code from the training set?
The vast majority of the code that GitHub Copilot suggests has never been seen before. Our latest internal research shows that about 1% of the time, a suggestion may contain some code snippets longer than ~150 characters that matches the training set. Previous research showed that many of these cases happen when GitHub Copilot is unable to glean sufficient context from the code you are writing, or when there is a common, perhaps even universal, solution to the problem.
The reality is that, nowadays, the overwhelming majority of developers touches FOSS code every day and just assumes they're entitled to use it as they see fit. The folks that came up with "copyleft" or care about licenses, are very much not in the driving seat. Blame FAANGs and their hatred for GPL.
But would it be enough to spur the open source movement on its own if you could legally decompile all binaries and redistribute that? Probably not.
Its not like source vs. binary is a clear distinction - between code obfuscation, generated code, transpilation, etc. there is a lot of wiggle room what should or should not be OK.
https://www.inventati.org/frx/essays/softfrdm/whatissource.h... https://b.mtjm.eu/source-code-data-fonts-free-distros.html https://wiki.freedesktop.org/www/Games/Upstream/#source https://compliance.guide/pristine https://opengameart.org/forumtopic/source-required-for-art-l... https://wiki.debian.org/rly-free-software
wait for Microsodlft lawyer to get answer to your original question
The difference is that AI isn't able to grasp concepts, it's only capable of rehashing patterns. If it is able to understand concepts then it should be shut down and researched immediately, because it's either close to gaining consciousness or already has done so.
The core of copilot is a file or a block of memory laying out a bunch of floating points that get processed and turned into code. This arrangement of floats is derived from source code, with licenses and copyright notices.
I don't think it's any different from turning code into a compiled program. Any developer will understand that a compiled version of GPL code is a derived work and subject to the GPL license. Why would a compiler that turns code into floats be any different? Sure, those floats get mixed up with the floats from other source code, but linking to GPL'd code does something very similar and is also covered by the license.
It's possible to consider copilot similar to hashing: a SHA hash of a binary isn't subject to the binary's license, that'd be silly. However, hashes are inherently one-way, and copilot isn't.
A question I'd like to ask Microsoft is "if I steal the Windows source code and train an AI on it, can that AI be freely distributed and used for Wine/ReactOS/etc?" If Microsoft sticks to the stance that AI isn't subject to the licenses on software then a leaked source AI should be fine, but if they want to protect their intellectual property then they will send cease and desist letters to anyone even thinking about using such an AI model for code completion. My expectation is that Microsoft will act against such an AI.
Regardless, the fact that Github did not ask permission or provide an opt out before training started is a huge middle finger to all open source developers. Even if they can get away with this stuff legally, this approach has surely offended many open source developers who want big tech companies to abide by their code licenses. I don't do much open source work myself but I've been offended by the whole process from the day copilot rolled out and I don't believe I'm alone in this.
Personally I'd say the difference is understanding why a certain pattern works rather than blindly inserting whatever works. It's the classic Chinese Room thought experiment.
A human would probably try to defend against a copyright infringement suit over that by arguing something like the following.
There isn't sufficient creative expression in fast inverse square root (FISR) to be copyrightable. There is plenty of creativity in that thing, but it is in things that are not copyrightable such as the underlying mathematics that it is using. Copyright covers expression of ideas, not use of ideas (that's patents) or the ideas themselves.
The expression in FISR that they probably are copying from is pretty much all just in choosing the names of variables, and most implementations I've seen just use pretty normal names that follow normal naming conventions that people use when they aren't putting any thought into naming their variables.
That level of expression is arguably not creative enough to support copyright, at least in the US after Feist Publications, Inc., v. Rural Telephone Service Co., 499 U.S. 340 (1991) [1].
(I'm assuming that the human didn't do anything stupid, like reproduce the comments too).
[1] https://en.wikipedia.org/wiki/Feist_Publications,_Inc.,_v._R....
Algorithms for generic mathematical operations such as the dot product or matrix multiplication are often trivial to deduce, though optimizer vectorized versions perhaps less so. Most helper functions are unoriginal enough that no reasonable copyright law would protect them, which is also the case for (too) many cases of patented code.
The copyright question does ignore the code license question, though. If a complicated algorithm like FISR is not original enough the what protects any boring old operating system code? What stands in the way of publicly hosting Microsoft's leaked sources, as clearly the code is all quite trivial? There is very little in an operating system that other operating system developers haven't thought of or would reasonably have come up with had they been constrained to the same restrictions.
The variable names are one thing, though they could be chosen much more descriptively. However, the system also output the comment "// what the fuck?" which is not only terribly nondescriptive, it's also something that the system couldn't have come up with if it would have learned from code in any practical form.
The suit you linked is about the difference between information and creativity. However, the case surrounds a data set, something simply factual, rather than a composed piece of information such as code or a book. Code listed on Github is not similar to the listings in a phone book. If they were, all software copyright, proprietary or otherwise, goes down the drain. I think that's impractical to say the least.
FISR could have been patented (and be now in the public domain anyway), but only it's specific implementation in DOOM is covered by copyright.
Also, your argument follows a composition fallacy: emergent properties exist, and thus you cannot simply say that because each individual piece of a whole is trivial, the whole is trivial. Heck, software pretty much by definition goes against that. For relevant precedent, there is no shortage of information that becomes classified when in aggregate. Knowing where a certain piece of infrastructure is isn't likely classified, but knowing where all the strategically important pieces of infrastructure are certainly is.
Which is why the question isn't whether the users of Copilot are infringing someone's GPL (they'd likely have a solid defense based on the individual piece not being sufficient to hold copyright protection), it's whether Copilot itself constitutes a derivative work of its input data, which it consumed as whole (copyrighted) works.
This is to avoid any contamination and verbatim copies of code. Once you have read a piece of code there is a risk of "contamination" and you will be influenced by it. It does not matter if you directly copy it, write it out from memory or use an AI to regurgitate it. It will be a copy of the code. To me this is very clear.
My limited experience: my then-employer opted not to acquire a company after doing due diligence. Ultimately we decided that the price of acquisition (both paid out, and also incurred in internal time) was below the cost of building a comparable product ourselves.
As the dev who did the tech portion of the due diligence I was now “tainted” by my knowledge of their system. As a result I could not work directly on the effort to build our own comparable solution.
[0] https://wiki.winehq.org/Developer_FAQ#Who_can.27t_contribute...
Can you please elaborate on this?
Definitely. Many of my acquaintances complaining about Github Copilot without trying it themselves regularly pirate movies, shows and music. They also always cheer if there is some court ruling against Facebook or Google, no matter what the actual case is even about.
> The question is ultimately going to come down to - "Is Copilot the same as a human programmer reading a lot of GPL code and rehashing, in a non-infringing way, the algorithms, functions, and designs used in a lot of FOSS? Or is Copilot performing more of a copy and paste pastiche of code that is protected by intellectual property law?"
It seems to me that the regurgitation only happens if you post the first half of the code, expecting the second half. I imagine that the software sees how several hundred repositories (which are all forks) have a very similar pattern and tells you the best fitting approximation of how they continue, which is again very similar.
In the future I can definitely see Github updating their license and some kind of exodus by FOSSers towards GitLab. But I believe that many open source projects will just put up with it, similar to how Youtubers and Twitch streamers want to stay on the premier platform.
The problem is that, iirc, GPL didn't consider this at all and still uses language focused on copying code, so something like copilot might slip through the cracks of those definitions.
Then again, the license uses this language when it allows usage of the code in the first place, so one could say that either a) this usage is covered by the license, in which case all conditions apply, or b) it is not covered by the license, in which case... github wouldn't be allowed to use the code at all.
To give an analogy: I think feeding code into an AI is essentially analogous to compiling the code. A machine turns it into something more usable and the original human-written content isn't part of the result anymore, but the intellectual property gets dragged through the process nonetheless. Why would it be any different just because the mechanism of transforming the code into executable software gets a bit more complicated through the usage of AI?
It literally can't do it in an "in a non-infringing way" as it wasn't made to do it "in a non-infringing way".
People were able to get copy-pasted code verbatim. It means it does not know whether what it does infringe on the GPL or not.
Let say you find a human that never knew anything about copyright and you show him a bunch of Disney movies and you ask him to make you a movie and he literally copy one of their movie. Does it make it non-infringing? (Funny thing is, even people aware of copyrights does infringe it... so yeah hard to say even a machine could make some non-infringing content).
The solution would be to at least make him aware of copyrights and works with that, but first is it even possible, and seconds, is it even enough...
Sadly nothing will ever be done, at least not until it we feed it Disney movies and it start to affect their bottom lines.
That is probably the way to kickstart a legal discussion about Copilot.
Copilot is completely depending on the legal theory of being effectively exempt from copyright, under fair use doctrine; if that legal theory falls apart, the entire space (and a lot of other machine learning stuff) is utterly doomed.
Will it, won’t it, should it, shouldn’t it? Dunno.
(And when people say that it should just say what license the code it generates is under and what attribution or similar is required: Copilot can’t tell whether it’s reproducing copyrightable chunks of code, or indeed where what it produces came from, by the very nature of machine learning techniques. The whole verbatim reproduction issue demonstrates this—they’re trying to avoid such reproductions, which a cynic might say is because it weakens their fair use claim, but it’s not easy to do.)
[0] https://twitter.com/mitsuhiko/status/1410886329924194309
Ironic that those who generally purport to champion FOSS fail to understand that Free Software was all about defeating copyright. The GPL was meant to turn copyright against itself.
Since I have two free complimentary months I decided to sign up even though I'm not super thrilled with it (see previous comments). I was given two options:
1. allow code from public repositories 2. allow copilot to learn from my code
I disabled both of these options. Presumably I am now using an AI model which learns and suggests based on the context of my project.
I just gave it a test run. I have a function with this code:
if (!card.IsFaceUp && !card.IsBlocked)
{
FlipTableauCard(card);
card.SetIsBlocked(false);
break;
}
// if the card is face up, flip it
if (card.IsFaceUp)
{
FlipTableauCard(card);
card.SetIsBlocked(false);
}
The "allow code from public repositories" doesn't do what you think it does. All it does is add an extra filtering step to avoid producing code found in it's training set verbatim. The model you are using was still trained on those repositories, it's not limited to your project.
But my comment still stands. You can turn off the verbatim copying feature that people keep talking about and the "AI model" will generate code based on your own codebase.
When I'm using it with Unity or a JS project that has NPM modules, does it use those as context to fill in some code as well? No clue.
Was it trained on open source code and is that ethically and legally shady? Yes.
Is it copying verbatim at this point? No.
Does it help me be a better programmer and will I pay for it? No and only if I forget to cancel my trial subscription.
US Copyright law states that fair use and derivative work are not infringing - and said law supersedes licensing.
It says no such thing:
That's not a fundamental statement about all machine learning systems. GPT-2 did a lot more direct regurgitation than GPT-3. GPT-3 tends to be much more transformative, but does still sometimes spit out code / text verbatim.
Copilot and codex spit out close enough to my own code that it's clearly creating a derivative work, at least by my read.
This is untrodden legal ground, but I think that a lot of this comes down to issues of reasonableness. The reason I used the AGPL license was to create a commons. If copilot played within some reasonably friendly way around that commons, I might not feel bad about it.
However:
1) Copilot wants me to pay to use something derived from my own code, where I stuck a license there designed precisely NOT to be in that position.
2) Copilot provides a competitive advantage to proprietary projects who are more likely to be able to afford it, over open-source / community ones. The reason I used an AGPL license was because I thought we needed this type of code to be open and transparent. I work in a domain where transparency is essential (I don't disclose domain, but you can think of transparency in government, education, voting, medical, police, etc.)
3) I have no way to have a conversation with anyone at github / Microsoft. They took my stuff, and they won't talk to me about how they use it. It's automated systems all the way down.
4) The whole Open AI nonprofit -> for-profit transformation is just sleazeball. Given all the talk about ethical use of AI, something like this really leaves a sour taste in my mouth. I don't mind DeepMind, FAIR, etc., since they're honest about their goals. Open AI feels like a Silicon Valley get-rich-quick scheme with a lot of nice marketing copy and legally-questionable tactics.
Jury, judges, and developers are swayed by common sense. People like me can be swayed to testify one way or another based on whether we feel cheated. What Microsoft / github / Open AI did here wasn't very reasonable, friendly, or sensible.
TL;DR: I support the concept of co-pilot in essence. The specifics here feel illegal and sleazy.
I was about to admonish you for phrasing it this way when we all uploaded code willingly to github, giving up certain rights according to the ToS, but then I remembered microsoft straight up bought all of github, so "took my stuff" is pretty accurate. I would be interested to see a diff of the ToS since the purchase.
A lot of code on github (albeit not mine) is uploaded without the original party's agreement. Richard Stallman doesn't use github, but a lot of his GPL-licensed code has been incorporated into projects hosted there. If the terms-of-service allowed github to violate GPL licenses, I think most projects would need to migrate to gitlab. It'd be neigh-impossible for project authors to know that no GPL code in their project came from someone who did not have a side-license to github.
Even if that argument fell apart somehow, their terms-of-service state (https://docs.github.com/en/site-policy/github-terms/github-t...):
This license does not grant GitHub the right to sell Your Content. It also
does not grant GitHub the right to otherwise distribute or use Your Content
outside of our provision of the Service, except that as part of the right to
archive Your Content, GitHub may permit our partners to store and archive Your
Content in public repositories in connection with the GitHub Arctic Code Vault
and GitHub Archive Program.
/joke
Has anyone tried taking source from leaked copies of old MS code and tried to get copilot to reproduce it?
If you learn to write novels by reading other authors, is that a crime? No.
If you reproduce their work, sometimes word by word, yes.
I'm less worried about MS getting sued for this and approaching 100% expecting that users are opening themselves up to legal exposure. I can't see any legal department saying go ahead with using copilot code, but by all means ask.
it depends on what this "bunch" means. It's not clear cut at what granular level does the copyrighted parts become so small, and sources so many, that the new works is considered transformative.
I wrote some code. Released it under the GPL, and my only expectation is that if you use my code in your product you make source available to users (and GPL does require you tell the user how to get that source code). That on small requirement, the one thing that I'm asking you to do if you want to use my code, is not being respected by Copilot. It recommends my code, and obfuscates it, and does not tell the user where it was synthesized from, nor provides a way to get to the original source. From a certain point of view, Codepilot could be seen as a willful infringement machine. It will be interesting to see how this gets sorted out.
I think the interesting legal question here will be, are *you* (the user of the service) infringing, or is *copilot*?
I suspect Copilot's legal team has already worked license terms such that they're passing the buck onto you.
it would seem reasonable that copilot should not be liable for anything that a user instructs it to do.
Copilot is a tool, much like the "copy" command.
If you choose to use what it's suggesting, then the fault is completely yours.
maybe my co-pilot reproduces code verbatim from your GPL'd project because you and a dozen other developers all copied the same solution from stack overflow.
Some open source developers are not allowed by their employers to read source code with a different license for fear of infringement.
If you were to write all those ideas and idioms down and pass them to someone who had not seen the Linux source code, who then used it to reimplement similar functionality, neither of you would probably be guilty of copyright infringement. (there are still patents of course). https://en.m.wikipedia.org/wiki/Clean_room_design
Copyright doesn’t protect programming idioms and concepts. It protects against verbatim copying, more or less.
It all comes down to how you characterize what Copilot does. We will just have to wait for new caselaw or even legislation that accounts for autonomous systems in defining legal wrongdoing.
This response doesn't relate to the example provided by CameronNemo, as it's a different scenario.
At any rate, there is no clean room because copilot has "seen" the literal source. It is not comparable to clean room implementation in any fashion.
the software running the model does not have access to source code
as the parent said, it depends how you characterize it, which is why this will be decided by whoever can afford the best lawyers.
Copilot has reproduced entire functions from existing codebases, which invalidates the idea it doesn't have access to source code.
But if that were the case, one could get away with copying music by merely compressing it - I am not copying the data, I have a totally different set of data that happens to get decoded into a similar performance.
That's roughly analogous to a machine learning model, isn't it? compressing an enormous dataset into a "model" that is capable of being decoded in myriad ways depending on context.
At best, if copilot told you explicitly (by doing the very hard work of identifying the likely sources of the code output ) you could make some (more) informed decisions as to if it's worth the risk to include it.
For this reason, I've worked at places that forbid employees even reading open-source code. If we were having difficulty with an open-soruce component to the point where we needed to look at the code, we'd hire a contractor, explain the problem, and then they'd explain a solution, and all the communications would go through a company lawyer.
In other words we're getting into cheat-detection software territory, which sounds difficult to get right in general.
I think the claim that it's "probable copyright infringement" is nowhere near proven.
GitHub likely gave that option to satisfy user's lawyers who might have a higher threshold for "clean room" implementations or "no open source". Not as any kind of implication of copyright infringement.
Fair use works as an argument to the usability of Copilot.
Derivative works, per US Copyright law, are not infringement, either.
is 'if err != nil {' your original work? Or is it 'commonly accepted knowledge' as a Go programmer?
That's really depends on the country.
For example Japan has a law[0] that's allows usage of any copyrighted materials for machine learning and other data analysis. You can also do it for commercial purposes. There are some limitations (you can't share the dataset itself, but you can share the model), but overall it sounds good.
So I think it is relevant here because there's a gray area around whether or not training a model is like linking to a GPL licensed software(not derivative with caveats) or deriving from one.
By the way, Free and Open-source software licenses are not public domain (or 'public-domain-equivalent'), The copyright holder of the software licenses it to whomever, but the holder still retains their copyright.
This keeps coming up, but if you look at the text of the GPL the word ‘derivative’ literally never appears. GPL in fact explicitly exempts code that is accessed over a web service from needing to be shared, as is the case with copilot.
Of course the code copilot generates may violate GPL, but that’s up to the tool’s wielder to determine. Just as it is when searching for code on the internet, consulting books, recalling past knowledge, etc.
I don’t even use copilot (I had early access and discovered programming languages are better for unambiguous encoding logic than English, go figure). I’m just sick of all these supposed craftsmen blaming their tools rather than holding themselves accountable for what they commit.
To "modify" a work means to copy from or adapt all or part of the work in a fashion requiring copyright permission, other than the making of an exact copy. The resulting work is called a "modified version" of the earlier work or a work "based on" the earlier work.
A "covered work" means either the unmodified Program or a work based on the Program.
If the fair use doctrine fails and the license is relevant, there’s still nothing special about GPL, because almost all licenses would be being violated in some way (most commonly starting with attribution requirements). In this situation, Copilot will certainly be discontinued immediately.
Copilot is thus completely depending on this economical bet.
It for those reasons why we could not write a "Cosinger" or Comusician" that is trained on music found on youtube. It would be sued into oblivion the first time any 2-3 notes could be linked to a specific copyrighted song. If copilot survive long term we might see a similar project trained on creative common music, including CC-NC no-derivs, but music labels might own a few of those and their guns would be quite large.
While technically any snippet of code can claim a copyright, it's sort of ridiculous to suggest the snippets that Copilot generates has any protectable value.
Practically Copilot is saving you the labor of writing fairly trivial but tedious code. I work on open source full time, there's no random snippet of code that I've written I'd feel upset about if someone copied.
Licenses on software are mostly performative.
I seriously doubt you actually believe this, at least in an equal way. The code to Windows XP is available on the 'net, but I can't just go and compile that and start giving out copies without serious legal repercussions.
Copilot is "copyright for me but not for thee" and it's bullshit.
It generates what a intern code monkey would generate after reading a few stack overflow posts and a few github repos.
Again, its copyright for them, but no protections for us.
I mean, it might. That's the whole concern. It's scraping random bits of code from all over the place.
1 month ago, HN front page https://news.ycombinator.com/item?id=31458635
"Q: Why is nothing from the Windows XP source code leak added?
A: Even though Microsoft has only taken down a few Windows modifications, they will most definitely take Windows XP Delta Edition down if there is a reference to the source code inside it. The Windows XP source code is illegal to download, fork, and redistribute, so nothing from it will ever be added."
But let's think of an example that may push the line. What would happen if someone wrote a closed source "Linux" kernel using the Linux kernel interface as a stub, and filled in the code using copilot. You'd expect some of the generated code to come from Linux, since that's the best code for the stub. Linux's use of GPL is not performative and this could create multiple instances of copied code.
But for everyone else who isn't building a closed source replacement of GPL software, I have a hard time thinking you'd be impacted.
Academic papers as a whole can include both "white papers" and "peer-reviewed journal articles" aka the "papers" that many think of. I would surely put a "white paper" onto my resume if I thought it was a great piece of work.
Something like the original BTC paper is a good example. It wasn't published in a conference or journal, but the level of rigor and the scale of the contribution is similar to what would be published in a conference or journal.
I think the crypto community was the end of it. Now "we are launching $CoolCoin" gets called "whitepaper."
I would still call my paper a whitepaper, though. While it's not rigorous, it's more about law where rigor doesn't apply in the same way. And it is not peer-reviewed, although it was reviewed and rejected by the FSF.
unless microsoft is doing work-for-hire for you via copilot, i highly doubt they are liable.
You, as the person who is claiming to have produced the work (even though you were using a smart tool to help), must be the person who also is liable. Otherwise, could you not claim that the auto-correct on your word-processor is liable for copyright infringement?
Running a copyrighted movie through a neural network compression algorithm and uploading it on bittorrent isn't going to stop you from being sued. Even if the output is produced by an AI.
If copilot allows you to type
// source code of linux kernel
And you get the whole code, then I would consider it unoriginal
The same way your movie example, if you told it
Avengers Endgame
And it gave you the whole movie, it would also be. But what if you type (like with DallE) Spiderman fighting Thanos and you get something different, but that resembles some Endgame scene. Would that infringe copyright, be fair use, or what?
For example
If we go back to the very basics, the purpose of copyright is to reward and incentivize content creators.
The purpose of fair use is to allow certain uses of a work that are beneficial to society, and don't substantially interfere with rewarding content creators. Consider for instance a parody. The parody does not substitute for the original work, so it will not interfere with the reward.
Looking at copilot through this framing, we can ask whether it will interfere with rewarding creators. If Alice writes a program, will copilot trained on that product allow Bob to create a competing product more easy, essentially freeload on Alice's creativity? Yes, to some extent it will.
On the other hand there are also advantages of allowing copilot. It promises to make software creation easier, for Bob, but also for Alice.
Now, as this is trained on code that is posted publicly but not private code it will advantage private code over publicly posted code. This has the potential to harm the free and open source software movement. This movement is a positive in the world, so harming it should be considered a bad thing.
But really, the big question is how big these effects are, and no one seems to have a good answer to that.
If it is nothing but a glorified autocomplete that inserts SHORT snippets of code, as in a quote from a book in a book review, or the above quote, that's fine. But when it is inserting whole pages, that's beyond Fair Use, and basically a nicely scaled and laundered version of plagiarism.
Another key is now much it is modifying the code to suit the situation. Is it generating entirely new synthesized output like DALL-E 2 or GPT-3 such that the output is closer to generative creativity, or is it merely pasting in code blocks found in similar situations?
It comes down to the question of how transformative is the output, which is a key concept in copyright law.
>>“What is the difference between this and someone doing it manually? "
Another key question here. If it is merely cut/paste beyond a line or two, it's plagiarism, but if it is synthesizing new works, it's good. Same as manually: am I using your GPL work for inspiration to generate new works, or am I copy-pasting pages of code?
Does anyone have any extensive experience with Copilot to be able to highlight these differences?
EDIT: fmt, clarity
Also I think it's important to note that Copilot isn't an independent AI; it's a human in the loop system. In my experience I always make adjustments to the code it generates to better fit my needs. So on top of the transformation that copilot does, I'm also stacking transformation on top of that. So the final, potentially copyrightable product is very far removed from the training data.
Copilot's training algorithm, OpenAI Codex, "is a descendant of GPT-3" under the hood. I think code is unique in that unlike text and images, there's much less variance in code. It's not as expressive as English, or visual art. So I think there might be a higher chance it'll generate something similar to its training data--but only because all code is inherently more similar.
Good points about you adding transformations on top of the Copilot output and less variation in code due to it's structured nature.
Interesting also that it responds to variable names and patterns in your writing/code. Does it also respond to comments?
Your answer points me more towards the not-infringing argument. Perhaps the best solution would be to have a companion plagiarism-checker tool that examined your code vs it's training set of GPL/MIT licensed code when you are nearly finished to flag significant copying. Shouldn't be too hard and would avoid the whole problem (and also maybe sometimes point you to a library you should be using instead of rolling your own).
If I understand your question correctly, it does respond to comments. One of the best ways to interact with copilot is to write a comment (eg "Read input until y/n is entered") and have it generate the resulting code. If you're asking if it matches commenting style, I think it does, but I haven't pushed it in that direction too much.
I think that's an interesting proposal, although it does place the onus on the developer. I would be curious to know how often developers would fail that check even without Copilot!
Overall I think there are definitely open questions, but I'm personally really excited by systems like GPT, Copilot, or Dallee. The future is not clear, but I think these tools to some extent make the internet make sense. There's way too much data online to make sense of--be it code, text, or images. Unlike a search engine which just links to hopefully relevant material, these tools "learn" from all that data and respond with an answer of sorts--not a list of references. I think one huge improvement would be making these systems more explainable. So getting a certain response, you can also see the thousands of references that were used to generate that response. That would help a lot in providing transparency in whether there is plagiarism happening, and also just be an immensely useful tool for humanity and the internet. It feels like the next logical step for the internet. I would even say it feels like the internet was built for this!
And a good point you make about not relying on the programmers to run the plagiarism checks - the tool should do it itself.
It sounds like this is indeed closer to an copyright-OK generating of new code, rather than mere laundering, and if it isn't yet there, it seems like the copy/paste paradigm would be a juvenile phase, and it should improve and get more "creative" and less copy/paste-ish with further development.
Old comment for documentation's sake:
Are licenses even enforceable by law? The idea of writing some mundane basic code and then wanting to sue someone for "stealing it" just sounds ludicrous to me. True copyright has barriers to make sure you actually invented what you're trying to patent.
That's not to say that there isn't a problem here -- there's definitely an ethical component to how this product works, but this whole code licensing thing never clicked for me. Does it hold any actual power?
Also Copyright isn't patent.
The popular open source licenses explicitly give permission for people to resell your work, it's not even buried in the small print or anything. e.g.
GPL: "Our General Public Licenses are designed to make sure that you have the freedom to distribute copies of free software (and charge for them if you wish)".
MIT License: "including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software".
Apache License: "each Contributor hereby grants to You a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable (except as stated in this section) patent license to make, have made, use, offer to sell, sell, import, and otherwise transfer the Work".
This should no more cause burnout than someone buying your old car and using it for an endurance race makes you tired. Where there's burnout involved it's more likely the demands for support and fixes that head back upstream without any associated money. More concerning is Copilot trained on code which isn't GPL licensed or similar. Sharing code doesn't automatically grant anyone any license to use it for anything at all.
> The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
I'm not confident Copilot will comply with this part of my code's license.
Here's a discussion about that usage in the MIT license: https://opensource.stackexchange.com/a/2188
(That a tool exists doesn't free one from responsibility for using it; CoPilot not citing original code and its license terms seems like it rules it out for anything beyond experimentation; that part is very arguable)
Yes but if I use copilot in a private codebase how sure am I that it has not copied GPL code.
So yes, using a substantial portion of GPL code in your proprietary software product is copyright infringement. Or even using a substantial portion of GPL code and not licensing the code under an appropriate license - https://www.gnu.org/licenses/gpl-faq.html#WhatDoesCompatMean
If I were to turn around and resell it, then you could sue me; but that wouldn’t be a chat bot’s fault.
Just like it’s not the clipboard’s responsibility to ensure I’m not violating licenses, I don’t think it’s copilot’s responsibility either. Use it at your own risk.
That is definitely not how US copyright law works (although you are welcome to argue that it should), that there is such a bright line around selling. It's possible for something to be a copyright violation even if you give it away for free (see torrent sites!), and it's possible for to qualify as fair use and not be a violation even if you sell it.
Under USA copyright law (others are similar but not quite the same), the first step would be deciding if the output of this chatbot counts as a "copy" or "derivative work" at all. If it does not, then there is no copyright violation whether you sell it or not. If it does, then it is a copyright violation (whether you sell it or not) unless it's use can count as "fair use". Whether the use is "commercial" is just part of one of four factors that are balanced to determine if it's fair use. For instance, if you are only using a tiny portion of the copyrighted work and it doesn't have much effect on the profits of the original copyright holder and the use is considered highly "transformative" too (sound like copilot?) -- it could well be fair use even if you are selling your output.
(Also... Github Copilot is literally selling it, right? You have to pay them to use Copilot! It would accordingly be considered a commercial use. If I make a copy of a hollywood movie and sell it to you, I'm probably violating copyright (unless I can convince a court it's fair use), regardless of what you do with it, it doesn't matter if you re-sell it or not. If you re-sell it or make another copy, you may be additionally violating copyright another time yourself.)
I do think there's a reasonable argument that copilot is fair use. It wouldn't mainly hinge on whether the output is sold or not. This is presumably the argument MS/Github would make if brought into court. Since Oracle v. Google, I've stopped trying to predict what courts will do on software-related copyright cases, the law seems to be pretty chaotic and the actions of courts unpredictable. (also i'm not a lawyer this is not legal advice).
In general, I am in favor of an expansive bounds of fair use, and think it serves "the people" to have such.
A FOSS license doesn't mean it's all fucking freebies all the way down as any big tech company is quick to remind you (by attaching trademarks all over the place, for example), but if it's a big tech company taking your stuff and running with it, it's suddenly all fair game.
All in all, it's how companies have been behaving since about forever. Fuck little people all the way.
I have not heard of what people have already been discussing about copyright legal issues around corpuses of Other People's Content, I am curious to read more if anyone has a link.
While I understand that in this case it seems like (or is!) a Big Corporation taking advantage of the Little People -- I would urge extreme caution in advocating for reducing and limiting "fair use" rights because in this case it will hurt the Big Corporation. It used to be clear to everyone that fair use helped the Little People against the Big Corporation copyright holders and should be encouraged. Lately, people (especially software devs, who of course write IP) have been excited about strenghtening protections for copyright in order to somehow limit the Big Corporations (see all the stuff around Amazon and open source), but this is a dangerous game. Fair use is one the only tools we have protecting us from a dystopia where you can't open your mouth or type on your keyboard without paying someone a license fee, which is what OTHER Big Corporations would love to see.
However I don’t quote verbatim which is what this tool has been doing.
Plenty of trigrams exist in many novels that are exactly the same, and i bet that there's plenty of n-grams in programming which would be construed as copies of each other.
The nature of using this tool could hang you in theory more than if you didn’t use it.
In several countries, humming a song - reciting a melody - is cut and dry de jure copyright infringement. It's just not rigorously punished.
Looking at the fact pattern in front of me, I can't imagine Copilot has gotten this far without Microsoft's lawyers being fairly confident that they have a compelling legal argument in their back pocket that not only protects their substantial investment in this project, but perhaps puts up a barrier on the limits of what can be claimed via a copyleft license.
This whole exercise screams "designed for win-win" to me. What that argument might be, I cannot guess.
If anyone shows an example of co-pilot automatically producing something too closely resembling copyrighted code that's too significant to be fair use, you got a problem regardless of what technology got you there.
If that lazy programmer "toils" away behind a corporate wall (sadly too likely), you might never know what he's up to. But you've put his "work" on display and let anyone on the internet record his behavior with this thing.
At some point, co-pilot is going to be able to "suggest" something that's traceable and someone's going to sue. I think it will come down to whether what gets produced is copyrightable or fair use, not whether it was an AI or not that scraped and copied it.
Models like GPT-J or DALL-E mini could not legally exist anymore. So sure, please, make this illegal, regress the whole field of machine learning 20 years back, and make it something that only billion dollar corporations can do.
We should be striving to make copyright less draconian, not more.
> We should be striving to make copyright less draconian, not more. Agreed, and I'm not sure why you think forcing GPT and Copilot to respect open licenses will make them illegal instead of more open.
I wasn't talking about Copilot. I was talking about the vast majority of other interesting models. It wouldn't make Copilot itself illegal because it was trained on explicitly licensed data (so it'd make it GPL-licensed), but it would make those other models illegal.
Take for example the GPT-J, which was trained on 825GB of data scraped off the Internet. If we assume the view that a machine learning model is a derivative work of its training data then that makes GPT-J illegal, because it was trained on a bunch of "all rights reserved" data, and there's no legal license under which it could be released. Most interesting models are like that.
> so Copilot must not charge users
That's not what the GPL says.
I was with you until then. Charging for GPL code is perfectly within the licence as long as you make the source available.
Regarding the first part, I certainly wouldn't argue that training Copilot is copyright infringement. However, the code it spits out in its current state can in some situations be infringement.
Copyright infringement doesn't happen when you read War & Peace, it's when you take that and reproduce it verbatim or very close to it.
So to that point, your doom & gloom absolutist scenario could not play out if the product of the model was sufficiently different.
We saw Google sued by Oracle for infringement over copying APIs. Now imagine you're a not-Google sized company, are you going to take the chance that Copilot will spit out something that they consider copyrighted?
I think in terms of legal/business risk, it's just too high as it stands now.
It would absolutely play out, because it is impossible to guarantee that such a model will always produce something "sufficiently different". These models are black boxes with billions of parameters. That's how they work. It's just as unrealistic as those politicians pushing through "lawful access to encrypted data" that'd effectively make strong end-to-end encryption illegal. There's no middle ground here. We either accept that such a model might sometimes output a snippet from its training data and benefit from the 99% of times it doesn't, or we can be copyright maximalists and ensure no one benefits. (Except ironically huge corporations like Microsoft, either because they can license the data for training, or because they have sufficiently well funded legal departments.)
> Now imagine you're a not-Google sized company, are you going to take the chance that Copilot will spit out something that they consider copyrighted? > > I think in terms of legal/business risk, it's just too high as it stands now.
This is a fair point, but you can say this about any other interesting model. Is that piece of text generated by GPT-NeoX-20B (which is a fully free and open model trained by essentially hobbyists) illegal to use because it might infringe on someone else's copyright? You don't know. And it was also trained on code from Github. Where are the posts calling for people to sue their authors because they're not respecting the GPL?
Here, I've just tried it and screenshoted it for you, spitting out GPL'd code: https://i.imgur.com/2T4uSJR.png
Again, this is not the Copilot. This is the free GPT-NeoX-20B model that anyone can download. The model's not under GPL, and yet it clearly "contains" GPL'd code. Anything which affects Copilot's legal status will also affect GPT-NeoX-20B, but even more severely since GPT-NeoX was also trained on a ton of "all rights reserved" data. So when you raise your pitchfork at Copilot you should also ask yourself the question - are you fine with also killing projects such as GPT-NeoX, or maybe a more lax copyright law is more beneficial to the society as a whole when it comes to machine learning?
Again, you misunderstand. Copyright violations don't occur just by using the system, and as I noted, it's not a problem to use ML models. The actual output content is what matters.
In your example, yes, absolutely, if you used some editor feature that spit that code out verbatim, sans license adherence, you will be violating copyright. The issue isn't specific to copilot beyond the fact that Github has created and offered this model trained on source code and begun selling it as a product.
TL;DR: Is it illegal to use it? Of course not. Is it going to get you sued for blindly taking its output, packaging it and selling it? Possibly. If I were trying to manage the risk to my business, I certainly wouldn't be allowing developers to use it.
Personally, I'm a bit bothered by this myself, but I'd be lying if I said I never once got any ideas by looking at the source code of a GPL project.
Copyright protects the specific, tangible expression, not the idea or concept itself. So in your case, having "got ... ideas by looking at the source code of a GPL project" doesn't necessarily mean there is any violation of copyright. Employers will often require that employees haven't seen sensitive code at all (clean room) as it avoids any possibility of copying the code itself and provides powerful evidence if they're ever sued, but that's not a legal requirement, just a cautious practice. To read code, learn a concept from it and then apply that concept in your own code is fine (as far as copyright is concerned) as long as you're not copying the code itself.
I am not a lawyer and the above is not legal advice, ofc :-)
When I look at the discussion here, it looks like most of the arguments are based on common sense or morality. While that is nice, I would have loved to hear a perspective on this based on law instead. Is there someone here with a more legal perspective/background that could comment?
In practice, you usually need to demonstrate that you were harmed by an action to bring it to court.
So we would need the author of a GPL'd repository to sue, with proof that copilot copied and pasted their code.
The moment that happens, MS would simply omit that single repository from the copilot dataset, tell the judge that they fixed a bug in their "content filtering algorithm", and have the case dismissed. They might also ban the developer's account for good measure.
I think this is true for copyright infringement damages. But is a GPL violation a copyright or a contract issue? And unlike copyright, the fix isn't damages to compensate the rightsholder, but to uphold the license.
> So we would need the author of a GPL'd repository to sue, with proof that copilot copied and pasted their code.
What happens if multiple authors wait and see, and then jointly sue? At that point, removing a single repo isn't enough.
That's not true. Courts won't consistently mandate that people relicense their code that includes GPLed stuff. They'll make you pay a fine and remove the code.
There are reasons why one opensource project(forgot the name probably redox os) bans people from contributing if they have read pirated source code?
Accidentally including the same material in something new is not immediately infringement.
Anyone who thinks Copilot is writing code with any comparable degree of novelty to Stephen King's prose clearly hasn't actually tried Copilot.
No, you don't. You don't do that for kids, not for animals, not for corporations, not for.... . We don't even do that for patents vs copyright vs ....
Altough, I realize that it isn't feasible.
Altough google suing for youtube videos is somewhat different to disney with SW as google doesnt have an exclusive license on those.
But yeah, you changed my kind on that one.
Don't get me wrong, it's problematic when the model spits out moderately sized, copyrightable, chunks, verbatim.
But I think the bigger issue for the free and open source communities is that Copilot is proprietary, not that it exists. We should create a FOSS alternative. Presumably we can even train it on (source available) proprietary code.
Separate from this, the violations are per instance, not the model itself. If I'm an author and write hundreds of non-infringing books over two decades and then suddenly write and sell a copy of an existing work, I'm not going to be sued for my other books or the reading I did to create any of it. I will be sued for the actual infringing book.
https://docs.github.com/en/site-policy/github-terms/github-t...
>This license does not grant GitHub the right to sell Your Content. It also does not grant GitHub the right to otherwise distribute or use Your Content outside of our provision of the Service, except that as part of the right to archive Your Content, GitHub may permit our partners to store and archive Your Content in public repositories in connection with the GitHub Arctic Code Vault and GitHub Archive Program.
Copilot is not part of GitHub's services or archival efforts. At least, it was not when they were training the model. And the TOS don't mention Copilot at all (or the terms "train" or "model").
Good paper, though I disagree with many points.
However, two things: you claim that Copilot might have infringed during training because of the GPL, but the GPL's clause only kicks in on distribution. If GitHub had trained Copilot and then did nothing else, that clause would not kick in.
Second, your Stack Overflow example of copying is not good because SO has an explicit license for all material given. You must agree to let SO and others have the material under that license, which is CC-BY-SA 4.0.
Don't get me wrong; I hate Copilot. But those two things you said were wrong.
I'm too lazy to check the GPL comment (I'll assume I made a mistake). But as far as I can tell my only reference to Stack Overflow was not about liability based on copying from SO. I was making a comment about a common industry practice.
> [GitHub Copilot] has been trained on natural language text and source code from publicly available sources, including code in public repositories on GitHub.
Like technically it's legal for multi billion dollar companies to turn MIT projects into hosted services.
They can do this without compensating the original devs, they can close source their additions, etc.
Just feels icky.
Co Pilot also feels icky, back when it was free, it looked like a neat project. For 10$ a month, one would hope the Microsoft overlords would offer some compensation.
Let's say you write a code snippet that Copilot copies a few million times. Maybe Microsoft can give you a bit of cash.
Then again, it's probably buried in GitHub's TOS that they can use tools like CoPilot to extract from anything you upload.
Edit: I see the possible confusion point here:
> would assume the copyright any liability attached
Should read:
> would assume the copyright and liability attached
Don't want people writing code similar to yours? Don't publish anything. Force people to sign contracts before having contact with your source code.
When I use GitHub I get some handy functions and sort of the style it is written in… but that’s it.
It’s not a lot of code.
Does someone own?
formatDateISOXXX() { a handful lines of code here }
2. It's in the user agreement
3. You weren't harmed in any practical way
4. It's not a real tool and nobody's using it in the long run
5. If GPL fans want to waste money on lawyers, focus on making the GPL have teeth
6. No, of course you don't get to impose a new license on your work retroactively
TiVo says it doesn't.
.
> Vendors are routinely sued over infringement.
The GPL has never one time been successfully prosecuted. Never. Show me a single specific case.
https://docs.github.com/en/site-policy/github-terms/github-t...
> This license does not grant GitHub the right to sell Your Content.
Charging $10/month seems like selling it. The only questions is whether it counts as your copyrighted content, if it's been chewed up by an AI.
if the resulting code is licensed under a GPL incompatible license: yes
if the resulting code is licensed under the GPL: no
If code is GPL, it's free to read, and learn from. That's kind of the point.
It also shows in global competition. Here in germany I have seen scientists trying to train a gdpr conforming data set....
Lets just say that you obviously get much better resuls, the bigger your training set.
Remember Bill Gates' open letter to hobbyists in 1976, in which he blamed pirates for ruining everything? This discussion is much more interesting and subtle than the old debate about piracy's place (and it definitely has a place) in user communities, but I find it humorous that Gates's behemoth, all these years later, seems to have circled around to doing the same thing.
As one who started out by going to swap meets with my Commie 64 and completely disregarding the hand-wringing, moustache-twirling frets of the people RMS (and I) would later refer to as Hoarders, I am pretty sure that I land in the Don't Care category of this debate. Specifically, I think the most compelling point made in the article is when it points out that the tools which were used to create CoPilot are available for free to all, so anyone who is willing to put in the effort of learning how to train models is free to recreate CoPilot for themselves, and share it with anyone.
We all have copypasted code from github or SO at some point without thinking about the license, I am pretty sure, in many cases while at work or otherwise engaged in commercial software activities. CoPilot cannot write entire programs (yet), all it can really do at this point is automate the searching and CTRL-C+V part for snippets and functions, and that is a net good for anyone who writes software.
If there is going to be a lawsuit, I would hope that it hinges strictly on whether github has the right to charge money for it, because essentially, what they are doing by increasing the speed at which people can work is lowering the number of coders required to finish a project on a given timeline. This is both a boon to anyone who is working on something they intend to release FOSS, and simultaneously a device by which greedy Hoarders who want to pay as few people as possible while profiting from the work of those they do pay can increase their profits immensely while creating fewer good jobs for humans.
To put it more simply, when Hoarders use this tech, it is a transfer of wealth out of the pockets of working programmers, into the pockets of shareholders in whatever corporation is reselling this technology. There is good reason, in other words, for working programmers to dislike this, and for FOSS volunteers to like it.
The lawsuit I would like to see happen, if they must, is one which simply seeks to prevent anyone from charging for this, as that would at least keep the investor class from further draining the economy of people who actually work.
Given that the tools for training models are freely available, what I would prefer is if someone with the time implements some form of free version of the same thing. The data they trained the model on for Copilot is freely available to anyone with an internet connection, as are the ML tools. It's extremely clear to everyone that the code copypasta'd by the model does not belong to Microsoft/Github, so if they can sell it, nobody can argue that a free program can't automatically search and copy it for you for free.
There are some that this solution, which I am pretty sure will come about, would not satisfy, and those are the people who can rightly be said to be against innovation and new technologies. It hinges on whether money is changing hands, and whose hands the money is passing into.
Back in those Commie 64 days, we all pirated freely with zero consequence, with one exception that I can recall: one dude in my city started selling pirated software at $5 per floppy. As it happens, he was a friend of my uncle's and I knew him personally. He was a poor man with very little going for him in life, suffice to say, but nonetheless, it only took a couple of months until the police were at his door, and they seized his 64 and all his software.
I disagreed with his actions but I never hated the guy, he was too sad and pathetic to hate if you saw the life he actually lived, but many, many people did, including all the local pirates, because we were firmly of the mindset that software was meant to be shared freely. We may have been wrong about that, but everyone, pirates and non-pirates alike, agreed that the dude taking money was 100% wrong.
I think the same dynamic is taking place here. You will never convince me that copying code from github or SO is wrong, but I agree that CoPilot should not be a commercial thing.
I am a human programmer. I work a lot with FOSS. I learn a lot from FOSS.
I go on StackOverflow and answer a bunch of questions.
The answers I give are based on all the shit I learned working with FOSS.
Is SO infringing copyright by publishing my answers?
I would think that "NO" would be the obvious answer.
Copilot is no different. There is no way this would hold up in court.
We like to imagine that's what it's doing, because the utility of what it actually does can have a similar end result.
What Copilot actually does is read every line of code from a project, and keep a dataset built around what code tends to have been written adjacent to what other code.
Copilot's output is literally a translation of the code it was "trained on", using data about that code to do its translations.
No it reads every line of code from EVERY project.
It is not reproducing code from individual coders, it is reproducing patterns deduced from the work of millions of coders.
It's like Smart Compose in Gmail. Copyright on anything I write is mine by default, but Gmail can of course take all things written by everybody and then use it to train a thing that detects what I probably mean and offer sentence auto completion in Gmails.
// fast inverse square root
float Q_
I can get access to the same code if I search “// fast inverse square root” in google:
https://www.google.com.au/search?q=%22%2F%2F+fast+inverse+sq...
The float Q_ is a partial code match so obviously that requires a more specialised search engine such as copilot.
But all this example demonstrates is that copilot is a great code search engine allowing me to easily find code that is published publicly on the internet
> It is not reproducing code from individual coders, it is reproducing patterns deduced from the work of millions of coders.
But now you say it is a search engine for finding code snippets from the internet. You really fail to see how this is a problem?
Sticking with your analogy:
> It's like Smart Compose in Gmail. Copyright on anything I write is mine by default, but Gmail can of course take all things written by everybody and then use it to train a thing that detects what I probably mean and offer sentence auto completion in Gmails.
How would you feel if I could reproduce your entire personal email verbatim just by starting it with "Hi <unique mother's name>" and letting Gmail complete it for me? How would that hold up in your analogy?
Can’t it be both? If you already know a specific piece of code well enough that you could find it in a search engine, then it reproduces that code.
If not it’s reproducing general code patterns. I don’t see why they can’t both be valid use cases and I don’t see how the latter use case infringes the rights of the Quake coders any more that search engines that already exist.
> How would you feel if I could reproduce your entire personal email verbatim just by starting it with "Hi <unique mother's name>" and letting Gmail complete it for me? How would that hold up in your analogy?
I think to match what’s happening in the quake example provided I would have had to publish a specific conversation online, and then you would have to start your text quoting my conversation exactly such that it couldn’t match any other published conversation, and then it would provide it as an auto complete.
This wouldn’t infringe on my rights any more than if you were to copy my conversation from the internet into an email and send it.
It would also be about as useful to you as the above Quake example which is not very.