This is my favorite question about Copilot ever.
OpenAI Codex (which copilot grew out of IIRC), Amazon and Salesforce versions of Copilot exist. Huggingface Bloom was trained on a sizeable amount of public code. Tab9, now behind, was one of the earliest to combine public code repositories with Deep learning for smarter autocomplete. The data requirements for Transformer scaling mean any and all public facing repositories will be assimilated, whether Github, Gitlab, Stackoverflow or so on.
Wish more energy was spent on how to fund pretrained models that will also run efficiently on CPUs, fine-tuneable to one's language and local environment. Removing reliance on cloud services.
Curious about people's opinions on Dall-E 2 or Google Image-gen, which parallel pretty much the same thing with Renders, Illustrations and Paintings, or upcoming models doing the same for voice acting and music. Coders seem more excited about the potential of those tools.
CoPilot is being offered for widespread commercial use, so it's held to a higher standard. Respecting copyright is much more important when you're building a business and not just sharing fun AI art on social media.
Because it probably mines one bitcoin block every time you click on something. No sane codebase could possibly be so abysmally slow.
This may be the dumbest move from M$ that I have read on this thread! Sure, companies need to protect their private IP, but this really feels like creating unnecessary friction for no good reason...
That's definitely something that most large corporations do.
I also want to know why people think their code is so special that no one else could have ever come up with it independently. Each and every opponent of Copilot is the best developer ever, I guess?
That said, I don't understand the choice to use GPL for any reason, so maybe I'm not equipped to understand the arguments against Copilot. Forcing your code to be open forever isn't freedom, it's the omission of freedom. Someone using your (for example) MIT-licensed code in a closed-source commercial software project doesn't "un-free" the code you released; your code is still exactly as open and as available as it was before, and zero freedoms were lost by anyone.
Please feel free to use my code in any way that its license permits: attribution for the permissive licenses, share-and-share-alike for the copyleft licenses. Those license terms are the price of the code, no different from a proprietary product's "this costs $x" or "this costs $x/month". I'm happy to give away most of what I work on every day, and I ask that people 1) give credit, and 2) in some cases, share under the same terms, and 3) in many cases, don't sue me or other users of code I've written over software patents (which shouldn't exist).
If the day comes that copyright goes away, and we can freely copy and share the code of any currently proprietary software and other works, I'd celebrate that. Until then, I don't want an asymmetric situation in which proprietary licenses must be adhered to but Open Source licenses are ignored.
a) Those who do have the source code to share it. Sometimes the source code is available but it can still not be freely used and/or shared.
b) Allow modifications and redistribution of the binary artefacts, which for is sufficient for many goals.
c) Remove all concerns with reverse engineering and allow e.g. decompiling programs and sharing that source.
Also, remember that the GPL already does not make source available externally if the modfications are only used internally.
Nobody has claimed that they want this. People just want derived work to adhere to the license they chose for their project.
>I also want to know why people think their code is so special that no one else could have ever come up with it independently. Each and every opponent of Copilot is the best developer ever, I guess?
Would you feel the same way about ripping off game assets, or music?
I think you just have an axe to grind with free software in general based on your messages and the general tone. Just because you don't understand it doesn't mean that the ideas are invalid.
I am also curious why copyright laws should protect proprietary software, music, games, writing, etc but not apply to my software, even if it isn't the highest quality work?
This has some interesting implications – for example, it means I can't mirror somebody else's (open source) code on GitHub without their explicit agreement.
So any code uploaded by someone other than the copyright holder renders someone liable to be sued for copyright infringement, AFAICS. The only question is whom it makes liable -- the uploader, GitHub (=Microsoft!), or both?
I can see arguments either way: The uploader is clearly infringing by giving away a right that isn't theirs to give. But so is GitHub / Microsoft, for using a "right" they haven't been properly given. So I'm provisionally leaning towards "both".
> I can't mirror somebody else's (open source) code on GitHub without their explicit agreement.
Who is doing the "mirroring" -- you, in uploading the code, or GitHub / Microsoft in actually hosting it, keeping it available for download from their "mirror"[1] site?
___
[1]: Is that even the correct terminology nowadays, when AIUI for lots of projects GitHub is their primary code repository?
Also, there's no way for anyone to know what portion of code that I commit was hand written vs. generated, so you kind of have to treat it all as written by the committer anyway.
Though this does bring up interesting questions about what happens with things like automated PRs that fix bugs / update dependencies... are those then non-copyrightable? ¯\_(ツ)_/¯
Just as much as a hero riding off into the sunset is not copyrightable in a movie script. However, a hero riding off into the sunset with bananas in the pistol holsters would be.
This is what I would want to hear more about when discussing if Copilot violates copyright.
This isn't true. As an analogy, consider that forcing people to not own slaves isn't the omission of freedom. See also https://www.gnu.org/philosophy/freedom-or-power.en.html
Like thinking you're ending slavery by freeing all the current slaves but not making it illegal to own, buy, and sell slaves, or capture previously free people into slavery. Guess if you'd have slavery again very soon?
The analogy is about freedom vs lack thereof, not manual labour vs software. And as you see, it works very well.
I'm worried about exactly the opposite: having Copilot help me write code that seems quite generic to me, but which in fact makes my code subject to a license I don't even know about, and/or simply violates copyright.
For an open-source project this could be embarrassing but probably fixable. It gets more complicated if FAANG is doing due diligence on your company. I can see Copilot being both an accelerant and, later, a liability for startups.
I guess time will tell how much acquiring companies (my worry) care about Copilot. Given the difficulty hiring good devs, and the productivity level of body-shop devs, I see it getting a whole lot of use very soon, acknowledged or not.
But also.. In that case, when I commission an artist to paint my portrait, surely I can't claim to be the artist.. But I'm no lawyer.
I'm not sure there is a contractual agreement in GitHub's co-pilot that says: "Any code you write here is commissioned work". But honestly I didn't read the T&C's.
So I think you MAY have debunked my analogy, but not the main reason for the analogy.
As software takes a back seat (or at least a "normal" seat) in society, would we see a normalization of income? Could this be hastened by the development and introduction of tools such as copilot?
Potentially, unless there are new / better things that humans can claim they can provide compared to AI tools. This is the point where I think you and I agree, and I think it's your primary argument in any case (unless I'm mistaken).
Compare the visible output of someone writing in assembly vs someone writing on top of a modern web framework. Is assembly harder? Yeah. But the web framework is going to give you a usable product in a fraction of the time with way more features. And that's worth more money to the company you work for.
It's always going to be a knowledge worker's job. It's always going to reward experience and creativity and attention to detail. A lot of programming is looking at the world, seeing a gap in what exists, and figuring out what best fits that gap. An AI can't do that. Programming is making 1000 tiny decisions that can't possibly be specified completely by a product manager and need a human to weigh the tradeoffs.
Thats what everybody in the chess world said: "AI can decide low level stuff. This one move. This small attack on a rook. What it can't do is conceive of how to take a bunch of different tactics and put them together to produce a game of chess."
...Until Deep Blue beat Garry Kasparov.
> It can't tell you if you should use postges or mongo.
Yeah, and then came: "It may be able to play chess, but it can't tell you how to play Go."
Look how that went.
So, unless you are a code monkey punching code into autogenerated skaffolding all day, your job is safe.
No, that’s forcing restriction on all users of your code.
Usage restriction is the opposite of freedom…
Forcing all of your code to be GPL is like saying “I am on a diet, so now I will force everyone else be on the same diet. Freedom!”
GPL license ensures that your users will keep the same freedoms that you got.
Of course, there's an inherent conflict - the freedom to oppress others is incompatible with freedom from oppression.
Nobody is forcing anyone to use the code.
If they chose to use it they have to abide by the licensing terms because that’s how it works. If the people laboring for free to produce this code don’t want it to be used in a proprietary application then tough luck, write the code yourself.
Every time the GPL comes up someone drags out this same old dead horse to beat on a little bit more.
until the time comes when a tax department gets the funny idea to use it, and forced you to use it, or people with guns come to your door and haul you away in the morning.
(edit: formatting)
This is a terrible analogy. Here’s a better one: I’m holding a potluck. If you decide to come, you can eat all you want. If you take food from my event, you can’t hoard it, you must share it, even if you’ve “made it better” by changing it somehow after you left.
Don’t like my rules? OK, don’t come to my potluck.
By analogy, there is a law against me putting handcuffs on another, and in fact the police would stop me from doing so. Did the police protect freedom? Aren't they restricting me from handcuffing others?
In a similar manner, under the MIT I can restrict my users from modifying and compiling my source code. Is a license that means I have to let my users modify code restricting freedom? Isn't it ensuring freedom of others, in the same way that making laws of "you shall not handcuff others for no reason" is ensuring freedom of others?
Suppose that there's a law that states that water and access to it is always supposed to remain public, because water is a public good.
Suppose that someone comes tomorrow and starts claiming ownership of all the water springs in your country, he becomes the only entry point to get water, and you have to pay him a fee every time you open a tap.
Is he still free to do so? In other words, is the freedom of someone who restrict the freedoms for everyone else still a form of freedom that is worth even considering, let alone respecting?
Because the foundation of your ideas is exactly the reason why capitalism fucked things up and just let a bunch of jerks get rich without merit.
What a disingenous reply. FOSS licenses do not grant ability to replicate "in any way" that you wish. You still have to comply with the license terms. What the hell is wrong with you?
> I don't understand the choice to use GPL for any reason ...
The reason is people like YOU.
either code is owned by its licensors or it isn’t.
> I also want to know why people think their code is so special that no one else could have ever come up with it independently.
I’ve never heard anyone argue this in the real world, ever —and i’ve been involved in this space for years.
If someone doesn’t want our code, then they can go ahead and write their own from scratch. We’re certainly not stopping them.
many people do seem upset at us that we’re sharing code, tho. particularly that group who primarily make their fortunes from other people’s work.
It’s the users of that closed-source commercial software that lose freedoms.
How many times does it have to be stated? GPL is for the users.
Do you feel the same about other creative processes as well? Can I rip a Justin Bieber's song and say that it's mine just because it's a shitty song anyway, so who cares? Or does this only apply to software because software is somehow an "inferior" art? Do licenses even have any legal value to you?
Boost Software License - Version 1.0 - August 17th, 2003
Permission is hereby granted, free of charge, to any person or organization obtaining a copy of the software and accompanying documentation covered by this license (the "Software") to use, reproduce, display, distribute, execute, and transmit the Software, and to prepare derivative works of the Software, and to permit third-parties to whom the Software is furnished to do so, all subject to the following:
The copyright notices in the Software and this entire statement, including the above license grant, this restriction and the following disclaimer, must be included in all copies of the Software, in whole or in part, and all derivative works of the Software, unless such copies or derivative works are solely in the form of machine-executable object code generated by a source language processor.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, TITLE AND NON-INFRINGEMENT. IN NO EVENT SHALL THE COPYRIGHT HOLDERS OR ANYONE DISTRIBUTING THE SOFTWARE BE LIABLE FOR ANY DAMAGES OR OTHER LIABILITY, WHETHER IN CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
Permission to use, copy, modify, and/or distribute this software for any
purpose with or without fee is hereby granted.
THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES WITH
REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY
AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY SPECIAL, DIRECT,
INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM
LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR
OTHER TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR
PERFORMANCE OF THIS SOFTWARE.GPL uses a different definition of freedom, which I prefer. They look at consequences of restrictions / permissions, and their implication on freedom (not just for me, but for everyone). So some restrictions can lead to actually more freedom, while some permissions can actually decrease freedom.
This is similar to gun-control. While it reduces freedom for gun owners, it allows everyone to be more free of hanging out anywhere they want without being afraid of being shot. Similar arguments can be made for vaccine mandates.
So GPL restricts usage of software because in the long term it gives back power to users, which will be more free.
Eh. I see what you’re saying about gun control, but the idea that “some restrictions can lead to actually more freedom, while some permissions can actually decrease freedom” is actually very American.
The free software movement says that everyone deserves software freedom. The Declaration of Independence similarly says “We hold these truths to be self-evident, that all men are created equal, that they are endowed by their Creator with certain unalienable Rights, that among these are Life, Liberty and the pursuit of Happiness.” While I haven’t found a source confirming it, I think that the founders believed that the freedom of speech was one of these unalienable rights.
The GPL puts restrictions in place to make sure that downstream projects give users software freedom. The Constitution put restrictions in place to ensure that the federal (and nowadays the entire) government doesn’t interfere with our unalienable rights.
Take a look at how the first amendment is worded:
“Congress shall make no law respecting an establishment of religion, or prohibiting the free exercise thereof; or abridging the freedom of speech, or of the press; or the right of the people peaceably to assemble, and to petition the Government for a redress of grievances.”
The first amendment does not grant the freedom of speech because it doesn’t need to be granted. From the founders’ perspective, god already grants the freedom of speech to everyone forever. The key phrase here is “Congress shall make no law”. The first amendment is restricting Congress to ensure freedom.
The idea that “some permissions can actually decrease freedom” is also present in the Constitution. For example, take a look at Article I sections 8 and 9. The framers of the Constitution could have given Congress the power to pass any law. Instead, they chose to specifically enumerate what Congress can and cannot do.
Perhaps, though, most Americans don’t know much about our founding and think that freedom=just let me do what I want. I don’t know.
For this reason, I had idea to make up a new license (although I will not write most of my ideas here but will do so elsewhere). But, its main working would be: mostly you can do whatever you want (including omitting attribution and copyright notices) without worrying about the license, but you cannot use legal processes (such as lawsuits, DMCA, etc) to prohibit these freedoms to any downstream recipients (regardless of how many). The license would also ensure patents can be used freely, disclaimer of warranty (if the license is included in the copy and the recipient has not paid for the copy), and some other things to ensure freedom (although there can be some restrictions on the use of trademarks (e.g. to avoid false advertising), and some things to avoid working around the freedoms in certain ways). You can be forgiven any number of times, though; the license will not be terminated. Furthermore, for a practical reason of license compatibility, relicensing by GPL3 and AGPL3 (and possibly also CC-BY-SA 4.0, for works other than computer programs) are also allowed, as long as you have a copy of the source code and can satisfy the terms of those licenses.
Really? What exactly does this CoPilot thing actually spit out? I can't help but think that it spits out near verbatim, which in the UK is probably dodgy on Copywrite.
You then go on to decide that the GPL isn't for you. That's fine. You even explain that you are ill-equipped for something. That too is fine.
You are not a fan of free or "libre" stuff. That comes across loud and clear. Thank you.
On the contrary, forcing your code to be open forever is the only way to preserve freedom.
A similar line of thought is the "paradox of tolerance", which posits that if a society tolerates the intolerant, the tolerance of that society will lessen.
> I don't understand the choice to use GPL for any reason
No, you just don't understand the GPL.
Or, for some reason, pretend not to.
Truly a classic.
Determining what freedom should mean is not, and has never been, a simple matter of "well, if you make any restrictions on it, then it's not real freedom, so everyone just gets to be free!" It's all about finding balance, and dealing with nuance, and all that frustrating hard stuff.
Which is a freer society: one that restricts late night partiers from playing loud music in residential areas, or the one that does not?
A better question would be whether they would take legal action against a competitor that creates a copilot equivalent and publicly states that they trained it on leaked, proprietary M$ source code. That would actually be an example of hypocrisy.
Because these models work better with more data and presumably this a lot of high quality data that they already have lying around anyway? Because there no downside according to their own reasoning? Because it would shut up a lot of these criticisms right away? Because marketing would be so much easier with that kind of dogfooding?
In short: because according to their own story there would be only upsides, no downsides.
Just because they use FLOSS licenses, does not allow them to evade things like Affero GPL3. And, to that end, if they are using Affero, I want the source to the whole copilot infrastructure -or- proof they used no AGPL3 code anywhere.
But that's not as damning as it sounds.
First, we know Copilot, if given the right prompt and told to autocomplete repeatedly without any manual input, can regurgitate bits of code seen many times in many different repositories, like the famous Quake fast inverse square root function and the text of licenses. That doesn't mean it does so under normal prompts and normal use. Perhaps it does sometimes, and that would be a real concern. But any regurgitation that isn't under normal use, which only happens if the user is trying to make Copilot regurgitate, is not a problem when it comes to copyright violations of open source code (since anyone trying to violate an open source license can do so much more easily without using Copilot), yet it may still be a problem when it comes to leaking confidential information.
Second, whether something is a copyright violation and whether it risks leaking confidential information are somewhat orthogonal. A copyright violation usually requires at least several lines of code, and more if the copying is not verbatim, or if the code is just a series of function calls which must be written near-verbatim in order to use an API. On the other hand, `const char PRIVATE_KEY[] = ` could hypothetically complete to something dangerous in just one line of code. That said, it almost certainly wouldn't, since even if a private key was stored in source code in the first place (obviously it shouldn't be), it probably wouldn't be repeated enough to be memorized by the model. Yet…
…third, the risk tolerances are different. If, to use completely made-up numbers, 0.1% of Copilot users commit minor copyright violations and 0.001% commit major ones, that's probably not a big deal considering how many copyright violations are committed by hand – sometimes intentionally, mostly unintentionally. (When it comes to unintentional ones, consider: Did you know that if you copy snippets from Stack Overflow, you're supposed to include attribution even in any binary packages you distribute, and also the resulting code is incompatible with several versions of the GPL? Did you know that if you distribute binaries of code written in Rust, you need to include a copy of the standard library's license?) But when it comes to leaking confidential information, even one user getting it would be somewhat bad (though admittedly Microsoft does distribute much of their source code privately to some parties), and taking even a small risk would be a questionable decision when there is a ready alternative.
If Microsoft/Github ever made that argument, that also means that when Copilot is using GPL software as input, the output can only be released under the GPL.
Copyright licenses don't apply to small snippets, no matter if you think they do, and learning and applying other people's code isn't prohibited by the license, and thank god, can't be prohibited.
There's also a setting at https://github.com/settings/copilot (link only works if you've signed up for copilot) that will check any suggestion on the server against hashes of the training set, and block anything that exactly duplicates code in the training set (with a minimum length, so very common code doesn't get completely blocked). Users must choose the value for this setting when they sign up for copilot.
source: I work on copilot at github
While GitHub might have a license to use that code to train the model, it’s debatable what license applies to the output of the model, and what users of the model can do with it.
It’s possible for an AI to reproduce something so close to the original that it would be considered an infringement on the original work.
I'm pretty sure there are a lot of those.
Past MS engineers have been commenting for a decade on how MS has grown too big, can't manage, and has become a monolith "too big to fail". By nature when engineers are small pieces of a giant machine, they don't do their best work. And those with the experience move on to better things.
>there is nothing really surprising in this leak. Microsoft does not steal open-source code. Their older code is flaky, their modern code excellent. Their programmers are skilled and enthusiastic. Problems are generally due to a trade-off of current quality against vast hardware, software and backward compatibility.
[0] https://web.archive.org/web/20040401115821/http://www.kuro5h...
I can think of annoyances but awful? Come on.
Specifically, clear anti-user choices that exceed by far being "annoying":
* Making it exceedingly difficult or impossible to use the OS without logging in with a Microsoft account.
* Forcing the user in various ways to surrender data to Microsoft. Some of them can be disabled if you really go out of your way, others can't.
* Prompting me again and again to switch to Edge and other MS defaults. I've had the same install for a few years now and NO, I don't want to change to "Microsoft recommended defaults", no matter how many times you ask me.
* Showing the same "OS setup" screen after some updates, requiring me to pay very close attention to what I'm clicking, lest I select something MS is trying to lead me to. The amount of attention required from the user on those screens corresponds quite well with anti-user behavior.
This is hilarious. I recently got a new laptop that has window$ 11. After setting it up with a Non Microsoft email (which required some good fight), I tries to install some random app from the Microsoft store, but got a "something went wrong please try again" on the first screen.
It's pathetic. I haven't used Windows since Win 7 , which I basically installed for gaming. Seeing the latest version of the OS makes me feel sorry for them. That's why Apple with all their assholery is eating their lunch (on the flip side my wife just got a MBP m1 and I was pleasantly surprised that it has hdmi port, magsafe, several USBc ports. Apple seems going in the right direction.)
The telemetry makes this clear. Reboots and updates even more so.
The UI lag and stealing of focus ("oh, you're typing a document... too bad, I want to launch a new Explorer window that will immediately steal focus") make it clear that the computer is in charge and will probably listen to your requests, but on the timeline it chooses.