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

GitHub Copilot

copilot.github.com

Post reply on HN
Print view View original

981–990 of 1001 posts

Re: GitHub Copilot

#981

Post by geekraver »

rictic wrote: post #779
billti wrote: post #688

Earlier quoted context omitted.

The landing page for it states the below, so hopefully not too much of an issue (though I guess some folks may find a 0.1% risk high). > GitHub Copilot is a code synthesizer, not a search engine: the vast majority of the code that it suggests is uniquely generated and has never been seen before. We found that about 0.1% of the time, the suggestion may contain some snippets that are verbatim from the training set.

If you pulled a marble out of a bag ten times a day, with a 0.1% chance each time that it was red: after a day you'd have a 1% chance of seeing a red marble, the first week you'd have a 6.7% chance, the first month you'd have a 26% chance, and the first working year you'd have a 92.6% chance of having seen at least one red marble. Probabilities are fun!

And you would have a whole lot of blue marbles.

Re: GitHub Copilot

#982

Post by bootlooped »

adn wrote: post #967
bootlooped wrote: post #663

Earlier quoted context omitted.

If I spent a lot of time reading open source repos on GitHub to teach myself to code, and then went out and got a high-paying job based on that knowledge, is that ethical? This seems roughly analogous to what the machine is doing.

Regardless of the legality, one of these situations is clearly ethical compared to the other. In the case were you get a job based on your knowledge of GPL software, you still must respect the license if you use that code commercially (i.e. at your new job). And yes, if you reproduce GPL code you "learned" from, you are violating the license. A company ingesting an entire GPL codebase without warning or any way to op…

But the points you listed are the same whether it's a person or a machine:

- if you reproduce GPL code verbatim, you're in violation*

- no warning that somebody/something is ingesting the repo

- no way for the repo to have opted out

- closed source (you can't get the source from somebody's brain)

- private profit

* Do we even know if the machine is more or less likely to do this? Humans are certainly capable of it.

Re: GitHub Copilot

#983

Post by dylannorthrup »

user-the-name wrote: post #926
natfriedman wrote: post #345

Earlier quoted context omitted.

In general: (1) training ML systems on public data is fair use (2) the output belongs to the operator, just like with a compiler. On the training question specifically, you can find OpenAI's position, as submitted to the USPTO here: https://www.uspto.gov/sites/default/files/documents/OpenAI_R... We expect that IP and AI will be an interesting policy discussion around the world in the coming years, and we're eager to…

> training ML systems on public data is fair use So, to be clear, I am allowed to take leaked Windows source code and train an ML model on it?

Or, take leaked Windows source code, run it through a compiler, and own it!

Re: GitHub Copilot

#984

Post by dylannorthrup »

natfriedman wrote: post #345
stefanos82 wrote: post #197

Earlier quoted context omitted.

Lots of questions: - the generated code by AI belongs to me or GitHub? - under what license the generated code falls under? - if generated code becomes the reason for infringment, who gets the blame or legal action? - how can anyone prove the code was actually generated by Copilot and not the project owner? - if a project member does not agree with the usage of Copilot, what should we do as a team? - can Copilot copy…

In general: (1) training ML systems on public data is fair use (2) the output belongs to the operator, just like with a compiler. On the training question specifically, you can find OpenAI's position, as submitted to the USPTO here: https://www.uspto.gov/sites/default/files/documents/OpenAI_R... We expect that IP and AI will be an interesting policy discussion around the world in the coming years, and we're eager to…

Fair Use is an affirmative defense (i.e. you must be sued and go to court to use it; once you're there, the judge/jury will determine if it applies). But taking in code with any sort of restrictive license (even if it's just attribution) and creating a model using it is definitely creating a derivative work. You should remember, this is why nobody at Ximian was able to look at the (openly viewable, but restrictively licensed) .NET code.

Looking at the four factors for fair use looks like Copilot will have these issues: - The model developed will be for a proprietary, commercial product - Even if it's a small part of the model, the all training data for that model are fully incorporated into the model - There is a substantial likelihood of money loss ("I can just use Copilot to recreate what a top tier programmer could generate; why should I pay them?")

I have no doubt that Microsoft has enough lawyers to keep any litigation tied up for years, if not decades. But your contention that this is "okay because it's fair use" based on a position paper by an organization supported by your employer... I find that reasoning dubious at best.

Re: GitHub Copilot

#985

Post by mitjak »

bogwog wrote: post #786
mitjak wrote: post #639

Earlier quoted context omitted.

the main argument against Copilot for me. it takes longer to grok existing code than just write it from ground up.

Plus if you use this for anything non-trivial, it's functionally similar to using a third-party dependency, except the licensing terms are vague and there's no one to ask for help. I could see it being useful maybe for testing. Even if it straight up copies code from the Linux kernel, you don't usually ship tests to customers so the GPL (probably) isn't a problem. But the problem with testing is that you need to be 1…

i'd argue the danger is even greater with tests bc the entire point of tests is to provide certainty. wrong tests are a false sense of safety that's unquantifiable without understanding every single line.

Re: GitHub Copilot

#986

Post by __MatrixMan__ »

SCLeo wrote: post #148
natfriedman wrote: post #107

Earlier quoted context omitted.

It shouldn't do that, and we are taking steps to avoid reciting training data in the output: https://copilot.github.com/#faq-does-github-copilot-recite-c... https://docs.github.com/en/early-access/github/copilot/resea... In terms of the permissibility of training on public code, the jurisprudence here – broadly relied upon by the machine learning community – is that training ML models is fair use. We are certain this…

> ...the jurisprudence here – broadly relied upon by the machine learning community – is that training ML models is fair use. To be honest, I doubt that. Maybe I am special, but if I am releasing some code under GPL, I really don't want it to be used in training a closed source model, which will be used in a closed source software generating code for closed source projects.

> a closed source model

It seems like some of the chatter around this is implying that the resultant code might still have some GPL still on it. But it seems to me that it's the trained model that Microsoft should have to make available on request.

Re: GitHub Copilot

#987

Post by nanis »

itsoktocry wrote: post #366
anthropodie wrote: post #335

Earlier quoted context omitted.

Few days back, Sam Altman tweeted this "Prediction: AI will cause the price of work that can happen in front of a computer to decrease much faster than the price of work that happens in the physical world. This is the opposite of what most people (including me) expected, and will have strange effects" And I was like yeah I gotta start preparing for next decade.

> Prediction: AI will cause the price of work that can happen in front of a computer to decrease much faster than the price of work that happens in the physical world. I'm skeptical. The envelope of "programming" will continue to shift as things get more and more complex. Your mother-in-law is not going to install Copilot and start knocking out web apps. Tools like this allow programmers to become more productive, wh…

For every programmer who can think and reason about what they are doing, there are at least 10 who just went to a bootcamp and are not afraid to copy and paste random stuff they do not understand and cannot explain.

Initially, they will appear more productive with CoPilot. Businesses will decide they do not need anybody other than those who want to work with CoPilot. This will lead to adverse selection on the quality of programmers that interact with CoPilot ... especially those who cannot judge the quality of the suggested code.

That can lead to various outcomes, but it is difficult to envision them being uniformly good.

Re: GitHub Copilot

#988

Post by thamalama »

so gh has trained their AI programmers as much as they could with shit code examples, now to bridge the last mile they need some hand holding? sign me up! will I get a real online gf out of this finally?

Re: GitHub Copilot

#990

Post by hakre »

Abimelex wrote: post #675
iandanforth wrote: post #619

Earlier quoted context omitted.

What is the licensing for code generated in this way? GPT-3 has memorized hundreds of texts verbatim and can be prompted to regurgitate that text. Has this model only been trained on code that doesn't require attribution as part of the license?

Not sure how this is handled in US, but in Germany a few lines of code have in general not enough uniqueness to be licensed.

So the corpus has been compiled under license and the derivative work is eligible for distribution?
Post reply on HN