Sounds like a big distraction.
Why are the green ones always angry?
haha, you asked a question that I've always wondered.
GitHub Copilot
171–180 of 1001 posts
Re: GitHub Copilot
#172Earlier 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.
Is it any different than training a human? What if a person learned programming by hacking on GPL public code and then went to build proprietary software?
Re: GitHub Copilot
#173Earlier 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.
Perhaps we need GPL v4. I don't think there is any clause in current V2/V3 that prohibits learning from the code, only using the code in other places and running a service with code.
Re: GitHub Copilot
#174Calling it now, there will be a "Copilot considered harmful" post. If you need to go through the suggested code to ensure it's correct, you may as well write it yourself? If you glance at it and it looks about right, you can potentially overlook bugs or edge cases, you'll lose confidence in your own code since you didn't properly conceptualise it yourself. Potentially for newer developers it robs them of active exper…
Honestly, this attitude (dismissing the feature even without trying it out) comes across as insecurity and gatekeeping.
Re: GitHub Copilot
#175Hi HN, we've been building GitHub Copilot together with the incredibly talented team at OpenAI for the last year, and we're so excited to be able to show it off today. Hundreds of developers are using it every day internally, and the most common reaction has been the head exploding emoji. If the technical preview goes well, we'll plan to scale this up as a paid product at some point in the future.
Have there yet been reports of the AI writing code that has security bugs? Is that something folks are on the lookout for?
I haven't seen any reports of this, but it's certainly something we want to guard against: https://copilot.github.com/#faq-can-github-copilot-introduce...
Re: GitHub Copilot
#176I wonder if CoPilot uses Github's private repositories to train itself, which would allow malicious users to somehow obtain code or designs that they otherwise would not be able to view.
Re: GitHub Copilot
#177Should I be impressed that the example parse_expenses.py on the home page doesnt include any error handling and uses a float for currency? This seems like it's going to revolutionize copy and paste programming.
The output is great for a quick one-off script. Maybe if you make the comments look more “enterprise-y”, it’ll go for more careful code?
I would say the use of float makes it a nonstarter - even for "quick one-off" scripts. That's a fundamental error in the generated code. It maybe looks correct at a quick glance but it's only introducing subtle errors to find down the line.
Re: GitHub Copilot
#178Earlier 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.
Would you be okay with a human reading your GPL code and learning how to write closed source software for closed source projects?
Re: GitHub Copilot
#179So if it was trained using "source code from publicly available sources, including code in public repositories on GitHub." was it also GPLv2? So everything generated also GPLv2?
No, a model trained on text covered by a license is not itself covered by the license, unless it explicitly copies the text (you cannot copyright a "style").
> you cannot copyright a "style"
This line of thinking applies to the code generated by the model, but not necessarily to the model itself, or the training of it.
Re: GitHub Copilot
#180Earlier quoted context omitted.
Yeah, running locally would be my preference. I get "Antitrust (2001)" vibes from this, but that's the tinfoil hat side of me.
Like this? https://visualstudio.microsoft.com/services/intellicode/ I bet it's the same people, trying to push their crap into all sorts of successful products.
The IntelliCode and Copilot teams have been collaborating closely together, since we want them to provide a "better together" experience. However, the underlying tech isn't the same. Copilot is powered by OpenAI Codex, and enables rich code synthesis via a cloud service. Whereas IntelliCode uses multiple local models, to enhance various parts of the editor (e.g. prioritizing the completion list based on your context, detecting "repeated edits" and suggesting additional refactorings).