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

71–80 of 1001 posts

Re: GitHub Copilot

#71

Post by not2b »

dekhn wrote: post #48
iwintermute wrote: post #21

So 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").

But it actually is explicitly copying the text. That's how it works. The training data are massive, and you will get long strings of code that are pulled directly from that training data. It isn't giving you just the style. It may be mashing together several different code examples taking some text from each. That's called "derivative work".

Re: GitHub Copilot

#72

Post by vgalin »

Deathmax wrote: post #13

GitHub says you don't need to credit GitHub for any of the code suggestions, but since it's trained on public sources of code, anyone have a clue on potential licensing pitfalls?

From the FAQ at the bottom of the project showcase page[0]:

"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. Here is an in-depth study[1] on the model’s behavior. Many of these cases happen when you don’t provide sufficient context (in particular, when editing an empty file), or when there is a common, perhaps even universal, solution to the problem. We are building an origin tracker to help detect the rare instances of code that is repeated from the training set, to help you make good real-time decisions about GitHub Copilot’s suggestions."

[0] https://copilot.github.com/

[1] https://github.co/copilot-research-recitation

Re: GitHub Copilot

#73

Post by fulafel »

CityOfThrowaway wrote: post #59

This is very impressive! OpenAI’s tech opens an ethical Pandora’s box: 1. It’s clear that the raw inputs to all of OpenAI’s outputs originated with real, human creativity. 2. So, in a sense, OpenAI is laundering creativity. It reads in creative works, does complicated (and, yes, groundbreaking) transformations, and produces an output that is hard to trace to any particular source. 3. Yet, isn’t that effectively what…

What are some ethical problems that could emerge from the box? Maybe unfair competition from having very good tools compared to other programmers, or havin irresponsibly shallow understanding of what the produced code does?

Re: GitHub Copilot

#74

Post by fabiospampinato »

adelarsq wrote: post #23
fabiospampinato wrote: post #3

I'd like to see something like that, but with knowledge about every single file in the codebase, and running locally.

There is Tabnine that can work like this

Last time I tried Tabnine it wasn't really of much use to me, the top of the line GPT-3 is a much much bigger model, it should be able to do much more intelligent things.

Re: GitHub Copilot

#75

Post by fay59 »

natfriedman wrote: post #60

Hi 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?

Re: GitHub Copilot

#78

Post by akersten »

patwolf wrote: post #37

I wonder if there's any potential for Copilot to suggest malicious code because it's been trained on an open source projects containing intentionally malicious code.

Maybe not malicious per se, but certainly I'd be concerned about seemingly-correct but actually-wrong code being suggested. Considering how often the top StackOverflow answer is slightly wrong or how often antipatterns crop up across various projects, I'm sure the training data is nowhere near "perfect code" - implying the output cannot be perfect either.

Re: GitHub Copilot

#80

Post by enriquto »

natfriedman wrote: post #60

Hi 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.

> Hundreds of developers are using it every day internally, and the most common reaction has been the head exploding emoji

Apart from developing this "head exploding" stuff, couldn't some of these incredibly talented hundreds of developers fix the github code search?

Post reply on HN