Live data from Hacker News

GitHub Copi­lot inves­ti­ga­tion

githubcopilotinvestigation.com

41–50 of 1001 posts

Re: GitHub Copi­lot inves­ti­ga­tion

#41
This is a bit off-topic, but I wonder if there are people/teams right now creating git repos, doing the source code equivalent of "SEO" on it, and embedding backdoors in stupidly overoptimized for the training process code?

I wonder when we'll hear about the first big hack that gets traced back to production code pushed live after CoPilot "suggested" eval(base64decode({webshell}))

Re: GitHub Copi­lot inves­ti­ga­tion

#42
It seems like there's no good license that places absolutely no restrictions or requirements on people using your code (such as attribution and respecting patent rights) worldwide.

I want my code to be used the way people treated text in the old days. There's texts that have been re-written, added to and edited by thousands of people over the centuries and yet they don't come with thousands of pages of attribution notices because why would they?

Re: GitHub Copi­lot inves­ti­ga­tion

#43
post #12

in countries where there is no fair use (most of the world outside the US) it seems quite likely copilot is willful, commercial scale copyright infringement

Fair use is unusually permissive in the US, but most countries have very complex copyright rules to allow e.g. a televised interview in a room with contemporary paintings, without getting permission from the copyright holders of those paintings. It'd certainly make for interesting cases.

Re: GitHub Copi­lot inves­ti­ga­tion

#45
The example given is "sparse matrix trans­pose in the style of Tim Davis", but someone who wanted something with such specificity would be able to just take it from Github anyway, perhaps with a little more searching.

Re: GitHub Copi­lot inves­ti­ga­tion

#46
100% correct takes in the piece, this is just ridiculous

>"Tim Davis gave numer­ous exam­ples of large chunks of his code being copied ver­ba­tim by Copi­lot, includ­ing when he prompted Copi­lot with the com­ment / sparse matrix trans­pose in the style of Tim Davis /."

Copilot regurgitates code and blatantly violates licenses, not even sure what there is to argue about. Not only does it seem straight up illegal and sideline open source communities, I think the next logical step of this is that people who want to avoid having their work vacuumed up and their rights violated simply to move to proprietary software, which would be a huge disaster for open source.

Re: GitHub Copi­lot inves­ti­ga­tion

#47
post #13

> Microsoft char­ac­ter­izes the out­put of Copi­lot as a series of code "sug­ges­tions". Microsoft "does not claim any rights" in these sug­ges­tions. But nei­ther does Microsoft make any guar­an­tees about the cor­rect­ness, secu­rity, or exten­u­at­ing intel­lec­tual-prop­erty entan­gle­ments of the code so pro­duced. Once you accept a Copi­lot sug­ges­tion, all that becomes your prob­lem: > "You are respon­si­ble…

Could you help me understand the link between the two?

Re: GitHub Copi­lot inves­ti­ga­tion

#48
post #2

Doesn't really explain how co-pilot is stealing your community. I've used co-pilot and it works great until you are past boilerplate than it falls apart.

"It's not a problem that it steals code because it doesn't work that well right now" is not a valid argument, don't pretend that AI doesn't advance on a daily basis.

Reminds me of people who defend AI art with "you can tell it apart from real art" yeah no, at the rate we're moving you won't be able to tell at all in a year or two.

Re: GitHub Copi­lot inves­ti­ga­tion

#49
There are lots of comments arguing for or against Copilot on a value judgment, and having an opinion on it being ethical or legal, etc isn't going to be the same for everyone. But I think regardless of where you stand, there should be some sort of legal ruling to clarify the gray areas that Butterick breaks down.
Post reply on HN