This investigation should not stop at GitHub Co-Pilot, large language models currently that are trained on huge amount of data should also be investigated as I'm sure there are lot's of problems to be found there.
GitHub Copilot investigation
801–810 of 1001 posts
Re: GitHub Copilot investigation
#802I'd be rather saddened if Copilot was shut down or neutered because of a few vocal few protesting against it. It's been a massive productivity improvement to our senior devs, and I got so used to it that it's an annoyance when Copilot doesn't respond.
Careful though, you are trading yours (and their) muscle memory and brainpower to be locked into a proprietary solution. Reread your post. Doesn't it sound scary? You are blocked from even thinking and crafting because a specific web service is down. Even if Google is down you can go direct to Stackoverflow and MDN, and have a choice of information sources. Also what is "productivity" ... as in features built / month…
Correction: even if Google is down, you won’t notice, because DDG works just fine. Last time I went to any of Google websites was 4 years ago.
Re: GitHub Copilot investigation
#803Earlier quoted context omitted.
Microsoft has owned Github for how many years... and this is the _first_ test?
Embrace, Extend, Extinguish. Takes a while to get to the Extinguish phase.
Yep. Four years might be a long time in Silicon Valley, but not in Redmond. Satya Nadella has worked at Microsoft 30 years. This is a company that can think past the next quarter.
Re: GitHub Copilot investigation
#804It's too bad we can't experiment with interesting things like Copilot without worrying about remuneration and the respecting of rights. But that's the way of the world - we must think of these things. MS/Github should give code copyright holders a simple and easy way to opt-out of contributing their code to the Copilot corpus. Currently the only way to opt-out is to make your repo private. That's not good enough. It…
They have done so already with their license and there's no legal reason for them to have to opt-out.
Re: GitHub Copilot investigation
#805Imagine you are a history or philosophy teacher in 2100.
How cool it is to discuss these kind of issues? What do you think about the "erasing open source community" argument from the historical perspective? What does it have in common with industrial revolution?
Even though the real life implications are real, I find it fascinating and not so simple to unravel.
Re: GitHub Copilot investigation
#806Here are a few thoughts I haven't formulated before: It seems clear enough to me that training AIs on copyrighted works is typically or commonly a fair use under existing law, because the AIs can and commonly do learn non-copyrightable elements and aspects of those works. It's very obvious from enormous numbers of examples that current AI systems are capable of learning much more abstract features of human culture (g…
I don't know why we should be concerned with the status quo of copy-write law at all with respect to AI. ML is categorically new in how it applies to these domains, and it's not clear to me at all that rules that apply to humans have much to do at all with rules which should apply to machines.
Imo it is very simple: IP law is intended to incentivize creative work, so that it remains possible to profit from one's creation in an environment where it might be easier to copy than it is to create. We just need to figure out what outcome we want to create: one which incentivizes human creations, or AI "creations" - and build a legal framework to support it.
Re: GitHub Copilot investigation
#807Here are a few thoughts I haven't formulated before: It seems clear enough to me that training AIs on copyrighted works is typically or commonly a fair use under existing law, because the AIs can and commonly do learn non-copyrightable elements and aspects of those works. It's very obvious from enormous numbers of examples that current AI systems are capable of learning much more abstract features of human culture (g…
>If human beings had much more accurate memories, copyright would be quite a bit more intrusive (and/or quite a bit less effective) because, following any exposure to some kinds of works, we could use our own memories to reproduce those entire works from scratch for our own use or pleasure without obtaining authorized copies from elsewhere. I don't know the name, but I remember some sci-fi story about some academy wh…
Plenty of living musicians today have this capability.
It turns out that humans can extrapolate generalisms to a degree we are currently unable to explain clearly enough as a model to imitate.
It turns out that much ML is merely referenced regurgitation.
Marketing and hype are rather advanced skills in 2022, however…
Re: GitHub Copilot investigation
#808Getty Images already handled this issue with graphics. Most of their catalog was scraped early on by AI art generators. 'Errbody knows this because the Getty Images watermark appears in a lot of AI generated art. Getty Images, in turn, banned the sale of AI generated art because it is legally tainted.
The same thing will happen to source code produced by AI code generators. Github itself, or some entrepreneur, will come up with a way to identify and flag projects containing AI generated code based on models constructed from open source projects, so that those derivative works will not inadvertently be incorporated into other software that is concerned with such a flag. (They probably will also come up with an NFT-based mechanism of some sort to allow open source project rights holders to authorize incorporation of their code into AI models such that derivative works containing those fragments would not be subject to flagging.)
Hey YCombinator, give me $10M to make a billion dollar company that "lives at the intersection of" blockchain and open source. (Haha, No.)
Re: GitHub Copilot investigation
#809Earlier quoted context omitted.
It seems pretty clear to me, training an AI on copyrighted materials is not fair use. I'm not sure why you seem to think it is fair use
It depends on what the AI is learning. If the AI is learning to repeat text (e.g. Copilot) or images (e.g. Dall-E), then that makes it possible to reproduce the copyrighted works, so I would agree that that case is not fair use. -- It would be akin to compressing and distributing those works. If the AI is learning patterns -- such as "muggle" being a noun that relates to Harry Potter, or that the lemma for "muggles"…
I'm not sure how Copilot works, but I don't believe Dall-E repeats images. From my understanding it creates visual concepts of words and uses them to create entirely new images. If Copilot works in the same way for code, I honestly don't see that there should be any copyright issues here.
Re: GitHub Copilot investigation
#810While the moral and legal discussions here are interesting and worth exploring, I find this text hyperbolic. Its premise is that the main way that people currently interact with open-source projects is by digging into their source code, copy-pasting away a snippet of code that solves a particular problem, and then of course giving the authors the required attribution. This is far from the truth. The main usage of mos…
The whole point of open-source is about re-using and modifying the source though. Sure it allows using the product, but that's hardly the defining factor of open-source.