> MS absolutely has the authority to copy, use, and even train their models on your GPL-license code, because you agreed to let them do that when you signed their EULA when you decided to host your code on GitHub.
What about GPL code which you don't own, but post to Github, Like the gcc mirror repo?
you must have the right to publish the code you put on github.com, and by publishing to github.com, you assert that you have the rights to do so. you also grant GitHub the right to show that code to others, no matter what license your code is licensed under.
why does no one read the terms of service or license agreements? these questions are answered there and this "copilot is stealing" stuff won't even make it to court.
Even if it is true that the ToS that noone reads allows GitHub to blatantly violate the license of your code that anyone could choose to upload to GitHub without the copyright holder's consent, people are going to do it anyway en masse, and that code is going to get eaten by Copilot. Even if copyright holders are constantly playing the game of reporting public repos to GitHub to remove it's not going to be enough.
>this "copilot is stealing" stuff won't even make it to court
IANAL but I am heavily skeptical of your confidence here.
Please. If someone uploads a pirate copy of a movie to GitHub, GitHub doesn't get an additional license to anything, on top of whatever license the uploader chose.
Maybe GitHub is saying it's not their fault and that they were misinformed.
Those types of arguments usually don't get all that far with copyright violations.
if someone uploads a pirated movie to github, github won’t get the license if it goes to court, but they won’t get in trouble for the assumption that they had the license, either. the uploader had both pirated and committed wire fraud as gar as the github terms of service agreement is concerned.
either way, github is absolved. if a DMCA claim is filed on the movie, then that gets quarantined and removed from the list of things that they can show their users.
the user that uploads stuff to github.com attests that they have the right to upload it. by being uploaded, github can assume that it has the rights it asks of users unless and until they are told that they do not have those rights. so, github are covered until they are formally told that they are not, at which point they must restrict access to that data to themselves and others, which they regularly do.
these types of arguments do indeed work very well if github reacts promptly when they are told that they are using rights they do not have. github is not primarily used for piracy, like thepiratebay, and thus has a valid claim that they were lied to by a user. this is when the user who uploaded gets involved legally if the true copyright holder chooses to involve them.
you being mad at github, microsoft, or me doesn’t make any of us wrong.
The confidence of it not going to court is wrong. It will go to court, but the people saying it's obviously infringing are wrong. And even if it is Microsoft can just change their terms to make copilot compliant.
> Even if copyright holders are constantly playing the game of reporting public repos to GitHub to remove it's not going to be enough
There is no copyright police outside of criminal infringement.
The DMCA gives you the tools to protect your copyright, if 1000 people infringe on your copyright you need to be ready to sue 1000 people OR attack the platform under safe harbor.
If someone uploads your code to github, you need to find it and ask them to remove it.
If someone uses your code via copilot, infringingly, you need to find it and ask them to remove it.
What about GPL code which you don't own, but post to Github, Like the gcc mirror repo?