Late Stage Capitalism

28 readers

1 users here now

founded 5 years ago

MODERATORS

GitHub Support just straight up confirmed in an email that yes, they used all public GitHub code, for Codex/Copilot regardless of license. (lemmy.ml)

submitted 3 years ago by AgreeableLandscape@lemmy.ml to c/latestagecapitalism@lemmygrad.ml

2 comments fedilink hide all child comments

Source: https://cybre.space/@tindall/106539167944483388

From the same Mastodon thread:

The model is known to reproduce some code, including GPL-licensed code, verbatim; therefore, it must contain verbatim copies of that code, however it is encoded.

[...]

the snippet in question is clearly, deeply original. it is a cursed coding crime that contains several "magic constants" with high entropy.

So it should be required to be open source now, right?

you are viewing a single comment's thread
view the rest of the comments

[–] xenith@lemmy.ml 0 points 3 years ago (1 children)

Could anyone provide context to this?

[–] ksynwa@lemmygrad.ml 1 points 3 years ago

GitHub have created an AI that can write code. To train this AI, they used all the public code uploaded to GitHub (if the above screenshot is correct). This could be a GPL violation though I don't know the exact legalese.

If you are unaware, GPL is a software license which basically says that if my software is GPL licensed and you use it as a derivative to create something else, whatever you create should be GPL licensed too. For GPL licensed software, it is mandatory to share the source code with the users, so it could mean that GitHub is obligated to share the code used to create this AI since it used GPL licensed work.

There are other licenses other than GPL with similar caveats. I am using GPL as an example because it is one of the most prevalent of such licenses.