undefined | Better HN

0 pointsenriquto1y ago0 comments

> The training data contains most likely insane amounts of copyrighted material.

If that is the case then the weights must inherit all these copyrights. It has been shown (at least in image processing) that you can extract many training images from the weights, almost verbatim. Hiding the training data does not solve this issue.

But regardless of copyright issues, people here are complaining about the malicious use of the term "open source", to signify a completely different thing (more like "open api").

0 comments

tempfile1y ago

> If that is the case then the weights must inherit all these copyrights.

Not if it's a fair use (which is obviously the defence they're hoping for)

anon3738391y ago

Also, fair use is just one defense to a copyright infringement claim. The plaintiff first has to prove the elements of infringement; if they can't do this, no defense is needed.

j / k navigate · click thread line to collapse