Hacker News Reader

Readability is one of the key goals of open-source software.

That said, the inclusion of GPL-licensed code in training sets may yet force the release of those models under the GPL.

But the GPL allows for learning from it. But is the learning limited to humans? Is the learning process a way of modification?

I'd really like to see this tested at a court.

Can you even call it "learning"? We use the word as a metaphor, but it's really not the same thing as what a human does.