zlacker

>>ssgodd+(OP)
Even if they win against openAI, how would this prevent something like a Chinese or Russian LLM from “stealing” their content and making their own superior LLM that isnt weakened by regulation like the ones in the United States.

And I say this as someone that is extremely bothered by how easily mass amounts of open content can just be vacuumed up into a training set with reckless abandon and there isn’t much you can do other than put everything you create behind some kind of authentication wall but even then it’s only a matter of time until it leaks anyway.

Pandora’s box is really open, we need to figure out how to live in a world with these systems because it’s an un winnable arms race where only bad actors will benefit from everyone else being neutered by regulation. Especially with the massive pace of open source innovation in this space.

We’re in a “mutually assured destruction” situation now, but instead of bombs the weapon is information.

>>dissid+B6
Trying to prevent AI from learning from copyrighted content would look completely stupid in a decade or two when we have AIs that are just as capable as humans, but solely due to being made of silicon rather than carbon are banned from reading any copyrighted material.

Banning a synthetic brain from studying copyrighted content just because it could later recite some of that content is as stupid as banning a biological person from studying copyrighted content because it could later quote from it verbatim.

>>logicc+Oa
It's not exactly a synthetic brain though, is it? LLMs are more like lookup tables for the texts they're trained on.

We will not have "AIs as capable as humans" in a couple decades. AIs will keep being tools used by humans. If you use copyrighted texts as input to a digital transformation, that's vopyright infringement. It's essentially the same situation as sampling in music, and imo the same solutions can be applied here: e.g. licenses with royalties.