Is it? In Sam’s case, we’re mostly talking about creative products in the form of text, audio, and video. If an artist releases a song and the song is copyrighted, it doesn’t hamper innovation and technological development. The same cannot be said when a company patents a sorting algorithm, the method for swiping to unlock a smartphone, or something similar.
In such a scenario, it will be worth it. Llm aren’t databases that just hold copy pasted information. If we get to a point where it can spit out whole functional githubs replicating complex software, it will be able to do so with most software regardless of being trained on similar data or not.
All software will be a prompt away including the closed sourced ones. I don’t think you can get more open source then that. But that’s only if strident laws aren’t put in place to ban open source ai models, since Google will put that one prompt behind a paychecks worth of money if they can.
I don’t see how you can write the law such that it allows training ai on copyrighted data without making it possible to train a special llm on a single github instead of the entire universe, and essentially treat it as a full compression of the source.
But Sam is talking about copyright and all your examples are patents
It just so happens that in AI it’s about copyright and with margarine (and most other technologies) it’s about patents.
But the point is the same. Technological development is held back by law in both cases.
If all IP laws were reformed 50 years ago, we would probably have the technology from 2050, today.
It’s all the same shit. No patents and copyrights should exist.
Is it? In Sam’s case, we’re mostly talking about creative products in the form of text, audio, and video. If an artist releases a song and the song is copyrighted, it doesn’t hamper innovation and technological development. The same cannot be said when a company patents a sorting algorithm, the method for swiping to unlock a smartphone, or something similar.
If copyrights are used to add a huge price tag to any AI development, then it did just hamper innovation and technological development.
And sadly, what most are clamoring for will disproportionately affect open source development.
If open source apps can’t be copyrighted then the GPL is worthless and that will harm open source development much more
I’m not sure how that applies in the current context, where it would be used as training data.
Because once you can generate the GPL code from the lossy ai database trained on it the GPL protection is meaningless.
In such a scenario, it will be worth it. Llm aren’t databases that just hold copy pasted information. If we get to a point where it can spit out whole functional githubs replicating complex software, it will be able to do so with most software regardless of being trained on similar data or not.
All software will be a prompt away including the closed sourced ones. I don’t think you can get more open source then that. But that’s only if strident laws aren’t put in place to ban open source ai models, since Google will put that one prompt behind a paychecks worth of money if they can.
I don’t see how you can write the law such that it allows training ai on copyrighted data without making it possible to train a special llm on a single github instead of the entire universe, and essentially treat it as a full compression of the source.