alt Hacker NewsThe problem is that with generative AI, I have no means of protecting my work from being stolen.
It does not matter what license I put up. It doesn't even matter if I make it publicly available or not. LLMs have been trained on pirated material, they don't even have the decency to buy a copy. Even if I show my project to no one and just have a private repo on Github the code might still be used to train LLMs.
Your GPL licensed library? Yeah, we used claude to rewrite it and released it under MIT.
Now that wouldn't be so bad. One could argue copyright has long held back progress in certain areas. The problem is, the rules only apply one way. The rent seeking oligarchs of the tech industry can steal everything but I can't.
They can just eat the cost of a lawsuit, I can't. They can just decide to make a special deal with Disney to use their copyrighted material, I can't.
Sure the days of free markets capitalism are long gone. A few monopolists controlling the market has long been the norm. But AI makes it even worse. So much worse.
Replies
> a private repo on Github
Delete your github repos and operate your own gitolite instance. Feed vibecode to GitHub so the LLMs coprophagically train on their own slop.
AI will be the straw that breaks the camel's back on copyright Imo. We've known since 90s Napster music piracy that copyright is broken in the information age, and its just a flimsy set of unprincipled edicts meant to protect those with power and money.
Nowadays AI companies have more money and lawyers than most movie studios, so
I predict that there will be a billion dollar company/ies (probably exist even now in stealth mode), whose business model will be to slopfork existing software - after all AI has proven to be very capable at that.
With trillions of dollars both supporting and opposing this business model, something will probably change in some way wrt copyright, and hopefully in a way that's an improvement to the average person.