3 ms·
Just so I understand the argument, OpenAI would be claiming that anyone using their model outputs to distill i.e. train a smaller model on their model is a viol
by janalsncm 3y ago
Just so I understand the argument, OpenAI would be claiming that anyone using their model outputs to distill i.e. train a smaller model on their model is a violation of copyright, but them training on the entirety of the internet (including copyrighted material) is not a violation?
- JimDabell 3y ago> OpenAI would be claiming that anyone using their model outputs to distill i.e. train a smaller model on their model is a violation of copyright OpenAI haven’t claimed this. They are refusing to generate new data for ByteDance by suspending their account.