> [OpenAI] definitely can't argue that using copyrighted material for training is fine, but using output from other commercial models isn't. That's too inconsis

> [OpenAI] definitely can't argue that using copyrighted material for training is fine, but using output from other commercial models isn't. That's too inconsistent. Well, they can argue that, if they're fine with being hypocrites. They're hypocrites.