I would agree that we're acting like hypocrites here. But unlike stable diffusion, GitHub didn't release their model. So it's extremely hard to know what's going on inside copilot, on the other hand we have the model of stable diffusion and we can see wether or not it has memorized copyrighted images.