I guess I'm missing who owns/developed Cuda, then. Like, why does Nvidia think they can disallow anyone else from using Cuda if Cuda was made and broadly used as the API before Nvidia.