Undermining the foundation of open source AI

While I understand the criticism from an open source perspective, I have mixed feelings about the argument that the nature of code/data (open or closed) should be the decisive factor, especially regarding deployment.