I think the challenge here is that you need buy-in from industry pros to make this really worth it. The platform can be best-in-class from a hardware/API persp

I think the challenge here is that you need buy-in from industry pros to make this really worth it. The platform can be best-in-class from a hardware/API perspective, but if no tools exist using in it, why am I going to give up my CUDA-based workflow to spend way more effort re-inventing the wheel on mac? I could imagine an open platform supplanting CUDA in the long term from a cost-optimization angle, especially on server workloads, but I just don't see serious work moving from one closed platform to a smaller closed platform.