That's a user point of view. But I'm a developer. So, from my point of view, an API that gets cornered into compatibility/support status is a dead API, because in the mid-term you won't be able to use it for accessing the latest hardware features. In that context, what Apple is saying today is that OpenGL and OpenCL are dead (in other words, if some future GPU supports any new feature, you can expect it implemented in Metal, not in OpenGL nor OpenCL).
A very disappointing bit about this are the users who purchased a new Mac Pro. An expensive machine, theoretically optimized for OpenCL, with expensive dual GPUs, whose degree of support under Metal is still uncertain. I'm glad I decided to wait for later generations of the Mac Pro, when the new strategies from Apple get clearer.
Regarding Vulkan, it's still unborn, just preliminary implementations, but we see Microsoft pushing for DX, and now Apple pushing for Metal. Who's going to push for Vulkan? And it didn't arrive to a public status yet.
I've read very good reviews of the internal design of Metal, although didn't use it myself yet. If it's such a good API, maybe a good approach could be to write a compatibility layer from Metal to other OSs APIs (if that's feasible to some level), and write everything in Metal.
Because I develop for Mac, but also for other OSs. I know Apple wants us to target Apple products only, and the new API strategies from Apple clearly show their will, but sorry Apple, we'll still develop for other OSs too (you know, no OS stays forever, and well designed software surpases OSs lifetime).
