... except that it still pretty much requires Nvidia hardware. Maybe not for edge inference, but even inference at scale (ie. say at companies, or governments)

... except that it still pretty much requires Nvidia hardware. Maybe not for edge inference, but even inference at scale (ie. say at companies, or governments) will still require it.