I imagine running ARM code on ARM cores would be more efficient than running them on x86 with or without compatibility extensions, and vice versa. Instead of a

I imagine running ARM code on ARM cores would be more efficient than running them on x86 with or without compatibility extensions, and vice versa. Instead of always playing catch-up platform-wise for compatibility, just embed the ISA platforms into the chips themselves. As for why: because I'm not abandoning x86 any time soon and this would ease ARM transition. I also think it would set a foundation for hardware-based secure virtualization across platforms.