If the vulnerability is found before 2018 and there are still a significant amount of people using a very old operating system that doesn't support UEFI, then the severity of the vulnerability is considerable. Considering that Windows 10's adoption rate is higher than any other version and that any future computer will require the latest OS and UEFI to run, I highly doubt this potential/theoretical vulnerability will be of concern to the majority of consumers (especially gamers - who most are already using Windows 10 or the latest version of the OS of their choice).
Again, it's still a theoretical vulnerability, and many governement systems/bank system/etc run on COBOL, windows 3.1, and MS-DOS, so that's also a massive concern, and even with the UEFI protection, it's still theoretically possible to be able to do it
This is just speculation here, but I heavily doubt governments will buy the latest technology to run such ancient operating systems, especially since they're unsupported and therefore drivers don't exist for modern CPUs running on them OS'. It's more likely they'll be running an emulator on a modern OS if they're going to buy the latest hardware. UEFI can calculate checksums etc from a dedicated chip on the motherboard as far as I'm aware. You would need to attack vulnerabilities in both the CPU (which could be patched without notice in newer generations or new produced batches of existing generations) plus the motherboard's firmware and BIOS. While it may be a theoretical vulnerability I heavily doubt that it will be of concern to the majority of people - especially as time progresses and the incentive decreases. Nobody really should be worrying about this at all, even if the theoretical vulnerability becomes a real discovered vulnerability
As i was saying, it's a theoretical vulnerability, and as the article says, it's unlikely to be a vulnerability within our lifetime, but it exists nonetheless, it's just there
