Because most of these CPU level attacks require physical access, the intel one fro a couple f weeks ago could be done over a LAN but the "unpatchable" element was with physical access.
that T2 chip life.
No, you DO NOT need PHYSICAL access, you need LOCAL access which range from tricking the user into running an executable to buffer overflows that allow arbitrary remote (relative to the attacker, which would be local relative to the target) code execution.
Wasn’t the intel vulnerability on doable on boot? So GL with that method.
This is the AMD exploit thread and the side-channel exploits for both AMD and Intel require some form of local code execution to work. For the Intel boot-time thing, there isn't much that can be done about security once you are at the point where hands-on access to the hardware within the first seconds of power-up is required.
I know I was just referencing the intel one as an example that most of these vulnerabilities aren't threatening to the majority of people here, they're a danger for people who run servers etc as most of the data people want from regular PCs are things like card numbers which can be attained in an easier manner.
I don't see anything about this exploit requiring any of those things. I don't remember Spectre requiring any of those things either. All that is required from what I can tell is being able to execute code on the target machine, E.g. through malicious code on a website that is visited, or from within a VM on a cloud hosting server.
Where do you see any mention of them modifying the kernel? And even if they did, what about someone attacking a cloud hosting server by renting a VM and running whatever they want (possibly including a custom kernel) on it?
AMD is still better since most of its flaws are less severe than Intel's and patches aren't as bad performance-wise either. So, not only does Intel have catch-up to do on core count and power-efficiency, it also has more catching-up to do on hardware fixes and mitigation.
Those that thought Intel was the only line vulnerable are the same nut jobs that believe Star Trek was filmed "On Location".
Install NoScript or similar, then look at how much stuff from other domains the pages you are regularly going to are loading. Every single one of those may have some degree of access to what you are currently doing, be it from being an active element on the page you are viewing or receiving metadata telling it what you are doing from the page. That's how searching for one thing on one site may "mysteriously" get reflected in ads you see on a seemingly unrelated site.
Facebook
Twitter
Instagram