In recent days, most were surprised about the reporting of Spectre and Metldown critical vulnerabilities in nearly all mainstream CPUs for the last 15 years are unfixable to a large extent, even through OS updates.
We are not surprised.
It was 2014, when Bruce Schneier, the World most recognized security expert - in reply to a declaration of Intel CEO that their chips had not been hacked - clearly stated that, after Snowden, we should be assuming all mainstream CPUs to be compromised in undetectable ways, due to design and supply chain complexities or state backdoors. See minute 32.40-34.00 of this video.
Nothing we can do about it? Nearly everyone seems to think so.
But, since 2014, at TRUSTLESS.AI and the Trustless Computing Association, we’ve been promoting an approach to CPU/chip design (and fabrication!) that removes upfront trust, and radically reduces complexity, and enables for an offline in-person privacy-respecting lawful access so states don’t have to backdoor it.
Also, our Kryptus SCuP architecture - the only secure CPU in the world publicly verifiable in HW and SW design according to the Head of Information Superiority of EDA - is immune from such kind of vulnerabilities "as the underlying core does not employ speculative execution".
But we go way beyond, implementing our Trustless Computing CivicFab fabrication oversight processes, that are well in excess in user trustworthiness than even NSA Trusted Foundry Program processes.