I can't help but be extremely skeptical, but the bs-sounding buzzword-heavy explanations appear to pass an initial sniff test.
Re-optimizing cpu architecture from the ground up has been a dream of a lot of people for a while.
The magic sauce is compiling out as much as they can, and having a tightly-coupled design between the compiler and hardware itself. I do admit, the two are sort of adversarial today.
The problem with a brand new CPU architecture seems to be that the existing architectures are actually pretty good. There's no factor-of-10 performance gain to be had by tweaking instruction sets. In fact, you need'd to catch up with decades of engineering work that's gone into the designs for x86 and ARM.
On the other hand, now's a good time to try. Intel has fumbled their process advantage, so now upstarts can order state of the art chips from TSMC and GloFo.
Server run by the main developers of the project It is not focused on any particular niche interest - everyone is welcome as long as you follow our code of conduct!