It's not mentioned in the article, but this hardware architecture may be better suited for brute-forcing password/crypto/key cracking than traditional computing architectures.

Is there a paper on that? I’m not I see how.