Andrew Lilley Brinker on Mythryl: “Regarding vulns decaying exponentially: AI pushes the reasonable…”

Regarding vulns decaying exponentially: AI pushes the reasonable limit you can climb that asymptote closer and closer to 0 by driving cost per discovery downward.

Andrew Lilley Brinker@alilleybrinker.com

1. Formal methods exist and are becoming much easier / cheaper with AI. We can prove strong properties up to the limit of the formal system. 2. Vulnerabilities decay exponentially (see link), and we can get asymptotically assured about the security of individual software.

security.googleblog.comEliminating Memory Safety Vulnerabilities at the SourcePosted by Jeff Vander Stoep - Android team, and Alex Rebert - Security Foundations Memory safety vulnerabilities remain a pervasive threa...

21/10/26, 7:14 pm0 reposts4 likes