formal verification is one way, minimizing the attack surface is another, zero trust is yet another
and more
defence in depth or not making it
and more
defence in depth or not making it
4/ What to do instead:View quoted post on X
Spend the AI dividend on rebuilding software on mechanisms that kill whole vuln classes.
Treat whitehat finds as intelligence, not a patching queue, and patch what is actually being exploited.
And assume breach: defense in depth for dense-in-depth code.