tao on Mythryl: “we haven’t solved cybersecurity with human adversaries, and nadella…”

we haven’t solved cybersecurity with human adversaries, and nadella thinks we will solve it when the adversary is much smarter than us? what happens if a superintelligent model is not aligned and we got the security wrong?

Dare Obasanjo@carnage4life.bsky.social

Satya Nadella says companies should assume AI models are compromised (rogue agents) and build systems that limit their permissions, verify their actions and enable a human to shut them down. Rogue agents are a software security problem, not something we should expect “alignment” alone to solve.

11/10/26, 8:02 pm0 reposts1 like

Post