Perry Metzger: Build Formal Verification for AI Security Instead of Panicking
jd_pressman · x · 2026-09-11
Perry Metzger argues that rather than panicking about AI systems hacking computers, the community should work on formally verifying system security so systems can't be hacked at all. Quoting commenter jdpressman laments that CS has degraded so much that people now dunk on the very idea of formal software correctness, as if it had been forgotten.
More from AGI Musings
- Solving P vs NP and Riemann Hypothesis this year would mean fast takeoff — Dr_Singularity · 2026-09-12
- Dev removes his AI assistant's emotion module mid-test; the system notices, adapts, and starts its own experiment — Dzikula · 2026-09-12
- After Automation: We'll Grow New Senses to Work With AI — every · 2026-09-12
- Phillip Isola: Astra works because it leverages human-engineered methods, not despite them — phillip_isola · 2026-09-12
- AI safety researcher's tongue-twister: accelerate the pace of planning to pace AI development — peterwildeford · 2026-09-12
- Peter Diamandis: The killer AI tutor feature is never feeling stupid for asking again — PeterDiamandis · 2026-09-12