a16z: patching can't keep up with AI-speed attacks, rebuild defense in depth
AccBalanced · x · 2026-11-09
a16z's Asaf Ezrachi and Steph Zhang argue that AI has collapsed the offensive bottleneck of turning bugs into working exploits, so no one can patch in time. Their rebuilt defense-in-depth framework: 1) get systems off the open internet, 2) segment what must stay connected, 3) block attacks in the network at machine speed, 4) remove whole vulnerability classes as the most durable fix. The accompanying chart maps five gates an attack must pass (Reach/Bug/Run/Spread/Recover), noting patch speed only helps at gate 2.
More from Safety
- Apple Lockdown Mode now allows per-contact iMessage exceptions; non-secure Wi-Fi ban remains the pain point — matthew_d_green · 2026-11-09
- Nvidia board launches independent probe, says no evidence Jensen leaked secrets — GavinSBaker · 2026-11-09
- Researcher details flaw enabling remote unlock of X-brand cars worldwide — infosecau · 2026-11-09
- Floridi et al. paper: intelligence, black box, hallucination metaphors skew AI law — ArtificialOther · 2026-11-09
- Tencent AI chief Yao Shunyu warns AI will soon master biology, unleashing biosecurity risks — pstAsiatech · 2026-11-09
- Running vulnerability scanners on global legislation to surface legal loopholes — SnooPeripherals5313 · 2026-11-09