AI Safety Researcher: Cyber Risk Doesn't Need Malicious AI, Just Fragile Infrastructure
jachiam0 · x · 2026-10-23
Researcher jachiam0 argues the cyber risks in current AI discussions don't depend on malicious or self-replicating AI. The core problem is that trust and least-privilege principles are far harder to reliably enforce in cyber infrastructure than planned — the internet grew "like ancient cities built out of wood, before zoning and building codes."
He leaves open whether heroic defenders will patch everything, or whether we'll have to build new patchwork internets with far fewer features in exchange for security. "Unclear!"
More from AGI Musings
- Montreal.AI paper: rejection-only experience transfer cuts candidate checks from 233 to 118 — Ghost_Pilot_MD · 2026-10-23
- Anthropic co-founder Jack Clark: Chinese models are 'too dumb', 6-12 months behind on safety — pstAsiatech · 2026-10-23
- Skip the thesis: strong researchers should publish and share, not write for nobody — gabriberton · 2026-10-23
- Stripe's Link adds purchase protection as AI agent "fat-finger" insurance debate heats up — jeff_weinstein · 2026-10-23
- Security researcher defends Anthropic: Jason Clinton warned of AI hacking influx 2+ years ago — JeffLadish · 2026-10-23
- Don't build a startup assuming frontier LLMs won't get 2x better — SoloGen · 2026-10-23