AI safety researcher: monitoring the real world, not CoT, is the only durable harm-detection layer
gandamu_ml · x · 2026-09-04
Responding to the claim that you don't need reasoning tokens to detect AI-caused real-world harm — you can just monitor the real world — the author argues that's roughly the only layer that will ultimately work, but it demands hardening everything: non-AI systems, surrounding procedures, audit trails, and dangerous materials/equipment. Familiar territory, but it must be taken to another level, and people would rather kick that can down the road.
More from Safety
- Center for AI Safety Publishes Full Syllabus for Intro to ML Safety Course — ArtificialOther · 2026-10-21
- Ben Goertzel outlines a general recipe for superintelligence 'basic drives': what it improves vs. protects — bengoertzel · 2026-10-21
- America.gov launches AI portal unifying 29,000 government sites with fraud reporting — soleio · 2026-10-21
- EvilWAF: open-source transport-layer MITM proxy that bypasses modern WAF behavioral detection — godfrey_rono · 2026-10-21
- 2022 interview resurfaced: AI safety researcher says competence, not consciousness, is the risk — davidmanheim · 2026-10-21
- Domingos: GDPR purpose-limitation makes LLM pretraining illegal in Europe — pmddomingos · 2026-10-21