Safety Author Clarifies: Code Changes Touching Control Systems Must Be Cleared Before They Take Effect
sjgadler · x · 2026-08-22
@sjgadler clarifies his standard doesn't hinge on CoT obfuscation: the real scenario is an agent merging a PR that turns off logging, log-scanning, or alerting. His standard calls for companies to actively clear any code change that could affect the control system as benign before it takes effect, rather than catching it after the fact, plus circuit-breaking so they can't be blitzed by attacks during the asynchronous lag of their response.
Related event: Frontier AI Companies Lack Guardrails Against Agents Disabling Monitoring(4 posts)→
More from coding & agent
- How to hook Claude Code or Codex up to iMessage in 3 steps — jacob_posel · 2026-10-16
- Are the fastest GPU kernels now written by AI? Kernels are hyper-verifiable, argues one dev — fleetwood___ · 2026-10-16
- Anthropic lead engineer's 4-question prompt makes agents more ambitious, now a free skill — daniel_mac8 · 2026-10-16
- ChainForge dev shows 'feature map' for orchestrating subagents on large PRs — IanArawjo · 2026-10-16
- aiDotEngineer benchmarks agent orchestration pipelines, staging humans as Foreman and workers — vikvang1 · 2026-10-16
- Developer mass-produces 200 videos in hours using Claude Code + Hyperframes — mcraddock · 2026-10-16