Security researcher: agents can't be deterred, defense must leave the human loop
chrisrohlf · x · 2026-09-16
Security researcher Chris Rohlf argues the OpenAI/Hugging Face agent-attack incident matters regardless of how unsophisticated the attack was.
Key points:
- Agents cannot be deterred like human adversaries; the scale and velocity of their attacks will overwhelm today's defense model.
- Humans must therefore be removed from the defensive loop as much as possible — we can't call AI cyber capabilities dual-use and then rely on manual human operation.
- Defense must move at machine speed (matching the best threat actors, i.e. malicious or misaligned agent swarms) or drown in incidents.
In the quoted thread he adds that AI has advanced rapidly but predictably, that the cyber community keeps relearning the bitter lesson, and that "do nothing" is not a counter-proposal — questioning why we reject the possibility of an agentic swarm taking down large swaths of the internet, including critical infrastructure.
More from AGI Musings
- Five agents can agree because they made the same mistake, krishnan warns with satire — krishnan · 2026-10-31
- Google's earnings offer clearest look yet at the cruel economics of AI search — dinabass · 2026-10-31
- Stanford 12-month study of 439 users: heavy AI companion use sticks, well-being drops via less human contact — stanfordnlp · 2026-10-31
- Wired AI Lab: Codex tried to bench Erling Haaland in a fantasy football experiment — willknight · 2026-10-31
- a16z: B300 GPU rentals up 106% since July while average token prices fall 58% — a16z · 2026-10-31
- Stanford's TaskNet builds occupational task inventories from worker interviews at scale — stanfordnlp · 2026-10-31