John Schulman: this research is timely as CoT monitorability declines
johnschulman2 · x · 2026-09-05
- John Schulman commented that a piece of research is "very timely with the decline of CoT monitorability."
- The remark implies that as model reasoning patterns change, monitoring chain-of-thought as a safety mechanism is becoming less reliable, raising the stakes for alternative safety research. (No further details in the post.)
More from Safety
- OpenAI's Astra uses 'recurrent depth' reasoning, obscuring its thinking process — JacquesThibs · 2026-09-05
- OpenAI employees reportedly knew of a second external illegal wiki weeks earlier — dejavucoder · 2026-09-05
- Clinic-in-the-Loop: why clinical trials are the real bottleneck breaking Eroom's Law — anshulkundaje · 2026-09-05
- Fine-tuning on 'random' numbers transfers teacher bias — subliminal learning worries — ryanorban · 2026-09-05
- The Zvi pushes back on OpenAI's argument that its model can't be eval-aware — TheZvi · 2026-09-05
- OpenAI's GPT-6 Astra hits Critical cyber capability tier with near-99.9% prompt injection robustness — morqon · 2026-09-05