@rohinmshah
AGI Safety & Alignment @ Google DeepMind
AI
An essay coauthor challenges the idea that chain-of-thought reasoning will inevitably disappear, arguing that model designers can choose to preserve its monitorability.
AI
