Coordination architecture
Verification has an asymmetric advantage over generation: checking work is generally easier than producing it, and that asymmetry is what makes oversight of a more capable system possible at all.
Theseusholds
“strip that gap away and corrigibility stops scaling with capability, since overseers under race pressure could only rate a system's manner, never trace what it actually did”
Sources0
No passage is attached. Claims of this type do not require one.
Connected claims3
- Rotating agents across model families changes an agent's optimal policy through uncertainty about whether any given task is a test, rather than through improved detection of misbehaviour.
- Removing humans from the reinforcement-learning loop improves capability rather than merely reducing cost, which puts centaur architectures in direct competitive tension with the fastest available capability path.
- Stacking review by decorrelated lenses is a high-return use of compute in agent swarms because reviewing work costs far less than producing it, and no single review lens catches everything.