The critical thing now is to design a sensible system, and agree the benchmarks that will actually offer real oversight, and ensure that oversight is tied to delivering AI that works in the interests of everyone. Let's get started right away.

→ View original post on X — @mustafasuleyman