Two AI CEOs argued about how fast models should improve. Nobody argued about how fast companies should adopt them.
Amodei's safeguard is a gate with a badge. Ahuja's is a duty of care with no evaluator. Both are called safeguards. They are not the same object, and the gap between them is where most companies work.
Two failure numbers, two different rooms
Output that looks finished, survives a glance, and quietly does nothing. What a dead button in a client tracker says about what AI pilots should actually be measuring.
The risk with AI at work isn't that it's wrong. It's that it's wrong quietly.
Output that looks finished, survives a glance, and quietly does nothing. What a dead button in a client tracker says about what AI pilots should actually be measuring.