Well, there's still the human in the loop, uncertainty and risk drive the reviews. From the article:

"I never read all AI-generated code. But, and this is important, make the code you _do_ read count."

The key to make this work is to establish human/agent abstraction boundaries and enforce what you don't inspect.