AI safety lags behind AI power

Anthropic's CEO admits safety depends on understanding AI reasoning, but current evidence is unsettling.

Also today

Today's top five AI stories, one line each, at 7am Eastern. Free. One-click unsubscribe.