This site uses cookies. By continuing to browse the site you are agreeing to our use of cookies.
Computer scientist Scott Aaronson says sources tell him several AI labs have begun privately testing whether their newest models can break important cryptographic protocols, an omission he flagged after
