“Shishir Mehrotra believes that for a large company like Meta, not making large, aggressive bets in a critical new field like AI is a worse risk than making bets that might fail.”