Foundation Models Aren’t for the Children

In this meeting, Shu channels his inner David Ogilvy, pulling wisdom from the father of advertising as we unpack how bold individuals can often push progress faster than committees —while recognizing the moments when group consensus still matters. We pivot to the New York City mayoral race, and the strange new world of AI toys that occasionally say the wrong things to kids.  Our main discussion digs into foundation models in machine learning, the expanding economic frontier these models are creating, the staggering costs behind building them, and why thoughtful data collection matters more than ever. It’s a conversation about hype, hard realities, and the discipline needed to build the future responsibly.