Microsoft's workforce shakeup, AI agents taking over analytics, and the changing value of expertise. Are category managers becoming more valuable or more replaceable?

Will AI agents do all our shopping, and who pays when they get it wrong? In this episode, we look at why AI's business impact still hasn't fully arrived, why teenagers are calling AI "bland," and how a 10-minute coding session with Cursor saved a family's Harry Potter audiobooks. We also cover Meta's new shopping agent, the new "deep review" agents, pre-trained forecasting models, and why JPMorgan is pushing back on agentic payments.

Would you connect your bank account to ChatGPT? Yves's answer sets the tone for an episode about how fast AI is moving from tool to teammate, and how little most of us have thought through the consequences.

Chelsea and Yves trace the jump from reactive chatbots to always-on agents that quietly monitor your inbox, your calendar and your spending and the eye-watering token economics driving it. Then the big one: if a bot does your shopping, nobody sees the ads, the bundles, the cross-sells or the $X.99 pricing. The digital shelf, as we know it, starts to disappear.

Along the way: who actually wins the agent race, whether fifteen years of emails can rebuild the colleague who just retired, what cheaper frontier models mean for anyone locked into one vendor, and a listener asking the question everyone's thinking is my company reading my AI chats, and am I automating away my own job?