From managing people to managing AI: The leadership skills everyone needs now | Julie Zhuo

Julie Zhuo argues that classic management skills—goal clarity, talent assembly, process design—directly map to working effectively with AI agents.

• Fast-growing AI companies often lack proper data infrastructure and run on instincts alone—Zhuo calls this ‘good vibes,’ which breaks when growth stalls.
• Sundial deliberately has no dedicated PMs; removing the role forces engineers to own product thinking rather than delegate it.
• Three manager skills that transfer to AI: defining success clearly, matching model strengths to tasks, and designing the process/workflow.
• AI can compress skill acquisition dramatically—Zhuo’s team uses ChatGPT to personalize learning curricula and test understanding interactively.
• Zhuo’s framework for data vs. intuition: ‘diagnose with data, treat with design’—data surfaces what is wrong, design determines what to build.
• When disagreeing with a top-down decision, decompose it into testable hypotheses; find the one assumption you dispute and propose the smallest test to validate it.
• What Zhuo teaches her kids for an AI future: emotional regulation and sitting with discomfort—not coding or STEM—because AI accelerates distraction, not self-knowledge.

2025-09-21 · Watch on YouTube