No Priors Ep. 128 | With DeepLearning.AI Founder Andrew Ng

In this episode of No Priors, Andrew Ng discusses the future of AI, emphasizing multiple vectors of progress beyond just scale, including agentic workflows and multimodal models. He clarifies his definition of "agentic AI" and addresses the obstacles to implementing true AI agents, highlighting the critical need for talent capable of systematic error analysis. The conversation explores the strongest examples of agency in AI, particularly in coding agents, and contrasts "vibe coding" with AI-assisted coding, which he sees as rapid engineering. Ng also shares insights on how AI is changing startups by accelerating the speed of coding, thus making product management the bottleneck. The discussion further covers the characteristics of successful founders in the age of rapidly changing AI technology, the importance of hard work and customer obsession, and the potential for AI to empower individuals and small teams.