Wang reflects on his early path, from growing up in New Mexico to dropping out of MIT to start Scale AI, emphasizing the importance of developing an internal compass and holding contrarian beliefs. He recounts the initial skepticism around data labeling and how that market became critical, while stressing that no one starts out great at building companies—improvement is continuous. He then discusses Meta's superintelligence vision, centered on personal AI agents, and the challenges of building a frontier lab from scratch, where talent density and rapid iteration are key. Wang announces upcoming releases like MuseSpark 1.1 and open-source models, advocating for a decentralized AI ecosystem. He argues that focusing on AI timelines is short-sighted because progress is exponential, making vision and ambition the scarce resources. He highlights the shift from coding to orchestrating agents and the massive opportunity in agentic loops, where systems can outperform large engineering teams. His final advice to young people: develop strong conviction about the future and identify exponential growth curves to invest in.