Building easy AI apps is a bad idea. Here's 8 better bets

The hosts argue that the best AI opportunities over the next five years are not in generic software, content, or coding tools, but in the physical world: drugs, food, warfare, housing, materials, autonomy, weather and oceans, and human body hardware. Their core thesis is that "AI in the world of things" will create more durable moats than building easy AI apps, because those markets combine hard engineering, regulation, manufacturing, and real-world distribution.

They point to examples already emerging: AI-assisted drug discovery and cancer treatment, genetically improved crops, drone warfare and directed-energy systems, factory-built homes, self-driving systems for machines, and prosthetics that respond to neural intent. The hosts repeatedly frame these as massive industries where the technology window is open and where a few smart teams can build something with unusually high impact.

The conversation also broadens into a political and historical argument: America’s past advantage came from resources, values, and industrial capacity, and the next era could resemble another industrial revolution if the country avoids self-sabotage. They close by urging ambitious people to think like investors, widen their field of view, and move toward the biggest bodies of water rather than staying in familiar lanes.