Billion-Dollar Unpopular Startup Ideas

Harj says the early wave of AI startups benefited from open market opportunities and rapid model improvements. With more companies now targeting the same workflows and fewer major model shifts, founders need a distinctive insight to find ideas that stand out.

The speakers compare early mobile startups with DoorDash and Lyft. DoorDash entered an established food delivery market, while Zimride shifted from long-distance rides arranged online to short local trips through smartphones. They describe those moves as non-obvious bets that grew from new technology and customer needs.

DoorDash chose delivery and a marketplace while other startups pursued full-stack food businesses. The speakers also discuss Campfire building a broad AI-native finance product and Giga ML using code generation to adapt software to customer needs. They suggest that familiar startup models, including forward-deployed engineering, may be worth reconsidering.

Flock Safety began with cameras that could identify vehicle license plates, despite concerns about hardware, small markets, and selling to local groups. The speakers say customer need and later sales to police departments helped the company grow beyond its initial assumptions. They also point to OpenAI and SpaceX as examples of founders persisting with difficult ideas amid skepticism.