The period of easily finding greenfield opportunities in AI has passed. The market is now saturated with startups tackling obvious workflows in major verticals, and the pace of foundational model improvement has plateaued, making it harder to find a unique edge.
The speakers argue that the most successful companies are built on 'secrets'—ideas that are non-obvious and often considered crazy or dangerous by the majority. They use historical examples like DoorDash's software-only approach vs. the 'full stack' trend and Coinbase's pro-regulation stance to illustrate this principle.
New technology platforms (like the smartphone and AI) render old laws and business models obsolete. Startups like Uber and Lyft succeeded by operating in the gray area of outdated taxi regulations, while the current debate around open banking presents a similar opportunity for fintech.
This theme highlights founders who pursue incredibly ambitious, long-term visions that seem like science fiction and face extreme skepticism from the establishment. OpenAI and SpaceX are cited as prime examples, where founders persisted for years against negative press and expert disbelief.
AI, specifically code generation, is fundamentally changing enterprise software competition. Startups like Campfire can now build comprehensive suites to challenge incumbents like NetSuite by automating complex data integration and dramatically reducing switching costs for customers.
Keep pulling the thread on Bob McGrew.