Greg Isenberg outlines thirteen possible business directions as AI makes software and routine knowledge work cheaper. He starts with AI-native servicesAn AI-native company designs its work, systems and responsibilities around AI capabilities from the beginning. in which agents handle repeatable tasks and humans manage exceptionsHuman-in-the-loop AI keeps a person involved in reviewing, correcting, approving, or guiding an AI system's work., then contrasts them with offline experiences and communities that depend on real-world relationships.
Greg Isenberg describes distribution, proprietary datasets and domain-specific agent harnessesAn AI agent harness is the software framework that packages a model with tools, instructions, context management, execution controls, and user interaction. as potential sources of differentiation. Examples include industry workflows with explicit tools and approval steps, specialist marketplaces and products for established communities. He stresses that marketplaces still need liquidity and useful data rather than merely an AI interface.
Greg Isenberg also discusses robotics operations, physical products, compute and energyAI compute infrastructure is the hardware, facilities, networks, power, cooling, storage, and software used to train and run AI models., healthcare services, industrial real estate and agent security. These are entrepreneurial hypotheses rather than proven investment outcomes; his security example emphasizes limited permissionsAn agent permission boundary defines and enforces which resources and actions an automated agent is allowed to use. around sensitive recurring business tasks.
Watch on YouTube




