The greatest mistake in a technological revolution is confusing its first applications with its ultimate potential. Personally, I think we are currently making that mistake with AI. We imagine agents as digital assistants: scheduling meetings, summarizing information, conducting research and executing instructions. Useful, certainly. But largely automating work humans already do. The more consequential opportunity may be less visible: machines paying machines. Imagine an AI agent that needs specialized information to complete a task. It discovers a service, evaluates the price, pays a fraction of a cent, retrieves the information and continues working. No subscription or negotiation and no human approving every purchase. Now multiply that by millions of agents and billions of interactions. Purchasing intelligence, accessing proprietary data, acquiring computing power, paying for specialized services - all within predefined permissions, spending limits and security controls. This is not science fiction. The infrastructure is already emerging. Micropayments may finally become practical, not simply because moving value is becoming cheaper, but because machines can make purchasing decisions programmatically. The opportunity is not merely cheaper payments. It is making previously uneconomic transactions possible. Not every agent will need blockchain. Within a single company, internal ledgers may be sufficient. But when independent agents transact across companies, currencies and networks, open, programmable settlement becomes increasingly valuable. Blockchain will not win because it is blockchain. It will win where it delivers better economics, greater interoperability and less friction. Especially where no single company controls the network. As @johnjdagostino once said, and as I often repeat: AI is scalable intelligence and blockchain is scalable truth. Neither guarantees perfect outcomes. But together, they offer the potential to scale economic activity across networks that do not depend on a single intermediary. Their convergence is agentic commerce; and its evolution may surprise us. Machines can evaluate price, quality and performance against defined objectives. People also bring emotion, habit, preference and expectations of protection. Consumers will increasingly delegate purchases to agents. But trust must be earned through authorization, security and accountability. Trust is not an obstacle to agentic commerce. It is the infrastructure that makes it possible. I learned something important during my years in media. New technologies initially imitate old behaviors. Early television borrowed from radio and early websites resembled newspapers. Only later did we discover what those technologies truly made possible. The Internet did not transform the world by putting newspapers on screens. It transformed the world by creating entirely new forms of communication, participation and commerce. We must distinguish between what technology can demonstrate, what markets are ready to adopt and what entirely new markets it can create. These are rarely the same thing. I believe AI and blockchain can do something equally profound for economic activity. As intelligence becomes more accessible, economic agency becomes programmable and trust becomes increasingly important, the cost of coordinating and exchanging value can fall dramatically. New businesses become viable. New services become affordable. Entirely new markets become possible. And the first wave may be largely invisible. We are not merely teaching machines to participate in the economy we built. We are creating the conditions for an entirely new economy to emerge. One in which intelligence can act, value can move and trust can scale. This is why we built @paybox here at @moonpay: to enable people to imagine a better tomorrow using agentic commerce without the limitations of yesterday.