TEXTS

   Bill Gates is right that some parts of human life will have to be protected from artificial intelligence. He calls them Human Reserved. But the phrase still imagines humanity as a protected enclave inside a giant machine economy. I think the opposite is more likely.

​

1. The Human Economy Will Be Bigger Than the Machine Economy

​

   The deepest economic effect of Super Intelligence is not that it replaces labor. It is that it makes complexity cheap. Law, medicine, engineering, software, finance, research and management are expensive today because intelligence capable of handling complexity is scarce. Super Intelligence destroys that scarcity. It may do to intelligence what industrial technology did to agriculture: increase output enormously while reducing its share of economic value.

   The machine economy may eventually perform most necessary work and still account for less value than the Human Only economy around it. Once complexity becomes abundant, money moves toward what cannot be multiplied: reputation, attention, beauty, authenticity, rare places, relationships, status and human achievement. This is not a romantic preference. It is basic human behavior. From the sandbox onward, people compare, compete, signal and try to distinguish themselves from the person next to them. Once necessity becomes cheap, that instinct becomes the economy.

   That is why Human Only is not a reservation for obsolete workers. It may become the premium layer of civilization. Super Intelligence will make complexity abundant; humans will keep manufacturing scarcity. The old feudal lord collected rent from scarce land. The future rentier may collect it from scarce attention, scarce access, scarce reputation or simply from being unmistakably human.

​

2. Democracy as We Know It Will Not Survive Super Intelligence

​

   The political problem is not that Super Intelligence will be smarter than voters. The problem is that unrestricted access to it may be incompatible with survival. A system capable of designing pathogens, weapons, cyberattacks or autonomous strategies cannot safely be available to everyone. There may be powerful AI for everyone, but there cannot be unrestricted Super Intelligence for everyone.

   The moment access becomes restricted, democracy changes. Someone must decide who may ask the dangerous questions and who may give consequential tasks. That group could be elected, appointed or selected through some new meritocracy. Perhaps the future elite will emerge through something like Olympiads in the ability to formulate the best questions for Super Intelligence. We do not know. But functionally it becomes a priesthood: a caste standing between ordinary humanity and an intelligence ordinary humanity is not allowed to command directly. The medieval priest claimed to speak with God. The future priesthood will have one advantage: its god will answer.

   A single global Super Intelligence might be safer, but it is hard to imagine the United States, China or any other major power surrendering its most important strategic capability to a world government. Nor can one state stop while believing another may continue. The likely outcome is therefore several competing Super Intelligences, each surrounded by its own restricted caste of access.

​

3. Super Intelligence Will Be Nationalized and the Human Economy May Become More Private

​

   Genuine Super Intelligence is also unlikely to remain private property. A company may create it, but once it becomes more strategically important than a nuclear arsenal and more productive than an entire scientific establishment, ordinary ownership starts to look absurd. Its founder may eventually discover that even he is no longer allowed unrestricted access to the system he built.

   Safety is only one reason. The other is economic. If a private owner can use Super Intelligence to discover new medicines, new forms of energy, new materials and fundamental technologies, society is unlikely to accept that one corporation should extract monopoly rents from an unlimited machine for scientific progress. At some point, the system and at least part of what it discovers will be treated as strategic public infrastructure. Call it regulation if you like. In substance, it begins to look like nationalization.

Paradoxically, that does not mean the whole economy becomes less private. It may mean the opposite. If the dangerous machine core is centralized, the vast Human Only economy around it can become more private, more competitive and perhaps less regulated than today. The state controls the one thing too powerful to distribute freely; people are left to compete over everything human beings actually care about.

   That, to me, is the world Gates has not taken far enough. Super Intelligence may do most of the work, while humans create most of the value. It may force a less democratic system around access to intelligence, while making ordinary human life more private and more economically important. The machine will make complexity cheap. Humans will make scarcity expensive. And the future, strangely enough, may become more human because of Super Intelligence.