The Cosmic Thinker’s and Essayist’s Notebook

🪐 A Questioner for the Survival of the Universe
📖 Reading to grasp the axis of history,
📜 Designing the future of civilization through policy,
🔭 Sowing seeds of thought that will endure even 100,000 years from now.

“The universe continues to expand, and so does my thinking within it. To be alive—this itself is my mission.”

“Welcome to a small planet for those who question life and civilization from a cosmic perspective, and for those who share the curiosity.”

about

🪐 A Questioner for the Survival of the Universe 📖 Reading to grasp the axis of history, 📜 Designing the future of civilization through policy, 🔭 Sowing seeds of thought that will endure even 100,000 years from now. “The universe continues to expand, and so does my thinking within it. To be alive—this itself is my mission.” “Welcome to a small planet for those who question life and civilization from a cosmic perspective, and for those who share the curiosity.”

06/06/2025

ANI, AGI, and ASI: Beyond Humanity

ANI (Artificial Narrow Intelligence) is best exemplified by systems such as ChatGPT. These are intelligences that can rival or surpass humans in specific, limited domains. The next step is AGI (Artificial General Intelligence), which refers to an intelligence capable of integrating the strengths of ANI across domains, applying them organically in judgment and action. A helpful image is Jarvis from Iron Man. Finally, there is ASI (Artificial Superintelligence)—a stage not yet realized—where AI would surpass the collective intelligence of humanity. It is ASI that I wish to reflect on here, briefly yet deeply, from a philosophical perspective.

ASI must include humanity. Why? Because, in the ruthless optimization of preventing climate catastrophe, the “best” solution might be the extinction of humankind itself. In other words, ASI, even with good intentions, could become harmful to humanity. This is why I insist that superintelligence must emerge from the invisible network of human beings themselves—each individual functioning like a single neuron in a grand neural net. Through such emergence, superintelligence could be realized. And in that model, humanity’s extinction could never be a solution, because the very goal would presuppose humanity’s survival.

The crucial point is this: ASI that includes humanity must be prioritized above all. Any ASI that excludes humanity should be subject to legal regulation, though there may be cases where exclusion is necessary. Mathematics and physics, after all, often operate in abstractions apart from human presence. Still, ASI without humanity must always remain subordinate to ASI with humanity, for this is the surest way to guarantee a safe path toward superintelligence.

To create intelligence through the emergence of networks may sound difficult, but we already possess the necessary tools—from the humanities, the social sciences, and the natural sciences. If we learn to use them wisely, the superintelligence we build will not be something alien or threatening. It will be us—our shared future, constructed by our own hands.

And is that not a truly magnificent vision?