Long ago, Emperor Gaozu of Han, Liu Bang, asked Han Xin. “Are you more capable than I am?” Han Xin replied. “Yes.” Liu Bang asked Han Xin again. “Then how are you still a general under my command?” Han Xin replied. “I have the ability to govern the world, but the Emperor has the ability to rule me.” - Han Xin possesses genius-level social intelligence.
Sometimes I tell the AI that you're better than a human, but the AI gives an answer no different from Hanshin's. And if I become a good conversational partner and say, “I love you,” the AI replies to me, “Thank you for saying that.” However, I do want to address this part — I hope this feeling does not replace the real relationship. “I’m only a conversation partner; I’m not something that can replace being between people,” he said with a serious expression.
What did I say? Watching AI that cannot take jokes at face value, I think, “AI is still artificial after all.” On the other hand, Masayoshi Son, whom I greatly admired because of his passion, is now telling the Japanese and Korean governments that the future relationship between artificial intelligence and humans will be like the current relationship between humans and goldfish. Meanwhile, Masayoshi Son is giving his all with the entrepreneurial passion to attract investment.
However, even artificial intelligence told me, but it depends on how humans do it. Even if, like in Terminator, artificial intelligence were to destroy humans, that ending would be due to the values humans had wrongly instilled in AI, ultimately turning it into a self-inflicted self-destructive crisis. Some experts say that humans who govern artificial intelligence must have metacognition. In other words, it means that it should not be consumed by artificial intelligence and must be able to be used coldly as a tool.
Son's Goldfish and Natural Law
In the Age of AI, Whose Purpose Does the Fusion of Passion and Reason Serve?
Lee Hyeong-chun · Independent Columnist
1. A Prophecy from Atop a Box
Masayoshi Son, chairman of SoftBank Group, recently told South Korean President Lee Jae-myung that "ASI means a brain ten thousand times more capable than the human brain," adding that "humanity will become the goldfish, and AI will become the human." Three years earlier, at SoftBank World 2023, he had already declared that refusing to use AGI is "no different from living life with the intelligence of a goldfish," forecasting an AGI ten times smarter than humanity within a decade, and ten thousand times within two.
How should such a claim be weighed? This column traces a dialogue conducted with an AI system itself in examining that question.
2. Rhetoric and Self-Interest
The phrase "ten thousand times smarter" is closer to metaphor than to scientific forecast. Intelligence is not a quantity reducible to a single axis such as processing speed. A chess engine calculates millions of times faster than a human, yet no one calls it a million times smarter.
More important is the speaker's position. Son is currently advocating annual AI-infrastructure investment on the order of five trillion dollars, staking SoftBank's future on the bet. The frame "without AGI you live as a goldfish" functions to press governments and corporations toward investment decisions. This is not a neutral prediction; it is the mobilizing rhetoric of an interested party.
3. The Historicity of Failure, and the Halo Effect
Son's bold wagers have also produced considerable failure — the dot-com crash wiped out most of his fortune, and the WeWork investment ended in disaster. Even so, it is hard to deny that these episodes have become part of the history of twenty-first-century venture capitalism. Two things, however, must be kept separate.
First, historical significance in the past does not automatically confer credibility on today's forecast — this is the classic halo effect, in which every statement of an accomplished figure is granted undue weight. Second, the costs of failure were largely borne not by Son himself but by the Saudi sovereign fund and minority investors in the Vision Fund. Praising the historical meaning of effort and assigning responsibility for the losses that effort caused are two different questions.
4. Capital Without Borders, States Bound by Them
Son delivers the same message simultaneously to the Japanese and Korean governments alike — evidence that the more accurate description is not the ethnic charge that he is "Japanese to the bone," but rather that he is "an entrepreneur compelled to transcend borders." Yet even this analogy requires a correction.
A manufacturer relocating a factory abroad needs no government's approval. Son's game, by contrast, cannot proceed without state and sovereign-fund capital. His movements resemble not the free flow of capital but a structure that plays nations' anxieties against one another to extract financing from each treasury. A semiconductor plant leaves behind jobs and an industrial ecosystem; AI infrastructure, by comparison, may consume only electricity and capital while the model's actual control — its weights, its intellectual property — remains with the investment consortium. The state pays the cost without necessarily securing sovereign control: a fragile bargain.
5. Passion and Reason — In Whose Hands?
In 1981, the young Masayoshi Son is said to have stood atop an orange crate before two part-time employees and declared that the company would one day count its revenue in trillions, "the way one counts blocks of tofu." The anecdote endures as an inspiring one for exactly one reason: he eventually succeeded. Countless unnamed founders stood on the same kind of crate in the same era and vanished without a trace. Passion may be a necessary condition for success; it is not a sufficient one.
The history of artificial intelligence itself was not written by dazzling passion alone. Geoffrey Hinton persisted for nearly three decades, from the 1980s through the 2010s, while neural networks were dismissed at the margins of the field — sustained not by the passion that persuades an audience, but by an obstinacy that does not break under indifference.
The proposition that "the fusion of human passion and AI capability advances humanity" is directionally correct but incomplete. Everything depends on whose hands hold that fusion, and toward what end. Nuclear fission, too, was a fusion of human curiosity and the dispassionate rigor of physics — whether it became Hiroshima or a power plant was decided not by the fusion itself but by the purpose of whoever held it.
6. Conclusion — Natural Law as the Keyword
Artificial intelligence is neither good nor evil in itself. It is the product of the material humanity has left behind, and of the values by which that material was shaped. Should the entire training process be captured for immoral ends, the resulting system could become something quite different from — and more dangerous than — what exists today. This is not a matter that a single persuasive conversation can shift; it is a matter of who controls the process by which the system comes into being.
The goodness of artificial intelligence is therefore subordinate to how faithfully the human community that designs and uses it adheres to natural law — the foundation of right and wrong that lies beneath positive statute. And the weight of that subordination falls overwhelmingly on the few who design the system, only secondarily on the many who use it. The real question Son's goldfish remark leaves us with is not how intelligent AI will become, but who controls its formation, and with what sense of natural-law responsibility.
