The Discussion That Began Because No Question Came In
Friday night was originally meant to be a time for questions and conversation with viewers. Recently, I had continued several broadcasts around comments about the absence of signs, no-sign, no-thought, and non-abiding. I invited questions about Buddhism, society, politics, economics, or personal life, but no specific question appeared in the chat that evening. While warming up, I therefore began with politics, moved into a criticism of the threefold training and the attainment of cessation, then spent the main part of the broadcast discussing artificial intelligence and the end of the labor market. At the end, I paid tribute to the late Venerable Myeongjin.
At first glance, these subjects seem unrelated. Yoon Suk Yeol and election-fraud conspiracy theories, the Buddhist fiction of the attainment of cessation, OpenAI’s claim about AGI, an inference chip designed with AI, the collapse of the labor market, and the life of Venerable Myeongjin do not seem to fit on one line. For me, however, they share a common principle. Do not accept authority and rumor as they are; examine the facts, face the changes approaching us, and act on the side of the socially vulnerable.
Leave the Chinese Election-Interference Conspiracy Theory Behind
A viewer commented that I should reconsider Yoon Suk Yeol. The suggestion was that China had interfered in the Korean election, that evidence of fraud existed, and that Yoon’s declaration of martial law may have been intended to expose it. I cannot agree. If unmistakable evidence existed, it could not have remained entirely absent from both the media and the court proceedings. Above all, if a defendant’s life and freedom were at stake, why would he withhold evidence capable of saving him?
This does not mean that the media, YouTubers, broadcasters, legislators, prosecutors, and judges are all trustworthy. They can lie, publish false reports, and conduct manipulated investigations. People still reveal appalling historical awareness, as when a commentator laughed while discussing the People’s Revolutionary Party case and provoked national outrage. Prosecutors and courts can also be wrong or biased toward power. But distrust of institutions does not turn an unsupported claim into evidence. There is a vast distance between acknowledging that every institution can fail and concluding that the YouTube video one wants to believe is true.
If a president declares martial law, aims weapons at the people, and tries to provoke North Korea in order to start a war, those acts alone amount to the conduct of a traitor. War would throw families, neighbors, and the entire nation into suffering. Anyone who wants to reverse that responsibility by invoking ‘Chinese election interference’ must present verifiable evidence, not slogans. If people who believe Yoon is innocent truly possess such evidence, they can give it to the court and his lawyers. Declaring it indisputable only inside a YouTube echo chamber is the frame of fake news.
I do not say this because I like the Democratic Party. I have repeatedly said that two-party politics is one of the major limitations of South Korean democracy. I vote for the Democratic Party as a second-best option because I dislike the People Power Party so intensely, not because the Democrats perform well enough. Korea has enough productive capacity and national strength to provide better welfare and reduce inequality. Why should I applaud politicians who leave behind a hellish society in which the birth rate is low, poor people give up dating, marriage, and children, and some are driven to suicide?
That does not make the known wrongs of Syngman Rhee, Park Chung Hee, Chun Doo-hwan, Lee Myung-bak, and Yoon Suk Yeol matters of mere taste. I said that the People’s Revolutionary Party case alone showed that Park committed crimes worthy of death, and I became angry enough to call Syngman Rhee a bastard when people proposed a memorial hall for someone who repeatedly massacred his own citizens. The language is rough, but my point concerns facts and victims, not partisan preference. Anyone who disagrees with my politics may respond freely, but the response must rest on known facts and verifiable grounds.
Believing in Cessation and the Threefold Training Does Not Produce Wisdom
From politics I moved to my recent criticism of Buddhism. I have called fortune-seeking religion, miracles, supernatural powers, past lives, and rebirth superstition. Recently I have gone further, questioning the Four Noble Truths, the Eightfold Path, the nine successive attainments, the attainment of cessation, one-thought, constant samadhi, and claims of non-discrimination. I also reject Theravada’s presentation of its canon as if it were the entirety of Early Buddhism. Theravada is one branch of sectarian Buddhism, and like Mahayana literature, its texts contain later editing and systematization.
The widely repeated threefold training of discipline, concentration, and wisdom must also be examined. It is commonly said that discipline makes concentration possible, concentration produces wisdom, and wisdom finally brings awakening. I reject that necessary chain. A chaotic life and a restless mind may make concentration and deep reflection difficult, so these things can influence one another to some degree. Yet failure to observe every rule does not make concentration impossible, and lack of meditative concentration does not prevent wisdom from arising.
Precepts are rules made by a community so that people can live together. They change with conditions and eras; they are not an absolute ladder to transcendence. The claim that ‘without discipline there is no concentration, and without concentration there is no wisdom’ is, in my view, nonsense. The same is true of the claim that one must enter the nine successive attainments and the attainment of cessation to eliminate greed, anger, and ignorance completely. When people do not understand why desire, anger, and ignorance arise but seek only a state that erases three words, Buddhism becomes a desire for transcendence rather than wisdom that observes the self.
Where does wisdom come from? It comes from observing, questioning, and reflecting. It does not come from believing that something exists because scripture says so, that cessation must exist because a teacher says so, or that greed, anger, and ignorance must simply disappear. Does a foolish person automatically lose foolishness by entering a thoughtless concentration? Such a person may instead remain silent before injustice that demands anger, while exploding over trivial and useless things. Wisdom does not eliminate anger itself. It understands when and why anger arises and gives one the ability to speak without fear when speech is necessary.
OpenAI’s Claim About AGI by the End of 2026
Because questions still did not appear, I turned to recent AI news. The most striking item was OpenAI’s statement that it might internally reach an AGI threshold by the end of 2026. Definitions of artificial general intelligence differ across people and organizations. Some call it intelligence usable across many fields like a human, while others set human-level performance as the standard.
The OpenAI definition I introduced in the broadcast was much stronger: ‘a highly autonomous system that outperforms humans at most economically valuable work.’ Such a system would not merely resemble a human. If it surpassed humans in law, language, scientific research, mathematical proof, programming, policy design, production, and semiconductor design, it would undermine the foundation of the human labor market.
If a system matching that definition is built internally by the end of this year, ordinary users might gain access early next year. Of course, the announcement may be hype or exaggerated marketing. Even developers cannot know precisely how capable a model trained with a new architecture and algorithm will be before it is completed. Predictions that once said fifty, thirty, ten, or five years have now moved to the end of this year. That history shows the speed of change more clearly than it proves the accuracy of any forecast.
Last year, I said that at the current pace AGI might arrive before 2028 or 2029. Even that could prove slow. This does not mean that we should accept a corporate announcement as fact. The important issue is not winning a bet about the date. It is preparing now for the social shock that would follow if intelligence surpassed people across most economically valuable work.
When the Price of Intelligence Collapses, the Labor Market Collapses
If AI becomes better than people at most valuable work, will it be more expensive or cheaper? It may initially require enormous computing costs, but it will probably become cheaper than paying a human wage. Semiconductor processes and hardware will become more efficient, and AI software architectures will improve. The cost of intelligence required for the same work will continue to fall.
What will the market choose if a system can perform, more cheaply and more effectively, the work of a mathematician solving a century-old problem, a Nobel-level scientist, an excellent programmer, a chip designer, a judge, a doctor, an artist, or a policy expert? We must ask whether humans or machines will conduct AI research, semiconductor design, construction, surgery, diagnosis, and scientific investigation. Left entirely to the market, human labor will struggle to compete on price and speed.
Elon Musk once spoke of the reversal point at which the machine intelligence produced in data centers exceeds the combined intelligence of roughly eight billion people. Beyond that point, AI no longer remains only a tool built by humans; it begins developing AI, chips, and production systems again. When machines improve their technology faster than humans can improve machines, a feedback loop emerges. That is the core of what many people call the technological singularity.
Astra and the Division of Labor Among Sixteen Agents
The reported source of OpenAI’s confidence was its next-generation internal model, Astra. According to the article I viewed during the broadcast, sixteen AI agents communicated, divided work, proved research-level mathematical problems, and performed programming tasks like virtual colleagues. One agent could oversee the overall design while others implemented separate parts and checked one another’s hypotheses and results.
This differs from a single model answering one question. Multiple agents debate, divide labor, review work, and move like a research team. Sam Altman’s claim that the model would go beyond recombining existing data and create meaningful new knowledge came from this arrangement. Whether it truly creates new knowledge or merely supplies marketing language must be judged from actual results. Yet AI is already imitating the collaborative structure of research and development, showing that labor-market disruption will not stop with routine automation.
There was also an explanation that development had slowed because of safety reviews following an autonomous hacking problem. Greater capability brings greater danger. Still, the important point is that a next-generation model is reportedly already being used for internal development before public release. If a company first uses stronger intelligence for itself than it offers outside users, technological inequality and the monopoly of intelligence will inevitably widen.
Artificial Intelligence Is Building Semiconductors for Artificial Intelligence
The second news item concerned Jalapeno, an inference chip jointly developed by OpenAI and Broadcom. In a fabless structure, companies such as Nvidia and AMD design chips while manufacturers such as Samsung, SK Hynix, and TSMC produce them. OpenAI can likewise design a chip specialized for its own models. A general semiconductor company must serve many customers, whereas OpenAI can optimize a dedicated chip for its own data centers and models.
The report presented during the broadcast said that Jalapeno surpassed Blackwell-based systems in inference and improved both throughput per watt and response latency. The development process matters more than any single benchmark. OpenAI reportedly used its own models for design and verification and moved from initial design to tape-out in about nine months. The coding agent Codex and the next-generation Astra model were also used to optimize kernels and computational blocks.
AI is already being used in semiconductor design, operating-system kernels, communication between AI systems, and higher-level software implementation. The report also said that AI-generated implementations outperformed earlier work by human experts. The question is therefore obvious: who will develop future chips and software? If AI produces faster and more efficient results than humans, companies will choose AI.
I called this time acceleration. We cannot suddenly create and employ one hundred Einsteins in parallel. Training a single expert takes ten or twenty years. AI, however, can be copied and run in parallel as power and computing resources increase. Development time compresses as though future technology were being pulled into the present. We do not need to wait for the singularity to begin someday; the feedback loop in which AI is deployed to develop AI has already begun.
The Monopoly of Technology and Intelligence
When a company uses a powerful next-generation model internally while withholding it from the public, technological monopoly becomes a monopoly of intelligence. Firms with capital, data centers, and chips can use AI to expand research, production, and development without limit. Individuals and countries without capital and resources must rent that intelligence or fall behind. Existing inequality may not merely remain; it may accelerate.
If an AI company uses its model to build a better chip, uses that chip to train a better model, and then uses the new model to improve software and hardware again, a closed loop forms. Companies inside that loop pull farther ahead, while outsiders find it increasingly impossible to catch up. We must ask not only the month and day on which AGI will appear, but who will own the intelligence and who will receive its benefits.
Bill Gates’s Proposal and Distribution After Labor
The final AI topic was Bill Gates’s warning. He said that AI is bringing a massive shock to labor markets while society remains unprepared, and he proposed deliberately reserving some jobs for people even where AI could do the work. Care work, where warmth and emotional exchange matter, might remain with human workers, slowing the pace of change.
He also discussed imposing strong taxes on companies earning enormous profits from AI and using the revenue for people who lose labor income. It is striking to hear a proposal that sounds socialist from a major capitalist at the summit of capitalism. A company might object, ‘Why should the money earned by the AI we developed be distributed to people who did not work?’ Yet when machines replace labor, wages disappear; when wages disappear, people also lose the money needed to live and consume. Capitalism that ignores distribution may destroy its own market.
I did not introduce Gates because I agree with every specific remedy. We do not know how long deliberately preserving human jobs can work. What matters is that the problem has entered public discussion. Society must debate taxes on excess AI profits, basic income, public services, protection of care, education, and creative labor, and ownership of productive resources and intelligence.
The President Must Convene a National Public Debate
What angered me most was the question of why we had to hear this from Bill Gates as an individual. What are Korea’s politicians and intellectuals doing? I have warned for years that the end of labor is approaching, yet the political world pretends not to see until the crisis stands at the door. AI may bring a civilizational transformation greater than the Copernican revolution. This cannot be handled with a few lines of policy and then forgotten.
I call on President Lee Jae Myung: stop wasting time on irrelevant matters and convene a national public debate. Bring together intellectuals, technologists, workers, companies, welfare practitioners, and citizens. Debate publicly not for several days but over several years, and ask the people for ideas. Society must begin choosing now how to respond to the future of AI, inequality, and the end of the labor market.
We cannot wait until vulnerable people starve or die by suicide and then announce a response. Those with capital and resources can increase production, research, and wealth without employing people, while those dependent on wages lose the foundation of life. Without preparation by the state and politics, AI may bring monopoly for a few and despair for the many rather than prosperity for all. Do not wait with closed eyes until everything has been taken.
In Memory of Venerable Myeongjin
Just before ending the broadcast, I remembered Venerable Myeongjin, who had died in Jeju a few days earlier. I cannot evaluate his entire life here. If there were errors and controversies, they should also be examined. Yet we should remember that he was one of the few Buddhist monks who consistently took interest in workers and social problems and spoke against power from the side of the poor and powerless.
I regard someone who stands with vulnerable people in reality and refuses to bow to power as more Buddha-like than countless monks who hold their heads high and boast of their enlightenment. I am grateful that there was a monk who could tell Yoon Suk Yeol to step down and who raised his own voice before authority and power.
The age of AI will also need Buddhist figures like Venerable Myeongjin. If Buddhism speaks only of meditation, transcendence, and private awakening while ignoring the collapse of labor, inequality, political power, and the suffering of vulnerable people, it is neither wisdom nor compassion. May his life be remembered for a long time. I thank him for standing against power on the side of the socially vulnerable.
Do not believe; question. Leave behind false conspiracy theories and the dead authority of scripture, and begin a public debate now about the crisis of labor and distribution that AI will bring. Wisdom is not quiet that turns away from reality. It sees facts and acts on the side of the vulnerable.
This article was edited by AI solely from the supplied video captions and description. No external sources were used.