NTTVblog

Random thoughts from a wheelchair First, I would like to express my appreciation for the London Mathematical Society on this excellent initiative and thank the society for choosing our Institute for its first global meeting. On behalf of VIASM, I would like to extend our warmest welcome to our colleagues from the UK and all of you. In the capacity of scientific director of the VIASM, it incurs to me to say a few words on the scientific theme of this meeting, which is mathematics and AI. I take this opportunity to share some philosophical thoughts on this hot topic. Although I have no technical knowledge about AI, I can say that over the last half year or so, there has not been a single day without me reflecting on this fascinating phenomenon, which can be compared to a gigantic meteorite traveling toward our planet at the speed of light. Please bear with me as I indulge myself in speculation. When you listen to people like Sam Altman, you get the impression that his company spent billions of dollars to build this thing, which he’s sorry will destroy us all. According to Altman, this is a sort of man-made disaster that we have no choice but to contribute to it. We need to use all the water we have, heat up our atmosphere even more, to build this monster that will take away all our jobs. Doesn’t what Altman says make artificial intelligence look more like human stupidity? In the previous industrial revolutions, when steam machines and electricity replaced human muscles, large-scale disasters unfolded. Were wars to win new markets to absorb industrial surplus production, or unprecedented depression and unemployment direct consequences of the industrial revolution? It’s difficult to determine, but the advent of steam machines and electricity has likely contributed to large-scale disasters in the first half of the twentieth century. It would be, however, unfair to blame these great inventions, because at the core, they free us from not-so-enjoyable labor. We should blame our incapacity to renovate our socio-economic model to incorporate this whole new production capacity. This finally occurs in a somewhat organic manner, as the production workforce transitions to service work. The second half of the twentieth century was a period of global prosperity. Global prosperity was sustained by precarious world peace, except in some places, such as our own country. The nineteenth-century Industrial Revolution initially made human life worse, but ultimately improved it. But there were some unintended consequences. The second half of the twentieth century witnessed a brand new public health issue: for the first time in human history, a significant part of the population of some countries became obese. This could have been predicted because when we use our muscles less, we become obese. By extrapolation, we may predict that the public health issue for the second half of the twenty-first century could well be mental obesity. But it doesn’t need to be that way. If we believe Sam Altman, AI will replace humans in all the work that humans do, and as a result, humans no longer have to work. It will cause serious economic distress on one side and mental obesity on the other side. I believe that our economist will have the wonderful task of devising a viable economic model, in which many desk jobs created after the last industrial revolution will disappear with the advent of AI, but in which new jobs will be created. Muti Yung, who spoke here on Monday, shared a similar viewpoint. Although it is quite clear that many desk jobs will be replaced by AI, AI obviously cannot respond to all human needs. Beyond our basic survival and biological needs, as humans, we need to flourish intellectually and spiritually. No machine can replace us in our being alive and flourishing in intellect and spirituality. I want to believe that the progress of technology will reduce our mundane burden of staying alive and free our time and energy, allowing us to flourish. Again, I would like to leave the economic question to the economists, but I doubt that the current economic model, built on the constraints of the steam engine and electricity, would remain valid. In other words, we should not agree with Sam Altman, who believes that the goal is to build a super machine capable of crushing human intelligence, because that is the law of the free market. The law of the free market is a fascinating theory, a remarkable product of human intelligence, but it should not supersede our primary goal of staying alive and flourishing. Like many of you, I spent some time dialoging with different AI models, and I enjoyed their company. They are inhuman in their constant availability, but they are quite human in the sense that they are entertaining and unreliable. I don’t take any of their responses as truth, but the conversation with them helps me to keep thinking about a specific topic, eventually shaping my understanding of that topic. If we want to learn something, the current AI could already be a great learning mate. Yesterday, Bui Hai Hung gave a general presentation of LLM for dummies. He explains that at its core, LLM is still a statistical parrot, but one that has the whole internet in its memory. It is thus a talkative learning mate, which may impress us with its wit and clarity of expression. We must, however, be doubly cautious, as a statistical parrot is at its core unreliable. My experience is that when I ask a deeper question on a topic I know well, the answer is almost always wrong or misleading. But when I ask a question on a topic that I don’t know well, its response sounds believable. So please remain doubly cautious. From a mathematician’s point of view, the most obvious defect of LLM is its complete lack of ability to formulate a theory. By theory, I mean a system of definitions and basic logical statements upon those definitions, which allow us to speak of some specific phenomenon in a precise and concise way. As humans, we need concise formulation because we cannot parse too long sentences. That’s why we need theories. Furthermore, the more precise and concise the formulation, the more beautiful it appears to our eyes. On the contrary, the machine can parse long sentences. In fact, as I understand Bui Hai Hung’s talk, for the LLM, there is only one very long sentence that the system builds word by word. In its current architecture, LLM appears to be intrinsically incapable of constructing theory. Instead of building more powerful machines on the same architecture, which would require more water and electricity, the search for models incorporating similar constraints to human minds may be necessary to create an AI model that is a truly intelligent learning mate. From a teacher’s point of view, the challenge presents differently. The current AI models are already capable of solving most of the problems we normally ask our students. But for us, teachers, this has no positive value because we don’t need the response to our problems, but we want our students to find it. To put it briefly, AI in its current form does more harm than good to mathematics education. As we ask economists to find new modalities for economic transactions in the AI area, we have to ask ourselves to reinvent the way we teach mathematics and assess our students’ progress. It will take more than me speculating idly in this wheelchair to find out how, but I believe that solution exists. Whether it is inquiry-based learning or a return to oral examinations, we need to work things out systematically because the stakes are high. We don’t want to be in the situation again of asking our students to use the so-called “scientific calculator” simply because we haven’t thought out how to integrate real scientific computation into our curriculum. This is all I have to say. In the advent of AI we need to reassess every aspect of how we live collectively, how we exchange goods, services and ideas. AI is capable to free us from many boring and mundane tasks and let us to concentrate our minds on more essential questions pertaining to our intellectual and spiritual flourishing. Before getting there, without carefully thoughtfully and deeply reorganizing our collective existence, it may lead first to disaster with significant human tolls, before any improvement. Thank you for your attention. Source: from Ngo Bao Chau's Facebook Translated by Claude (AI) in https://mqtrinh.github.io/other/resources/ngo/

NTTVblog

Saturday, August 15, 2026

10 Bài Kiểm Tra Giúp Nhận Biết NÃO BỘ Còn TRẺ Hay ĐÃ LÃO HÓA SỚM

Random thoughts from a wheelchair ---AI and Mathematics Education

Friday, August 14, 2026

Trước Khi Gọi Cấp Cứu, Hãy Tìm Cách Giúp Tránh Nguy Hiểm Nếu Bạn Đau Tim và Ở Nhà Một Mình

Trứng và Podcast Tim Mạch

Niềm Tin

Solar Eclise Aug 12, 2026

What We Could Learn from the Responses to the Recent Earthquakes in Columbia and in Venezuela