Recently, I was watching the new dialogue book "the future of mankind, the future of AI" by shinmi Yamanaka and Yoshiko. Yamanaka is the director and professor of iPS cell research institute of Kyoto University and the winner of the 2012 Nobel Prize in physiology or medicine. Yu Shengshan is a chess player of general chess, a winner of the eternal seven titles (all won the "eternal" title in the seven events of general chess) and the National Honor Award. One is a strange man in the field of life science, the other is a superman in the field of human intelligence; A question about whether cutting-edge medicine can cure all diseases, and a reflection on why chess players lose to AI (Artificial Intelligence). This has triggered hot topics such as the future that mankind will be controlled by AI, AI will eventually replace mankind, and what the world will look like in a hundred years. Although we can't escape talking about the future with the knowledge source generated by the prior knowledge structure, and we still use the ancient mental structure of "nothing outside the heart" to discuss AI, the conversation itself is smart and wonderful enough, and the thinking circuit is clear and clear enough. Yamanaka announced to the world in 2006 that he had successfully created mouse iPS cells. In 2007, it announced the successful manufacture of human iPS cells. In the autumn of 2014, retinal cells were cultured with iPS cells for the first time in the world and surgically transplanted into patients with senile macular degeneration. IPS cells, fully known as "induced pluripotent stem cells", in short, are cells that have the ability to become any cell. Because it is made manually, it is called "induced". The brilliant prospect of iPS cells is that they can eventually cultivate internal organs. Therefore, Yamanaka said, "using iPS cells to cultivate organs is no longer a dream". But even so, Yamanaka believes that it is almost impossible to completely analyze a person's 3 billion base human gene. So far, in 30 years, human beings have only understood 10% of the biological mechanism of fertilized egg development, that is, 1%, and 99% still know nothing. It seems to be an unexplained "black hole", so it is meaningless to human beings. In the world of intelligent general chess, Yusheng good governance admits that the speed of AI catching up with human beings is indeed "much faster than expected". In March 2016, AI's alpha go defeated the world's top go player Li Shishi with an overwhelming victory of four wins and one loss, shocking the world. The terrible thing about alpha go is that it has a "heuristic", that is, it can give an answer that is not necessarily correct, but it can be estimated to determine the answer that roughly meets the requirements. Yusheng said that this is equivalent to "human sixth sense". In April 2017, the general chess software "ponanza" won two battles and two victories against the famous Sato Tianyan. Yusheng said that I was given the same hand by the software, and was modified step by step by it, saying "this is the correct solution, what a terrible world". As early as 1997, IBM's "dark blue" defeated world champion Kasparov for the first time. In short, machines can fight at full load for 24 hours and accumulate hundreds of thousands or millions of game data at the speed of one second game or 10 seconds game, which is completely beyond the reach of human beings. However, in Yusheng's view, the chess score of AI game first lacks continuity or inertia, which is a human beauty. Secondly, there is no concept of process, and each step points to the best hand in the current situation. Although it seems that every step is shocking, human beings will never choose this dangerous method. Because human beings have a steady psychological image of fear of losing, which AI can't have. Thirdly, AI will not play "entertainment chess" (or comfort chess). It is difficult for AI to lose the chess game after a stalemate, and to lose without showing landscape and make the opponent satisfied. The difficulty lies in how to cover up the intentional behavior. Therefore, in Yusheng's view, although random variables are constantly introduced, AI may also make creative work, but it has nothing to do with our real creation. Some creations can only be realized by human beings. AI is good at predicting what people like based on data, but human beings like unexpected nature, incredible intuition and inexplicable possibilities. These AI can't predict. Now AI can write poetry, but it has no textual meaning. "Sunset clouds and solitary ducks fly together, and the autumn water is the same as the sky" and "nostalgia is a small stamp. I'm here and my mother is there". Can AI write it? This shows that people are still people and things are things in the ultimate sense. Things need reductionism, and human things need phenomenology. In the book, Yusheng and Yamanaka share the same feeling that AI can make Bach's music, but can't write Haruki Murakami's novels. The reason why AI can't write Murakami's novels is that Yamanaka thinks it mainly lies in what we call intuition. AI can't have. Yusheng believes that AI is good at optimization, and it does not have what we call aesthetic power. It is almost impossible for AI to learn the beauty consciousness of "really beautiful" when human beings see natural scenery. This is reminiscent of Haruki Murakami in the Norwegian forest, who wrote that Lvzi asked Watanabe: what is the white smoke in the sunshine at the end of summer? Watanabe said he didn't know. Green son said that it was a cigarette burning tampons. It turned out that Lvzi went to a girls' school. She knew that at least 180 sanitary quilts were burned as garbage one day. Here, white smoke is not smoke, white smoke is girls, white smoke is physiology. Very reverse, very Zen. Can AI express this reverse and Zen? The two authors also mentioned that the "xiaodongjun" (whether the robot launched in 2011 can be admitted to the artificial intelligence project of the University of Tokyo) promoted by Professor Kiko Shinai of the National Institute of information of Japan gave up the examination of the University of Tokyo in November 2016. Why did "xiaodongjun" abandon the exam? The author did not explain. According to the author, it is because "xiaodongjun" has limited reading comprehension ability, that is, it lacks human common sense and logic. For example, human students can easily connect "so I want to drink cold" and "so I want to turn on the air conditioner" in English, but "xiaodongjun" makes a sentence "very cold to drink water" according to the basic principles of English grammar and the frequency of Internet appearance, so English scores can't go up all the time. In short, the book is lively and interesting in the form of dialogue, which collides with the thinking circuits of two authors in different fields. Although it is a serious topic such as "will mankind become extinct", it is easy to read and stimulates imagination. At the end of the book, the author's final thinking is: Although AI technology can't replace mankind no matter how advanced it is, it may help us solve other risks threatening human civilization, such as climate change, giant meteorite impact, volcanic eruption, and even nuclear war. AI may issue a severe warning to humans, so as to help humans restrain themselves, that is, because of AI, humans will be saved from extinction. Humans will coexist with AI. This reminds me of a picture in the book. In Yamanaka's research room, the plaque hanging on the wall is "everything in the world is a blessing in disguise". What is happiness and what is disaster in the world is really unpredictable. This is the wisdom of the Chinese ancients and, of course, the wisdom of mankind. "Unpredictable" is the human's mood towards AI. Shanghai Translation Publishing House can translate and publish this book in time and include it in the "translation vision" series, which may indicate this mood. In this way, it is necessary to talk about the future of mankind and the future of AI. Because the law of cause and effect only exists in our hearts. (outlook new era)